The method: where HypnoSelf comes from

It started with seven books, which we read for ourselves. They convinced us of one thing: what you tell yourself every day ends up changing what you believe about yourself, and you can choose what you tell yourself.

We wanted to live it, not just read about it: write our own sentences, record them in our own voices, and listen to them every evening as we fell asleep. We built the tool to do that, first for ourselves, then for everyone. That’s how HypnoSelf was born.

Then we wanted to know what research says about it. This page follows that path in order: the books, then the studies, and at each step what we did with them in the app.

In short

• What has been measured. A placebo can do you good even when you know it contains nothing active: trials have measured this, with a modest effect on how people feel. For a ritual, that’s good news: being open about how it works can leave its effect intact. The moment you fall asleep is a state of its own for the mind, and one team measured it using arithmetic problems. What you say out loud is remembered better than what you read silently, and a study using word lists finds that a word you hear again in your own voice, on a recording, is remembered better than a word read by someone else. A habit often takes more than two months to take hold, and practising at the same time each day helps.
• What remains open. HypnoSelf as a whole has yet to be studied. Experience often comes before science: Coué had people repeat a sentence morning and evening a century before us, and the hospital in Liège was operating under hypnosis, with a local anaesthetic, years before studying what happens in the brain in that state (§2). Where there are no studies yet, we tell you where our choice comes from: most often, the books in §1.
• What HypnoSelf is. A personal development ritual: a conversation with our AI helps you write your own sentences, then you record them in your voice, and you listen to them every evening. If you are receiving care for a health condition, HypnoSelf complements that care.

Throughout this page, we speak in two ways. For the books, we say what their authors propose, staying close to what they wrote. For research, we say what has been measured: in whom, compared with what, with a link to the study.

Contents

1. At the start, seven books
2. Then, wanting to check
3. The conversation: looking at what holds you back, and choosing what you want instead
4. Personalisation: sentences that are yours
5. Your voice: what you say yourself sticks better
6. “I”, “you”, or a question? What we know about phrasing
7. The act itself: listening to a recording at bedtime
8. The evening and the window of falling asleep
9. Brainwaves: what these rhythms really are
10. Plasticity: why repetition matters
11. Habit: how long, really
12. Attention: what you notice depends on what you look for
13. Our commitment
14. All our references, described one by one

---

1. At the start, seven books

The essentials

From these seven books come the daily appointment, the bedtime timing, the conversation where everything begins, and the idea of recording yourself in your own voice.

Why these books come first

Because that’s where it all started. They are personal development and popular science books, widely translated and reprinted. Putting them first is our way of telling you where we come from before showing you what we found.

Success Through a Positive Mental Attitude, Napoleon Hill and W. Clement Stone

What they propose. That mental attitude can be learned, and that you change it through short phrases repeated often: self-affirmations repeated frequently, rapidly and with conviction, they write, reach the subconscious and make it respond. In their view, you also need to act straight away, without waiting until you feel like it. They explicitly link repeating these phrases to Émile Coué’s autosuggestion, and advise saying his formula several times a day, on waking and as you fall asleep. One of the businessmen whose stories they tell, the Australian Bill McCall, remembers his father reading Coué aloud to him when he was a child. We come back to Coué in §2.

What it became in the app. The daily appointment itself, and the idea that repetition builds the habit. The book gives ready-made phrases to learn by heart; we chose to have you write your own instead. You keep them, and you listen to them again evening after evening.

The Power of Your Subconscious Mind, Joseph Murphy

What Murphy argues. That the subconscious accepts whatever the conscious mind holds to be true, without sorting the true from the false, and that falling asleep is the most favourable moment to address it: the best time to impress your subconscious, he writes, is as you drift off to sleep. He pairs it with the moment of waking. He also takes up a technique from Charles Baudoin: condense your intention into a short sentence, in the present tense, and repeat it slowly, like a lullaby, in a state of drowsy relaxation. Our home page quotes one of his sentences, “Whatever you impress on your subconscious mind is expressed on the screen of space as conditions, experiences, and events”. From him, we took the moment and the method.

What it became in the app. The choice of moment: Murphy is the reason we placed the ritual at bedtime, with short sentences in the present tense. To that we added listening every evening and, when a belief pushes back, sentences that describe a path, whereas Murphy wants a sentence that states the change as already accomplished (§6). What research has found about this moment is in §8, and it is richer than we imagined.

The Biology of Belief, Bruce Lipton

What Lipton proposes. That our beliefs, true or false, have more influence on our biology than our genes do: beliefs, he writes, control biology. He describes the subconscious as an autopilot, far faster than conscious thought, running programs learned mostly before the age of six, starting in the womb. Between the ages of two and six, he writes, children spend more of their time in theta waves, a state he compares to the one hypnotherapists lead their patients into, where the mind is more suggestible and programmable. He also warns that a positive affirmation fails when it contradicts a program installed in childhood.

Where the 95% comes from. Our home page quotes two authors who each put forward a figure of 95%. Lipton gives it on his website rather than in this book: by his account, we spend at best 5% of our time in our conscious mind, and 95% in our subconscious. Dispenza gives his in Breaking the Habit of Being Yourself: by midlife, he says, about 95% of who we are is made up of programs that have become automatic. In both cases it’s a figure of speech, and research has measured something close to it. Researchers asked participants to note, hour by hour, what they were doing and what they were thinking about: roughly four in ten of their everyday behaviours were repeated almost every day in the same context, while they were thinking about something else (Wood, Quinn & Kashy 2002).

What it became in the app. The initial spark: the idea that working on your beliefs takes more than an intellectual exercise. We concluded that this work deserved proper time every day, and we drew two design choices from it. First, a conversation before you write a single sentence, to take account of his warning about affirmations that clash with what you believe deep down. Second, listening to your sentences in a relaxed state, without hammering them into yourself.

The Brain That Changes Itself, Norman Doidge

What Doidge describes. A psychiatrist and psychoanalyst, he tells the story of the researchers who showed that the brain changes its structure with activity, thought and experience, including in adulthood. He calls the other side of this adaptability the “plastic paradox”: the same plasticity that makes us more flexible can also make us more rigid, and some of our most stubborn habits are its product. He borrows an image from the researcher Pascual-Leone: by going down the same slope again and again, a sledge carves tracks, good or bad, that become hard to get out of. From a study of people learning Braille, where lasting changes only appeared after months of daily practice, he draws a simple lesson: you have to study slowly, regularly and in depth.

What it became in the app. The framework: a regular ritual, focused on what you have chosen. If repetition carves tracks anyway, you might as well choose which ones: that’s the conclusion we drew. More on this in §10.

The Spontaneous Healing of Belief, Gregg Braden

What Braden proposes. That most of the beliefs that limit us are subconscious, and that they take hold early, when we absorb the beliefs of the people around us. He writes that until the age of seven, our brains are in a “hypnogogic or dreamlike” state (the half-asleep state we pass through as we fall asleep). To change a belief, he suggests bringing it up into awareness, naming it, putting its opposite into words, then backing that opposite up with facts drawn from your own experience, in your own words. Each of us builds this program for ourselves, he explains, because everyone works a little differently. He goes so far as to see belief as the “program” that shapes reality itself.

What it became in the app. The conversation where everything begins: naming what holds you back, then looking for the sentence that replaces it for you. Braden is also where the “before you turned seven” on our home page comes from.

Breaking the Habit of Being Yourself, Joe Dispenza

What Dispenza proposes. A two-step approach: become aware of your automatic patterns, the thoughts you keep repeating and the emotions your body has memorised, then install something else through daily practice. When you become aware of who you are, he writes, you become conscious of your unconscious self. This is the book in which he writes that by midlife, about 95% of who we are is made up of programs that have become automatic. He names bedtime and waking as the two moments when the subconscious is easiest to reach. He says he studied hypnosis in 1981 and went on to teach it, and he himself likens meditation to hypnosis. Finally, he suggests saying certain sentences out loud, and recording the text of your guided meditations yourself.

What it became in the app. The order of the steps: first the conversation, which brings your automatic patterns to the surface, then repeated listening. This first step echoes widely studied approaches, including Beck’s cognitive model (Beck & Haigh 2014), on which our conversation is based (§3). This book also gave us the idea of having you record yourself.

You Are the Placebo, Joe Dispenza

What Dispenza argues. He starts from something medical research knows well, the placebo effect, and from a 2010 study by Kaptchuk and his colleagues that made a strong impression on him: placebos, he writes, worked even when patients knew they were taking one. Hence his question: if the body responds to an expectation, can we learn to produce that state without a pill? He argues that a thought repeated every day, paired with a strong emotion, ends up conditioning the body, while making clear that emotion speeds up the change without being essential to it. To change a belief, he suggests writing it down, then writing opposite it the belief you want to adopt and how you would feel. He also invites you to record yourself reading the text of your meditation.

What it became in the app. Sentences worded to be felt. The structure of the conversation, which puts what you want to leave behind side by side with what you choose. Recording your own voice: Dispenza suggests recording yourself, and we made it the heart of the app (§5). Above all, this book led us to the trials where people knew they were taking a placebo, which we describe in §2.

References for this section

10.1037/0022-3514.83.6.1281 · 10.1146/annurev-clinpsy-032813-153734

---

2. Then, wanting to check

The essentials

It was after building HypnoSelf that we turned to the research. The act at the heart of the ritual goes back a long way: as early as 1920, Coué had people say a sentence out loud, at bedtime and on waking. Trials have also measured the effect of a placebo people know they are taking: it helps them feel a little better, most often compared with taking nothing. For us, this means that a ritual can keep its effect when people are told plainly how it works.

The order matters

We designed HypnoSelf based on these books, and then we turned to the research, because we wanted to know, for ourselves, what it said about the tool we had built. What it taught us changed some of our choices, and we’ll tell you which ones.

This page is called “The method” because it tells the story of our approach. Here, research serves as a bridge between a conviction and what we know today. When it has measured something, we give you the figures. When something has yet to be measured, we tell you what our choice rests on.

What you practise here, your sentences recorded in your voice and listened to every evening, has yet to be studied as a whole. People often practise before they understand: experience raises questions, and research then comes along to explore them. Coué and the hospital in Liège, just below, are two examples.

A morning and evening ritual, a century before us

Émile Coué was a pharmacist in Nancy. In Self Mastery Through Conscious Autosuggestion, a short treatise already in circulation by 1920, he describes a ritual that thousands of people would go on to practise. Every morning on waking, and every evening as soon as they are in bed, he tells his readers to close their eyes and, without trying to focus their attention on what they are saying, to say one sentence with their lips, loud enough to hear their own words, keeping count on a string tied with twenty knots. That sentence is his formula, known in English as “Every day, in every way, I’m getting better and better”.

At bedtime and on waking, a sentence said loud enough to hear yourself, repeated every day without effort: here you can already recognise the moment, the voice and the repetition of the HypnoSelf ritual. We added your own sentences, which you record so you can listen to them again.

Coué knew nothing of what research measures today about falling asleep (§8), about memory for what you say yourself (§5), or about the effect of a placebo you know you’re taking (further down). He made progress by practising and observing, and others took up this kind of practice: Hill and Stone quote Coué, and Murphy takes up the technique of Charles Baudoin, who was working at the same time.

Coué had started with hypnosis. He stopped putting the people he worked with into a hypnotic sleep after noticing that suggestion also worked on someone who was wide awake. This autosuggestion is what we call “self-hypnosis” on this page: you practise it yourself, and you stay in control of what you listen to.

Coué also presented his practice as a complement to medical treatment, and he warned that it takes time. On both points, we tell you the same thing: HypnoSelf complements any health care you are receiving (§13), and it too takes time.

In hospital too, practice came before explanation

In the 1990s, at the university hospital in Liège, patients had thyroid and parathyroid surgery under hypnosis, with a local anaesthetic, instead of a general anaesthetic. The first paper, on parathyroid surgery, reports that 31 of 121 consecutive neck operations were done this way (Meurisse et al. 1999). A second paper from the same team reports 218 operations under hypnosis: in only 2 of them did the team have to switch to a general anaesthetic (Defechereux et al. 1999).

These two papers report on the operations carried out. The patients chose hypnosis themselves; nobody was randomly assigned. The papers show a practice already established in the operating theatre, where hypnosis almost always held until the end of the operation. Only in 2000, after years of operations, did the same team look at what happens in the brain when hypnosis eases pain, by scanning volunteers under hypnosis (Faymonville et al. 2000).

The word “hypnosis”, and what it refers to

The word is in our name, so we might as well say what it covers in research. Brain imaging observes changes in activity during hypnosis. A meta-analysis, that is, a piece of work that pools several studies on the same question, brought together 8 of them, 75 healthy volunteers in all: it identifies the brain areas that become active when a hypnotic suggestion targets pain (Del Casale et al. 2016). An imaging study of 57 participants, selected from 545 because they were either highly responsive or barely responsive to hypnosis, describes reduced activity in an area that monitors conflicts and errors, and increased communication between an area that helps you concentrate and direct your actions and one that perceives the body’s sensations (Jiang et al. 2017). This result has yet to be replicated by another team. Finally, a suggestion given under hypnosis can change how unpleasant a pain feels without changing how strong it is: this is the landmark experiment in the field (Rainville et al. 1997).

Imaging results vary from one study to another, and the most recent review calls for them to be replicated (De Pascalis 2024). As for how well hypnosis works, an overview brings together 49 meta-analyses and 261 studies: it covers clinical hypnosis, in sessions with a practitioner, and its strongest results concern pain and supporting patients during medical procedures (Rosendahl, Alldredge & Haddenhorst 2024). That’s a different setting from ours, where you practise on your own.

Research continues, notably in Liège, where a team is studying self-induced trance: a trance that engages the body, drawn from shamanic practices. It’s a recent field. The three papers we cite give an idea of it: a study of a single person, an expert practitioner of this trance (Gosseries et al. 2020); a study that gathers trained practitioners’ accounts of what they experience during the trance (Vanhaudenhuyse et al. 2024); and the detailed plan of a trial with people who have had cancer, published before its results (Grégoire et al. 2022).

A ritual can do you good, even when you know how it works

This is the result that shaped our approach the most. It takes Dispenza’s question further, with figures to back it up.

A team of researchers pooled 63 trials, 4,554 people in all, both patients and healthy people. In each one, people were randomly assigned either to receive a placebo, a product with no active ingredient, or to be in the comparison group. The placebo was given openly: people knew what they were taking. This is called an “open-label” placebo. This random assignment, which is what makes these trials randomised, means a difference can be put down to the placebo rather than to the profile of the people taking part. Pooling 63 trials, as here, gives an overall result that is more solid than that of a single trial.

The people who received the open-label placebo did better than the others. The effect size is 0.35, a modest effect: if you pick one person at random from each group, the one who received the placebo is doing better than the other about 60 times out of 100 (Fendel et al. 2025).

Two pointers for reading the figures on this page

The effect size is a number that tells you how big a difference is, whatever the number of participants. Around 0.2, it’s small; around 0.5, medium; around 0.8, large. To picture it, pick at random one person who practised and one from the comparison group. With no effect, the first would be doing better than the second half the time, like tossing a coin. At 0.2, it’s about 56 times out of 100; at 0.5, 64; at 0.8, 71. In the areas this page covers, effects are most often small or medium. Depending on the study, this number is written d, g or SMD: it’s the same idea.

The comparison group changes everything. Imagine a pupil who goes from 8 to 12 out of 20: four points of progress. But if the whole class gained three points because the exam was easier, the pupil’s real progress is one point. The comparison group is the rest of the class. That’s why, for every figure, we tell you what it was compared with.

This effect mainly concerns what people say they feel: when they are asked how they are doing, it is 0.39 (55 results of this kind in these trials). On objective measures, those that don’t rely on what the person reports, it is 0.09 (17 results), which is almost nothing. In most of these trials, the comparison group received nothing.

For example, one trial followed 173 students who were anxious ahead of an exam. For three weeks, some swallowed an open-label placebo pill every day; others simply imagined taking it. Compared with a group that did nothing in particular, both groups were less anxious, and the merely imagined pill did as well as the one people swallowed (Buergler et al. 2023). Another followed 68 people preparing for their driving test: those who received an open-label placebo were less anxious before the test than those who received nothing (Schaefer & Enge 2024).

For us, this result is freeing: a ritual can keep its effect when people are told plainly how it works. So we can be precise on this page without losing anything. We tell you what we suggest you do, and what similar practices have produced when they were measured.

Where your own experience comes in

That leaves one question only you can answer: does it work for you? The effect sizes on this page are averages, calculated across hundreds of people. Try it and keep it up for two months, our benchmark for a habit to take hold (§11), then see what changes.

References for this section

10.1097/00000658-199903000-00014 · 10.1089/acm.1999.5.509 · 10.1097/00000542-200005000-00013 · 10.1016/j.jphysparis.2016.01.001 · 10.1093/cercor/bhw220 · 10.1126/science.277.5328.968 · 10.3390/brainsci14020115 · 10.3389/fpsyg.2023.1330238 · 10.1016/j.clinph.2019.11.011 · 10.3389/fpsyg.2022.807741 · 10.1093/nc/niae036 · 10.1038/s41598-025-14895-z · 10.1038/s41598-023-29624-7 · 10.1038/s41598-024-56600-6

---

3. The conversation: looking at what holds you back, and choosing what you want instead

The essentials

Before your sentences are written, you talk about what holds you back and what you would like to experience instead. Several of our books start there: looking at a belief before choosing another one. One study finds that an argument that starts from your own concerns is more convincing. Meta-analyses show that reflecting on what matters to you has a small effect, found in very different areas. On its own, the effect of a motivational interviewing session fades over time. In HypnoSelf, the conversation has a follow-up: your sentences come back every evening.

It all starts with a five-to-ten-minute conversation with our AI, before your sentences are written. In it, you talk about what holds you back, what you feel, and what you would like to experience instead. Then, every evening, you listen to these sentences. This conversation grew out of our books. Several of its steps resemble practices that research has studied, and we then looked at what they produce.

What the books propose: looking at a belief before choosing another one

Braden proposes identifying a belief before replacing it with another. His six-key method, which the reader fills in with their own words, essentially consists of naming the belief that holds us back, stating its opposite, then gathering indisputable facts, drawn from our own experience, that support what we are aiming for. He sets it apart from a simple affirmation, because it asks us to state facts from our own lives. Faced with a belief that no longer serves us, he suggests that we “replace it with something new based in an experience that’s true for us: our inarguable truth.” Changing a belief is, in his view, perhaps the hardest thing there is, because it touches on our identity. To bring these beliefs to the surface, he has us list the qualities and flaws of the people who looked after us when we were children, and what we expected of them.

In Breaking the Habit of Being Yourself, Dispenza makes awareness the first task of change: observing your automatic thoughts, reactions and emotions until you know how to spot them. He warns that positive thinking never works on its own when the body has been feeling the opposite for years.

In You Are the Placebo, he writes that changing a belief means changing a subconscious program. To help you spot your own, he gives a list of common limiting beliefs: I’m shy, I never have enough time, life is hard. Before his meditation, he has you take a sheet of paper. On the left, you write two beliefs you want to give up. On the right, the ones you want to adopt, and how you would feel if you believed them.

Lipton takes the same line: in his view, there is no point repeating a positive affirmation that contradicts a message programmed in childhood, because that message will undermine all your efforts. Yet he leaves a door open: the conscious mind, he writes, can observe a programmed behaviour, interrupt it and create a new one. This is what he calls free will, and he adds that it is difficult.

Even Coué, whose formula is meant to be general, starts by making you think: he asks you to weigh up rationally what you want to suggest to yourself. The repetition comes afterwards.

Braden, Dispenza and Lipton reach the same place by different paths: you look at the belief and name it before choosing what to set against it. That is where our conversation comes from. It starts by helping you describe what holds you back: in which situations, with which emotions, and what you tell yourself in those moments. It also asks you how you would like to feel instead, and what would be different in your life. It is Dispenza’s sheet of paper, with what you are leaving on one side and what you are choosing on the other.

What is measured: reflecting on what matters to you

Here, research had a pleasant surprise in store for us. We had thought of the conversation as just a warm-up. Yet it always includes a moment when you say what matters to you, and that moment corresponds to a step that research has studied a great deal.

In psychology, this step has a name: self-affirmation. The word brings to mind the affirmations people repeat to themselves, but for researchers, self-affirming means reflecting on what really matters to you: your values, the things you care about. That is the meaning Claude Steele gives it in the theoretical chapter that founded the field (Steele 1988).

In studies in this field, people are asked to reflect on their values, most often before being given a message from outside, for example information about their health. What is best established is a form of protection: reflecting on your values helps you cope with a one-off threat, such as being assessed. In HypnoSelf, this reflection has its place in the conversation. The evening repetition draws on other research, which you will find in §10 and §11.

Most of the figures that follow come from meta-analyses of randomised trials (see §2). In these trials, one group reflects on its values; the other goes through the same steps, without that reflection. The effects are small, and they turn up in very different areas.

• Among secondary school and university students, the effect size on academic results or well-being is 0.41: between small and medium, and the authors themselves consider it weak (Escobar-Soler et al. 2023).
• In health studies, all participants read a message about a risk to their health, and some have first reflected on their values. In that group, the effect shows up at every stage: they accept the message more readily, have stronger intentions to change, and change their behaviour a little more (an effect size of 0.32 on this last point). The authors describe these effects as small but reliable (Epton et al. 2015).
• On well-being, in the general population, reflecting on your values improves self-image, general well-being and social well-being, and lowers what the authors call psychological barriers. According to the authors, these effects are small, and chance alone cannot explain them. They persist over time: for the drop in these barriers, the effect measured later even exceeds the effect measured straight away (Zhang et al. 2025).

In one study, students were randomly assigned to two groups: some began by reflecting on what mattered most to them, then everyone did a problem-solving task in which they knew they were being assessed. Those who had reflected on their values did better than the others, and the benefit was concentrated among those who reported high chronic stress, meaning stress that has set in for the long term (Creswell et al. 2013).

Two brain imaging studies link this reflection to something measurable. The first follows sedentary people. Those who reflect on their values before receiving a health message show more activity in an area at the front of the brain, behind the forehead, linked to thinking about oneself and to the value we place on things. In the weeks that follow, their physical activity also increases, measured objectively rather than self-reported (Falk et al. 2015). The second compares two ways of reflecting: on your values, or on your everyday activities. Thinking about your values produces more activity in brain areas linked to thinking about oneself and to reward, all the more so when those values look towards the future (Cascio et al. 2016). In both cases, people reflect on their values before receiving a message: that is the order of the conversation, which comes before your affirmations.

These results guided two choices in the app. The conversation always asks you what matters most to you right now; if values come up, several of your affirmations are anchored in them. And the idea of looking towards your future shaped the way we write: alongside sentences that say what you are (“I am…”), the conversation writes some that describe a path (“I’m learning to…”, “I’m becoming…”), and more of them when a belief pushes back.

What is measured: what you put into words yourself

Researchers ran two experiments, the first on physical exercise with students, the second on quitting smoking with smokers. Participants were randomly assigned to two groups: some wrote arguments, the others read someone else’s. In the first, those who had written their arguments rated them more favourably than those who had read another person’s, with a medium effect size of 0.63. The second shows what really makes the difference: the content matching your own concerns, more than having written it yourself (Baldwin et al. 2013). In the app, your sentences start from what you have just said, in your own words.

Other researchers looked at who you address your own arguments to. When the message goes against what you already think, you convince yourself more when you address your arguments to yourself than to someone else. When it goes along with what you think, it is the other way round (Briñol et al. 2012). The first case is the one that applies to your affirmations: “I am capable”, for someone who doubts it, goes against what they believe.

A third line of research compares going through an exercise with hearing it explained. A meta-analysis brought together laboratory studies that test, one by one, the ingredients of a single therapeutic approach. One of them is cognitive defusion, which means stepping back from a thought instead of fighting it: compared with conditions with no exercise, it helps people believe a painful thought less and be less bothered by it. Across these studies, the effects are larger when people actually do the exercise, through a metaphor or hands-on practice, than when it is simply explained to them (Levin et al. 2012). They are found just as much in people in distress as in volunteers recruited without any particular criteria. These studies take place in a single session, with measurements taken just afterwards. That is the spirit of the conversation: having you go through the process, on your own situation, rather than explaining it to you.

What is measured: what a motivational interviewing session produces

In its format, our conversation draws on motivational interviewing, a way of guiding a discussion that helps the person put their own reasons for change into words. Cochrane, an independent network whose reviews of trials are known for their high standards, has evaluated it in people with problematic use of substances such as alcohol or drugs (Smedslund et al. 2011). Compared with no treatment, the effect on substance use is large right after the session (0.79). Then it fades: small in the short and medium term, almost nil in the long term, to the point where chance could explain it. In HypnoSelf, the conversation has a follow-up: the appointment you keep every evening, where your sentences come back day after day.

In a motivational interviewing session, a person can talk in favour of change, or in favour of the status quo, that is, of what holds them back. Two meta-analyses brought together studies that analyse what is said during these sessions. The more time the person spends talking about what holds them back, the poorer the outcomes. What counts is the share of talk directed towards change, relative to the share spent on what holds them back (Magill et al. 2017; Pace et al. 2017). These associations are weak, and were found by observing sessions. The conversation takes this into account. It listens to what is blocking you, then brings you back to what you want to build.

What we built into the app

Over the course of this section, you have seen where each step of the conversation comes from. Two further choices complete the picture.

• It follows a well-known psychological framework. To identify what holds you back, the conversation draws on Beck’s cognitive model, while talking to you in everyday language. A review describes its current formulation: the way we automatically interpret a situation influences what we feel and what we do (Beck & Haigh 2014). The conversation then helps you find the exact words for a few beliefs that limit you, four at most.
• It rephrases, and you confirm. Before a single sentence is written, the conversation sums up what it has understood and asks you whether it is right. Your affirmations start from what you have confirmed.

References for this section

10.1016/S0065-2601(08)60229-4 · 10.3390/healthcare12010003 · 10.1037/hea0000116 · 10.1037/amp0001591 · 10.1371/journal.pone.0062593 · 10.1073/pnas.1500247112 · 10.1093/scan/nsv136 · 10.1037/a0029937 · 10.1037/a0027231 · 10.1002/14651858.CD008063.pub2 · 10.1037/ccp0000250 · 10.1037/adb0000280 · 10.1016/j.beth.2012.05.003 · 10.1146/annurev-clinpsy-032813-153734

---

4. Personalisation: sentences that are yours

The essentials

Your affirmations are written with you, from what you said in the conversation: that is what we added to what the books proposed. Research measures a small advantage for what is tailored to the person, compared with other active programmes that are the same for everyone.

Your affirmations start from your situation, in your own words. The books that inspired us went about it differently.

What the books propose: formulas, or a template to fill in yourself

Hill and Stone give ready-made formulas. They call them self-motivators: short phrases, like a slogan or a proverb, such as the one telling you that you can do it if you believe you can. In the book, trainers even choose the student’s self-starter for them, the formula meant to spur them into action. You choose yours according to your goal or the flaw you want to correct, learn it by heart, repeat it to yourself often during the day, and make it your own: Hill and Stone want it to become personal and, for a salesperson, to symbolise their own aspirations. They also have you write down your goal, with a deadline, and even a letter to yourself that describes your plans as already accomplished.

Coué takes the idea of the ready-made formula further: he gives just one, for everything. It is known in English as “Every day, in every way, I’m getting better and better.” Since it covers everything (“in every way”), he considers specific suggestions unnecessary. In remarks recorded in an appendix to his book, however, he advises adding, after the general formula, a sentence specific to your own case.

Braden comes closer to what we do. His method, mentioned in §3, is designed for your own use: he warns that “your program might not be effective for someone else.” It is up to the reader to fill it in with their own words.

So the books give either ready-made formulas or a template to fill in on your own. Writing these sentences with you, from what you have just said, is our contribution. In other fields, research has measured what tailoring to the person achieves.

What is measured: a small advantage, compared with other programmes

An intervention tailored to the person does better than a standard programme that is the same for everyone. That programme is itself a real intervention, so the comparison is demanding, and that is what gives this small advantage its weight. It amounts to an effect size of around 0.2 (see §2): if you pick one person at random from each group, the one who received the tailored intervention is doing better than the other about 56 times out of 100, compared with half the time if the two were equally good.

The key figure comes from a 2023 meta-analysis. It brings together randomised trials of structured psychological therapies, delivered by trained therapists, with people being treated for a disorder. Compared with a standardised programme, therapy tailored to the person comes out ahead with an effect size of 0.22, a small advantage (Nye, Delgadillo & Barkham 2023).

Independent researchers find the same order of magnitude in adolescents: interventions tailored to the person do better than standardised interventions (Li et al. 2024).

An older meta-analysis makes the same comparison for printed health prevention materials. A message tailored to the person does better than a generic message, with an advantage of the same order (Noar, Benac & Harris 2007).

Finally, a review published in 2026 looks at mental health apps used by students. Personalisation stands out as a factor that strengthens the effect of these apps on depression (Jabs et al. 2026).

These results concern therapies, prevention materials and mental health apps. They say one simple thing: what is tailored to the person does a little better than what is the same for everyone. It is the same order of magnitude as for self-affirmation in §3 (Epton et al. 2015).

What we built into the app

• Sentences written from your words. Your affirmations use your exact words, your context and your emotions, as you expressed them in the conversation. They start from what you want to experience instead. When values have come up, several of them are anchored in those values.
• The last word is yours. You can edit each sentence, remove some, add others. If a sentence rings false to you, change it.
• Then your voice. Once your sentences are ready, you record them yourself. What research says about this step is in §5.

What we take from this is that a sentence written in your own words, about your situation, speaks to you more than a generic sentence. §3 pointed the same way: content is more convincing when it matches what you are going through.

References for this section

10.1037/ccp0000820 · 10.3389/fpsyg.2024.1470817 · 10.1037/0033-2909.133.4.673 · 10.1111/eip.70207 · 10.1037/hea0000116

---

5. Your voice: what you say yourself sticks better

The essentials

In HypnoSelf, you record your sentences in your own voice. What you say out loud is remembered better than what you read silently, and a study using word lists finds a ranking: you remember best what you say, then what you hear yourself say on a recording, then what another voice reads to you. Our bet is that your voice also helps you feel what you are saying.

This is our most distinctive choice: you record your sentences yourself, in your own voice, rather than listening to an actor’s voice. Research backs it on one specific point, memory.

What the books say

Saying your sentence out loud is an old practice, and several of our authors recommend it.

Coué has people say his formula loud enough to hear their own words. For him, the sentence reaches the unconscious by being heard; you’ve already seen his full ritual in §2. Murphy advises saying your sentences out loud, slowly, feeling what you say. According to him, repeating them aloud, or simply mouthing them, keeps the mind from wandering and helps them sink into the subconscious. Hill and Stone describe a method in which you read your wishes out loud, to become fully aware of what you want.

Dispenza goes further, all the way to recording yourself. In Breaking the Habit of Being Yourself, he has you say certain things out loud: what you have kept hidden, or a “Change!” called out as soon as an old thought comes back. For his guided meditations, he first suggests downloading the sessions he makes available on his website, or else reading the text given in the book and recording it yourself. He says it again in You Are the Placebo, where he tells readers to feel free to record themselves reading the script he provides. For him, these guided sessions are like stabilisers on a bike, which you take off once you have made the approach your own.

In these authors’ work, the sentence can also come through another voice: in Coué’s sessions, it was the practitioner’s, monotonous and lulling, and Dispenza first suggests the sessions on his website. At bedtime, Murphy suggests hearing, in your imagination, the voice of a friend or loved one congratulating you.

Dispenza also stresses emotion: for him, a thought should come with an emotion, which he sees as an accelerator of change rather than a precondition. The link with your voice is our own: we think that hearing yourself say your sentences helps you feel what they say.

The production effect

Memory researchers know this effect well: what you say out loud is remembered better than what you read silently. It is called the production effect, and several meta-analyses, studies that pool the results of many others, support it. The most widely accepted explanation comes down to one word, distinctiveness. What you have said out loud stands out from the rest in your memory, and standing out helps you find it again (MacLeod et al. 2010).

The figure that matters to us comes from experiments where one group reads everything out loud and another reads everything silently. That is your situation when you record: you say all your sentences out loud, and you are compared with someone who reads them silently. The effect size is then 0.37 for recognition, meaning when you have to spot a word you already studied among new words (Fawcett 2013). It is an advantage between small and medium. If you pick one person at random from each group, the one who read aloud recognises the words better than the other about 60 times out of 100 (§2 explains how to read these figures).

An update of this meta-analysis finds the same advantage when it comes to recognising what you read (Fawcett et al. 2022). When you have to recall what you studied from memory, with nothing to prompt you, the advantage comes mainly from the contrast between saying some words out loud and reading others silently (Jones & Pyc 2013). In your ritual, where everything goes through your voice, it is therefore recognition that benefits: you recognise your sentences better when you hear them again.

This advantage lasts. It holds for up to a week, with word pairs, sentences and texts (Ozubko et al. 2012), a span that fits well with a daily ritual.

The same idea shows up in the generation effect: a word you came up with yourself from a clue is remembered better than a word you simply read. Compared with simple reading, the effect size is 0.40, between small and medium (Bertsch et al. 2007). Your sentences grow out of what you put into words in the conversation, which is a closely related act.

Your voice rather than someone else’s

Two results speak directly to the choice of your voice.

One study, run with students on word lists, separated two things that are often confused: speaking, and hearing your own voice. To do so, it added a condition very close to what the app does, where you listen back to a recording of your own voice made a little earlier. The result is a ranking. Words said out loud are remembered best, then those heard again in your own voice, then those another person read (Forrin & MacLeod 2017). It is the only study that separates the act of speaking from the identity of the voice in this way.

Another series of experiments looked at which voice you read with. Participants read words out loud, either in their usual voice or imitating a character’s voice. In their usual voice, words said out loud were remembered better than words read silently. In the character voice, this advantage disappeared, except when participants read more words in that voice than in their own (Wakeham-Lewis et al. 2022). With your usual voice, the advantage appears straight away, and that is the voice you use when you record your sentences.

Then there is the self-reference effect: what you relate to yourself (is this like me?) is remembered better than what you process for its meaning alone, or relate to someone else. A meta-analysis shows this (Symons & Johnson 1997). The advantage is smaller when the comparison is with thinking about someone else than when it is with processing meaning alone. Your sentences are about you, which gives you one more advantage.

What your brain does with your voice

Your brain handles your voice separately from other voices, even from a voice it knows well. An imaging study observed this by comparing, in each participant, their own voice with a friend’s (Kaplan et al. 2008). Another recorded the electrical activity of the brain in 15 adults while they heard a single vowel, spoken either in their own voice or in the voice of someone they knew well. One of their brain’s responses was weaker for their own voice. The authors take this to mean that the brain needs less work to pick out its own voice, even before it pays attention to it (Graux et al. 2014). Recognising a familiar voice, meanwhile, draws on a wide network across the brain, according to a review of imaging studies (Sun et al. 2023). What these studies measure is how a voice is perceived and recognised. The brain areas involved are described in §14.

Why your recorded voice surprises you

On a recording, your voice sounds different from when you speak. When you talk, you also hear it through the bones of your skull, and that part is missing from any recording (Orepić et al. 2023). That is why people often find their voice strange the first time they hear it. Some researchers compare this unease to what we feel in front of a robot or a doll that is almost human: it is almost your voice, but not quite (Kimura & Yotsumoto 2018).

One simple instruction helps. In one study, 41 socially anxious people gave a speech, then listened to the recording of it. Those who had been asked to listen to it as if it were a stranger’s showed a greater improvement in how they judged their own voice than those who listened with no instruction, and the effect still held at the next speech (Nilsson & Lundh 2016). It is a small study, but the advice costs nothing.

What we built into the app

You record your sentences in your own voice, and you listen to them again every evening. Coué made sure that whoever said his sentence heard it, and Dispenza suggests recording yourself. We made this the heart of the app. Research backs this choice on memory: what you say, then what you hear yourself say, is remembered better than what another voice reads to you.

A simple tip

For the first few evenings, or at the start of each listening session, say your sentences out loud along with the recording, if you are on your own. Then let go and simply listen: the rest of the session is there to help you relax.

We start with speaking because, in Forrin and MacLeod’s ranking, what you say out loud is remembered better than what you hear yourself say, which in turn is remembered better than what someone else reads to you. The app reminds you of this on the listening screen for the first few evenings.

If your voice bothers you at first, that’s normal: on a recording, it sounds different. Try listening to it as a stranger’s voice. The app suggests this when you play back your first recording. This instruction helped socially anxious people judge their own voice more favourably.

References for this section

10.1037/a0018785 · 10.1016/j.actpsy.2012.10.001 · 10.1037/cep0000302 · 10.1080/09658211.2012.699070 · 10.3758/bf03193441 · 10.1080/09658211.2017.1383434 · 10.1080/09658211.2022.2115075 · 10.1037/0033-2909.121.3.371 · 10.1093/scan/nsn014 · 10.7717/peerj.14976 · 10.1093/scan/nsu031 · 10.1098/rsos.221561 · 10.1371/journal.pone.0199443 · 10.1080/16506073.2015.1126633 · 10.1037/a0033337

---

6. “I”, “you”, or a question? What we know about phrasing

The essentials

HypnoSelf mixes several ways of phrasing your sentences: with “I”, with “you”, with your first name, as a question. Talking to yourself as “you” or by your first name helps you step back in a difficult moment; for succeeding at a task or motivating yourself, the studies that compare the forms with one another give mixed results. One study carried weight in our choices. In people with low self-esteem, repeating a ready-made positive sentence left them in a worse mood than repeating nothing. Your sentences are different: they are written from your own words, in a kind tone, and when a belief pushes back, they describe a path rather than a state already reached.

Within a single ritual, your sentences take several forms: with “I”, with “you”, with your first name, as a question, or with a subject other than you. The same idea can become “I’m making progress”, “You’re making progress” or “Why am I making progress?”

What the books say

Our books move from one pronoun to another. Hill and Stone mix “I”, “you” and impersonal maxims. They take up Coué’s formula, in the “I” form. They tell of a sales manager who told himself every morning that he felt healthy, happy and terrific. But the formula they put forward most is in the “you” form: “You can do it if you believe you can!” Their formula for taking action, the one meant to become a reflex, is also in the “you” form: you tell yourself to do it immediately.

Murphy writes his examples in the first person, in the present tense, and asks you to feel the sentence as already accomplished. Yet he also suggests imagining, at bedtime, someone close to you congratulating you: words addressed to you, coming from outside.

Coué prefers sentences that describe a change under way. Rather than an “I want”, he has people say that such-and-such is in the process of disappearing. He also rules out “I’m going to try”. His formula, which you’ve already read in §2, expresses progress: getting better and better. When he speaks to a child, he says “you”, and always in positive terms: the child is told they are becoming hard-working, never that they are no longer lazy.

The question form comes from personal development, outside our seven books. Noah St. John popularised it under the name “afformations”. He suggests replacing the affirmation with a positive question, for example “Why am I getting better and better at this?” The mind, he argues, then looks for an answer instead of resisting a statement.

Hill and Stone, Murphy and Coué agree on the form: short, positive sentences, in the present tense, repeated often. Lipton, Dispenza and Braden add a note of caution, which you’ve already come across in §3: a positive sentence that contradicts your experience too much doesn’t take hold.

Distanced self-talk

Addressing yourself by your first name or as “you”, rather than as “I”, means talking to yourself the way you would talk to someone else. Researchers call this distanced self-talk. In a difficult moment, it helps you step back.

Kross and his colleagues tested it in seven studies, with 585 participants in total, in situations of social stress: speaking in public, making a good impression at a first meeting. Compared with those who talked to themselves as “I”, the participants who distanced themselves in this way were judged to perform better by outside evaluators. They felt less distress, experienced the situation more as a challenge than as a threat, and ruminated less afterwards (Kross et al. 2014).

More recent work points the same way. Orvell and colleagues show, in two experiments, that the benefit holds whether the difficult situation you reflect on is mildly or highly emotional, including in more vulnerable people; they add that its use in therapy remains to be examined (Orvell et al. 2020). Another study observed the brains of people talking to themselves by their first name. An area towards the front of the brain, along its midline, which is heavily engaged when we think about ourselves, became less active, and an electrical signal that accompanies emotional reactions also dropped. The signs of effort to control oneself stayed at the same level: the authors see this self-talk as a low-effort way to regulate your emotions (Moser et al. 2017). The part of this experiment that measured the brain’s electrical activity was later repeated exactly, with 67 participants. It was the first time anyone had done so, and the result held (Webster et al. 2022).

The broadest meta-analysis in the field covers psychological distancing in general. Compared with conditions without distancing, it finds a medium effect on felt emotion (0.52), and a smaller one on self-directed emotions such as shame (0.29) (Moran & Eyal 2022).

The pronoun itself

The “you” form has its founding study. In three experiments, giving yourself advice as “you” rather than as “I” helped people solve anagrams, puzzles where you unscramble the letters of a word. “You” also strengthened the intention to solve more of them and, in another experiment, the intention to exercise (Dolcos & Albarracín 2014). Other studies that compare “you” and “I” directly give mixed results (Magata & Tanaka 2024), which is one of the reasons your sentences mix the forms.

Talking to yourself with kindness

In this field, self-compassion has stronger support than positive affirmation. It means talking to yourself with the kindness you would show someone you love.

The trials that test it offer structured programmes, led by practitioners, often in groups. A meta-analysis of 21 randomised trials, 1,285 participants in total, finds medium effects on depression, anxiety and well-being against comparison groups, and an effect between medium and large on self-compassion itself (0.70) (Kirby, Tellegen & Steindl 2017).

On HypnoSelf’s exact target, the critical inner voice, another meta-analysis, pooling 19 articles and 1,350 participants, finds a drop in self-criticism: a medium effect (0.51), against comparison groups. Longer interventions do better than short ones, and the effects are larger against groups that did nothing than against groups doing another activity (Wakelin, Perman & Simonds 2022). Duration matters, which supports a ritual kept up over weeks, like our two-month benchmark (§11).

We also came to this framework by an unexpected route. In two experiments, participants wrote for three minutes about a value that mattered to them. They then showed more kindness towards themselves, and it was this kindness that led them to behave in ways more focused on others (Lindsay & Creswell 2014). This is what links the conversation about your values, in §3, to self-compassion.

A ready-made sentence, or a path

One result made us prefer kindness to the positive slogan. In people with low self-esteem, repeating a ready-made positive sentence, “I’m a lovable person”, left them in a worse mood than repeating nothing. When they were asked to think only about what made the sentence true, their mood was worse than when they could also think about what made it false. In people with high self-esteem, repetition improved mood, but only to a limited extent (Wood, Perunovic & Lee 2009). In this study, the sentence was the same for everyone, imposed, and repeated during a single session. Yours come from what you said in the conversation, and our bet is that a sentence born of your own words spares you this effect. This result also shaped the form of your sentences, described below.

Mental contrasting played a part too. It means picturing your goal and the real obstacle, for example seeing yourself running three times a week, then thinking about how tired you feel in the evening. A meta-analysis studied it in 1,528 participants, for changes in health behaviour. Against comparison groups, its effect is small at four weeks (0.28); in the fewer studies that followed participants for up to three months, it reaches 0.38, between small and medium (Cross & Sheffield 2019).

What we built into the app

Several forms, mixed together. Within one set of sentences, we mix “I”, “you”, your first name if you have given it, questions with “I”, and sentences whose subject is something other than you, such as “My needs matter”. “You” and your first name are there to help you step back, since those are the forms research used to measure that effect in difficult moments. “I” is there because it is the most natural to say and to hear yourself say. The question form comes from Noah St. John, and we keep it because it speaks to some people. This mix is a design choice, drawn from the books as much as from research: the books themselves move from “I” to “you”, and the studies on the pronoun give mixed results. If one form doesn’t speak to you as much, edit or remove those sentences.

Sentences that leave room for the path. When a belief pushes back, the conversation adds to your sentences versions that describe a path, such as “I’m making progress, even if I’m not there yet”, alongside sentences such as “I’m a confident person”. They fit better with what we saw about the ready-made sentence and about mental contrasting, and they ring truer. On this point, we are closer to Coué and his change under way than to Murphy, who wants the sentence to state the change as already accomplished.

The tone of kindness. The conversation helps you write your sentences in the tone you would take with someone you love.

The goal. The ritual aims to strengthen the feeling that you can act, rather than to persuade you that everything is fine. This is the sense of self-efficacy, described by Bandura 1977. §11 puts figures on how this feeling changes what you do.

References for this section

10.1037/a0035173 · 10.1177/2167702620951539 · 10.1038/s41598-017-04047-3 · 10.1016/j.ijpsycho.2022.05.003 · 10.1177/10888683211069025 · 10.1002/ejsp.2048 · 10.1371/journal.pone.0305251 · 10.1016/j.beth.2017.06.003 · 10.1002/cpp.2586 · 10.3389/fpsyg.2014.00421 · 10.1111/j.1467-9280.2009.02370.x · 10.1080/17437199.2019.1594332 · 10.1037/0033-295X.84.2.191

---

7. The act itself: listening to a recording at bedtime

The essentials

What you do each evening brings together advice from Coué and from our books: a daily appointment, short sentences repeated like a lullaby, on a recording you make yourself. In adults who slept poorly, listening to 25 to 60 minutes of recorded music every day improves sleep as they perceive it, and no adverse effects were reported. It is a bedtime aid; to treat insomnia, there is a dedicated therapy, carried out with a professional.

What the books say

Every evening, with HypnoSelf, you listen to sentences you have written and recorded in your own voice. This act brings together four pieces of advice taken from our books, and from Coué before them.

First, the daily appointment. Hill and Stone advise repeating Coué’s affirmation several times a day. Coué himself set two times: in the morning on waking, and in the evening as soon as you are in bed (§2).

The lullaby comes from Murphy. He suggests slowly repeating a short sentence, like a lullaby, in the evening and in the morning, or just before falling asleep.

The ear matters too. Coué asks you to say his formula loud enough to hear yourself: that way, he argues, it enters the unconscious mechanically, through the ear. Murphy also advises saying your sentences out loud.

Finally, the recording comes from Dispenza, the only one of our authors to suggest recording yourself reading the text of his guided meditations, then listening to that recording while you meditate. §5 covers his suggestion in more detail, along with what research has measured when you hear your own voice played back.

What is measured: listening to a recording every day

This is the evidence that comes closest to what you actually do. A Cochrane review, one of the gold-standard research syntheses in medicine, brought together 13 trials and 1,007 adults who slept poorly in total (Jespersen et al. 2022). In each trial, chance decided who would listen to recorded music every day, for 25 to 60 minutes, over a period of three days to three months, and who would listen to nothing or receive their usual care.

Sleep quality, as participants perceive it, improves by 2.79 points on the PSQI, a standard questionnaire scored from 0 to 21, where a lower score means better sleep. Relative to the usual differences between people on this questionnaire, that is a large effect, according to the pointers in §2. The authors rate the certainty of this result as moderate. The time it takes to fall asleep, as people report it in the same questionnaire, also goes down, with low certainty.

Another analysis points the same way. It brought together 20 trials and 1,339 patients with insomnia, and compared twelve approaches with one another, including where no trial had put them head to head. Listening to music comes out on top, with a medium effect on sleep quality (effect size of 0.61) compared with usual care, obtained mainly through these indirect comparisons (Feng et al. 2017).

So the perceived benefit of listening to music every day is documented, and no adverse effects were reported, in trials that together included more than a thousand people. Your ritual borrows the same act, with your sentences in place of the music.

If your sleep worries you, there is a dedicated treatment: cognitive behavioural therapy for insomnia, carried out with a professional, which works on the habits and thoughts linked to sleep. Trials find it has a large effect on the severity of insomnia (van Straten et al. 2017). The therapy treats insomnia; evening listening keeps you company at bedtime.

A recording alone can be enough to help you relax. In one trial, 60 healthy students were randomly assigned to four groups of 15. Three groups listened just once to twenty minutes of recorded instructions (muscle relaxation, deep breathing or guided imagery); the fourth was the comparison group. In the first three groups, relaxation rose more than in the comparison group (Toussaint et al. 2021). It is a small laboratory trial, but it shows that an audio recording can relax you without a practitioner beside you.

What we built into the app

The pieces of this act come from the books, and putting them together was our work.

Three choices that research sheds light on.

You know what you are listening to. Your sentences are spoken plainly, with nothing hidden, and you are the one who chose them. A ritual practised in full knowledge of what it is can do you good: the trials in §2 measure this, with a modest effect on what people feel.

Listening every evening. In the music trials, people listened to a recording every day, for days or weeks, and the improvement was in sleep as people perceived it. Your ritual follows the same rhythm.

A simple act. In the evening, all you need to do is start the recording of your sentences, made in your own voice (§5). A simple act becomes a habit faster than an elaborate routine (§11).

References for this section

10.1002/14651858.CD010459.pub3 · 10.1016/j.ijnurstu.2017.10.011 · 10.1016/j.smrv.2017.02.001 · 10.1155/2021/5924040

---

8. The evening and the window of falling asleep

The essentials

Four of our sources point to bedtime and waking as the moments to speak to yourself. One study shows that the threshold of sleep is a state of its own: 83% of those who had spent a little over a minute there on average then found the hidden rule in an arithmetic problem, compared with 31% of those who stayed awake. Once asleep, the brain still processes sounds, and it is when awake that it understands sentences: so your ritual stays with you up to that threshold.

Four sources, the same moments

Murphy, Hill and Stone, Dispenza and Coué all point to the same moments: in the evening as you fall asleep, and in the morning as you wake up. Each has their own reasons.

For Murphy, as you read in §1, the most favourable moment to address the subconscious is when you are falling asleep, along with the moment you wake up. Elsewhere, speaking of going to bed at night and waking up in the morning, he writes that these are the best times to impress the subconscious. In his view, when you are drowsy, effort is kept to a minimum, and opposing thoughts stop getting in the way. His technique comes from Charles Baudoin, who described it in 1910: repeating a short sentence over and over, like a lullaby, in a drowsy state.

Hill and Stone name two moments for repeating Coué’s affirmation: when you wake up and as you are falling asleep. For them, bedtime is simply a set time to do it, whereas Murphy and Dispenza see it as a moment when the subconscious is more receptive.

Dispenza talks about it in both his books. In Breaking the Habit of Being Yourself, he writes that the door to the subconscious opens at two moments: when we go to bed at night and when we wake up in the morning. In You Are the Placebo, he explains it in his own way: as we fall asleep, he says, the brain passes through slower and slower rhythms, and it travels the same path in reverse on waking. §9 explains what these rhythms really are. He himself says he prefers the morning.

Coué puts the morning on waking and the evening in bed on the same footing. At first he believed that suggestion only worked well during sleep, and he would put the people he worked with to sleep. He gave this up when he found it was unnecessary, and that telling people they were about to sleep triggered fear and resistance. He noticed, though, that people, lulled by the monotonous tone of his voice, often fell asleep of their own accord. He also took insomnia as an example of what effort makes worse: the more you want to sleep, the more restless you get.

Murphy, Hill and Stone, and Dispenza talk about what you do just before sleeping or on waking; in their books, the night is for sleeping. In Murphy, for example, you repeat or imagine your sentences yourself, then fall asleep. We wanted to know what research finds at that moment.

What is measured: one minute on the edge of sleep

One study shows something precise about this moment: the first stage of sleep is a special state for thinking. Those who spent a minute there found the hidden rule in a problem far more often than those who stayed awake. It measured this on arithmetic problems.

At the Paris Brain Institute, researchers recruited 103 healthy participants, chosen because they fell asleep easily, and who had been partly deprived of sleep the night before. They gave them arithmetic problems that could be solved much faster using a hidden rule. Between two sets of problems, each person took a 20-minute break, semi-reclined, eyes closed, holding an object in one hand, following a technique attributed to Edison. During that time, the electrical activity of their brain was recorded, to tell whether they stayed awake or fell asleep.

Of those who had reached stage N1, the very first stage of sleep, the one you pass through as you fall asleep, 20 out of 24, or 83%, then discovered the hidden rule. Of those who stayed awake, 15 out of 49, or 31%. That is a success rate 2.7 times higher. Among those who had slipped further, into stage N2, where sleep settles in, only 2 out of 14 found it, or 14%. In this study, the advantage lies at the threshold of sleep, and it is lost in those who fall more deeply asleep.

This time in N1 was brief: participants had spent a little over a minute there on average, with wide variation from one person to the next (Lacaux et al. 2021). When you listen to your ritual as you get into bed, it keeps you company on the way to this moment, as you tip over into sleep.

A more recent study, on a small group of participants, looks instead at the thoughts that come up as you fall asleep: it has participants hear a specific theme during stage N1. The more often this theme comes back in those thoughts, the higher the creativity measured after sleep, and the more original the answers, far from the most obvious associations (Horowitz et al. 2023). Its principle is close to HypnoSelf’s: content heard while falling asleep.

Asleep, the brain still hears; it understands sentences when awake

A study of 22 participants looked at what the brain does during the transition into sleep: it goes on correctly sorting the sounds it hears, real words on one side, made-up words on the other, and it keeps an implicit trace of them, unconscious but measurable. On waking, though, these words had been forgotten, and were confused with new words (Andrillon & Kouider 2016). The sleeping brain even processes sounds all the way up to the cortex, its outer layer, during the brief bursts of activity typical of light sleep (Jourde & Coffey 2024).

Understanding what you hear is another step. The brain tracks the sounds of speech in a comparable way awake and asleep; it tracks words, phrases and sentences when awake (Makov et al. 2017). An affirmation is a sentence: you understand it when you are awake. That is why your ritual comes just before sleep, while you still understand what you hear.

And the morning?

The morning has consistency on its side. One study followed the users of a meditation app for 180 days. Over the months, practice drops at every time of day, but it drops least in the morning. The more of their sessions a person did in the morning, the longer they kept going; the later in the evening they did them, the sooner they stopped (Berardi et al. 2023). The details are in §11.

One study describes a window as you fall asleep, another a practice that holds up better in the morning: that is why the app offers you both moments.

What we built into the app

The ritual at bedtime. Murphy is the one who led us there, and Hill and Stone, Dispenza and Coué also point to bedtime. The Paris study looks at that same moment and observes that those who had spent a minute on the threshold of sleep then found the hidden rule in an arithmetic problem more often than those who stayed awake.

A ritual that stops by itself. It accompanies you as you fall asleep, then the sleep timer stops playback after 30 minutes: you understand your sentences while you are awake, and after that, sleep takes over. You can change this setting.

The morning, too. The app offers you both moments, just as our four sources already set them side by side: the evening to take you up to the threshold of sleep, the morning to help you stay consistent. Dispenza, for his part, says he prefers the morning.

References for this section

10.1126/sciadv.abj5866 · 10.1038/s41598-023-31361-w · 10.1093/nc/niw014 · 10.1093/pnasnexus/pgae479 · 10.1523/JNEUROSCI.0168-17.2017 · 10.2196/42482

---

9. Brainwaves: what these rhythms really are

The essentials

Brainwaves are real, measurable rhythms, and theta goes with drowsiness and falling asleep. Two of our authors see the slow rhythms, theta above all, as a door to the subconscious. In the evening, your ritual keeps you company through exactly this passage into sleep.

Two of our authors, Joe Dispenza and Bruce Lipton, give brainwaves an important place.

What the books say

In Breaking the Habit of Being Yourself, Dispenza devotes a whole chapter to brainwaves. He describes five rhythms, from slowest to fastest, from delta to gamma, with their frequencies. For him, theta in adults is the in-between state, half awake and half asleep, in which a hypnotherapist reaches the subconscious. From this he draws his thesis, that we are easier to program in the theta state. According to him, meditation takes the brain down from beta to alpha, then to theta, a state in which the body sleeps while the mind stays awake. He sees childhood in the same light: between the ages of two and six, in his view, children are mostly in theta, and so they accept what they are told.

In You Are the Placebo, he returns to these rhythms and goes further. The slower the waves, in his account, the easier it is to reach the subconscious, and theta is the state in which we are most suggestible, that is, most open to suggestion. He warns that you can’t slip into it on command. He describes falling asleep as a descent, from beta to alpha, then to theta and delta, and waking up as the same path in reverse: for him, that is what makes bedtime and waking the two most favourable moments (§8).

Lipton, for his part, relies on data from Rima Laibow, whom he cites. Up to the age of two, he writes, a child’s brain works mostly in delta waves; between two and six, it spends more time in theta waves. He links these frequencies to hypnosis: hypnotherapists, he argues, bring their patients’ brains into delta and theta because these low frequencies make the brain more open to influence and to programming. This is how he explains why young children take in so much information.

What is measured: real rhythms

Your brain is constantly producing rhythmic electrical activity. It is measured on the scalp, with electrodes: this is the electroencephalogram, or EEG. This activity is then sorted into frequency bands, expressed in hertz (Hz); one hertz is one cycle per second. These are measured rhythms, regular patterns that we observe and associate with certain states.

The boundaries of these bands are shared reference points, set by the standard glossary of the International Federation of Clinical Neurophysiology; clinical neurophysiology is the medical discipline that records the electrical activity of the nervous system (Kane et al. 2017). Here they are, with the state each one is associated with:

• Delta, below 4 Hz (from 0.5 Hz in practice), dominates deep sleep.
• Theta, from 4 to 8 Hz, goes with drowsiness, daydreaming and falling asleep.
• Alpha, from 8 to 13 Hz, rises when you close your eyes and your attention relaxes.
• Beta, above that, dominates when you analyse, plan or talk things through.

Dispenza’s boundaries are very close to these, and his description of falling asleep matches what these bands cover: theta goes with drowsiness and falling asleep, and delta dominates deep sleep.

What we built into the app

Listening while relaxed, rather than forcefully repeating your sentences to yourself. We took this idea from Lipton (§1). Your ritual keeps you company as you slip towards sleep, that drowsy moment which measurements associate with theta (§8).

References for this section

10.1016/j.cnp.2017.07.002

---

10. Plasticity: why repetition matters

The essentials

Intensive motor or spatial learning, such as juggling or learning the map of a city, reshapes the brain, including in adulthood, when it is practised for months or years. These changes need upkeep, and partly fade when you stop. What we take from this is consistency. Your evening listening relies on another effect, measured in the laboratory, within a single session, with sentences that state facts: a sentence you hear again becomes familiar, and what is familiar seems truer.

Plasticity is the brain’s ability to change its structure and the way it works according to what it experiences and what it practises. This is well established, and research has also measured what it takes.

What the books say

Doidge, whom you met in §1, calls the other side of plasticity the “plastic paradox”: it makes us flexible, but it also carves tracks, like a sledge going down the same slope again and again, and those tracks are hard to get out of. He counts this among the most important lessons of his book. To change tracks, you need constraints, which Pascual-Leone, from whom he borrows the image of the sledge, calls roadblocks.

Another of Doidge’s ideas: plasticity is competitive. What you stop practising gives up its place to something else, and a bad habit gains a little ground every time you fall back into it. That is why, he argues, unlearning is harder than learning.

To make the case for consistency, he draws on a study by Pascual-Leone, who followed young blind people as they learned Braille, two hours a day, five days a week. Measured on Fridays, the changes in their brains came quickly, but remained fragile. Measured on Mondays, after the weekend break, they took six months to appear, then remained stable. For Doidge, this is the fable of the tortoise and the hare: lasting changes come from slow, steady work.

Finally, he sets a condition, which he takes from the researcher Michael Merzenich: attention. A task done mechanically does change the cortex, the brain’s outer layer, but only temporarily; for changes that last, you need sustained attention. In the chapter he devotes to Merzenich, he writes that with enough work and attention, we are able to rewire our brains.

For Dispenza, repetition is the central mechanism. In You Are the Placebo, he sums up Hebb’s rule by saying that when nerve cells are activated together, they connect together. He argues that repeating a thought strengthens the connections between neurons, and that without repetition, new connections disappear within three weeks; for this he draws on Kandel’s work. In Breaking the Habit of Being Yourself, he takes up the same law, with its flip side: cells that stop being stimulated together disconnect. There he lists four ingredients for changing the brain: learning, receiving instruction, paying attention and repeating. Attention is on the list, as it is for Doidge.

Lipton, for his part, talks about conditioning. This learning by association, of which Pavlov’s dogs are the classic example, makes the neural pathways between a trigger and a response permanent, in his view, and these pathways are our habits. Murphy, Coué, and Hill and Stone simply talk about habits, formed through repetition.

What is measured: learning that changes the brain

Take London’s taxi drivers. To get their licence, they have to learn the layout of the city’s streets.

A team of researchers followed 79 trainees, scanning their brains at the start of their training, then three to four years later. Other people, who were not taking this training, were followed alongside them for comparison. At the start, the trainees who would go on to qualify, those who would go on to fail, and this comparison group looked alike on these measures. At the end, in those who had obtained their licence, grey matter, the part of the brain rich in the cell bodies of neurons, had increased on both sides of the brain, at the back of the hippocampus, a deep structure used for memory and for finding your way around. In those who had failed, although they were enrolled in the same training, it had stayed stable (Woollett & Maguire 2011). Since they all started from the same point, it really was the learning that reshaped the brains of those who qualified.

This reshaping took a great deal of work. Those who qualified reported around thirty-four hours of study a week, for three to four years. What set them apart from those who failed was precisely the amount: they reported twice as many hours of study a week.

Another example covers a shorter period. Of 24 adults, 12 spent three months learning to juggle; the other 12, with no training, served as the comparison group. The jugglers’ grey matter increased, then, three months after they stopped training, it had partly gone back towards its starting level (Draganski et al. 2004). What the brain builds this way needs upkeep, and that is an argument for consistency rather than for an intensive burst.

“Cells that fire together wire together”

You have probably come across this phrase before. Dispenza puts it in his own way, as we saw above. Doidge attributes the short version to the neuroscientist Carla Shatz, and points out that Freud had put forward the idea sixty years before Hebb.

In 1949, the psychologist Donald Hebb put forward a postulate, a working hypothesis: if a nerve cell repeatedly takes part in making another cell fire, that is, in making it send out an electrical signal, the connection between the two grows stronger. It was a theoretical idea about how learning might be written into nerve tissue, and it guided half a century of research.

Experiments have confirmed a more precise form of the principle: a synapse, the point of contact between two neurons, changes according to the order in which the two neurons fire, within a window of a few tens of milliseconds (Caporale & Dan 2008).

What the image gets right: what is practised together, repeatedly, ends up holding together. That is true of synapses, on the scale of milliseconds. It is also true, on an entirely different scale and through other mechanisms, of your habits, and for those we have figures (§11). The image helps you understand why consistency matters.

What is measured: what is familiar seems truer

Another well-documented mechanism needs no change at all in the structure of the brain. Repetition makes a sentence more familiar, and what is familiar seems truer (Udry & Barber 2024). Researchers speak of a fluency effect: a sentence you have already come across is easier to read and understand, and we take that ease as a sign of truth.

This effect has a rhythm of its own: within a single session, the biggest gain comes from the second encounter with a sentence, and each repetition after that adds a little less (Hassan & Barber 2021). These studies were run in the laboratory, with sentences that state facts.

What we built into the app

Consistency rather than intensity. From Doidge, we keep the image of the sledge, from which we drew, in §1, the idea of a regular ritual, focused on what you have chosen. The Braille study as he tells it, and the juggling study, where the changes faded once practice stopped, point the same way: what lasts is built over time, and needs upkeep. Familiarity, meanwhile, grows most with the second encounter within a single session, then a little less with each repetition. Hence an appointment that comes round every evening, so you can come back to your sentences day after day.

Attention, when you talk things through, write and record. Doidge makes attention a condition for lasting change. In HypnoSelf, attention comes into play earlier on: when you talk things through in the conversation to name what holds you back, when you write your sentences, then when you record them in your voice. Evening listening has a different role: it happens every day, it makes your sentences familiar, and it fits into a moment of relaxation (§9).

References for this section

10.1016/j.cub.2011.11.018 · 10.1038/427311a · 10.1146/annurev.neuro.31.060407.125639 · 10.1016/j.copsyc.2023.101736 · 10.1186/s41235-021-00301-5

---

11. Habit: how long, really

The essentials

The “21 days” to form a habit are a myth. In the landmark study, half the participants needed more than two months, with huge variation from one person to the next, and missing a day hardly mattered. Three things help, and research has measured them: practising at the same time each day, keeping track of your progress, and feeling capable of getting there. A morning practice holds up better over time. We chose the evening, because Murphy sees falling asleep as the most favourable moment and bedtime comes round every day in the same context; the app leaves the choice to you.

You’ve probably read that it takes 21 days to form a habit. It usually takes longer, and the time varies a lot from one person to another, which is rather reassuring.

What the books say

For Hill and Stone, the key principle is a positive mental attitude, combined with immediate action. Repetition is the means they propose to anchor it and make it a habit. They draw on William James, for whom a repeated action becomes a habit, a bundle of habits makes a character, and character decides destiny. Hence their conclusion that you are in fact the product of your habits, and that your habits are yours to choose. They also suggest working on one principle a week, until the subconscious responds to it of its own accord.

Murphy describes the same path with simple examples. Swimming, dancing or driving are learned by consciously repeating a movement until it becomes automatic, second nature. He also argues that you break a habit by building a new one.

Doidge too, as we saw in §10, argues for slow, steady work.

These books agree on one point: repetition builds the habit. The app owes them its emphasis on consistency. That left the question of how long it takes, and research has measured it.

What is measured: a median of 66 days

The landmark study followed 96 volunteers for 12 weeks as they repeated, every day, a simple action from daily life: drinking a glass of water, eating a piece of fruit, going for a walk. Each of them rated, day after day, how far the action was becoming a reflex (Lally et al. 2010).

Automaticity rises quickly at first, then more and more slowly, until it reaches a plateau. For each person, the researchers calculated the day they reached 95% of that plateau. In shorthand, that day is called the time it takes for an action to become automatic.

That time is 66 days, and it is a median. The median splits a group in two: half the people took less than 66 days, the other half longer. Depending on the person, it took 18 to 254 days. These 66 days are calculated from 39 of the 96 people recruited, those whose progress closely followed a smooth curve. For the slowest, the calculated day falls after the end of the 12 weeks, and it is obtained by extending their curve.

A 2024 meta-analysis, covering 20 studies and 2,601 participants, does report this result as a median, with this wide range. In the studies that measure this time directly, the medians range from 59 to 66 days. For most health behaviours, the authors conclude, it takes two to five months (Singh et al. 2024).

A very large study points the same way. It analysed what gyms and hospitals were already recording, as the actions happened: the gym visits of more than 30,000 members, and the handwashing of more than 3,000 healthcare workers. For the gym, the median ranges from 122 to 226 days, or four to seven months. For handwashing, it takes about two weeks. The authors conclude that there is no magic number for forming a habit (Buyalskaya et al. 2023).

As for the “21 days”, they are a myth, according to two of the authors of the landmark study (Gardner, Lally & Wardle 2012). They trace the figure back to anecdotal observations, mere accounts of individual cases, in a 1960 book in which cosmetic surgery patients got used to their new appearance in about three weeks. They also note that a simple action, such as drinking a glass of water, becomes automatic faster than a more elaborate routine, such as doing fifty sit-ups.

Everyone at their own pace

Above all, these studies show how widely people differ, and that is good news. In the landmark study, the fastest took 18 days and the slowest 254, fourteen times as long. In another trial from the meta-analysis, with 192 people, those who reached the automaticity threshold set by the researchers took between 4 and 335 days to get there. If the act becomes automatic for you in three weeks, you’re within the range. If you’re three months in and it still takes effort, you’re within it too.

The same team also gathered, through interviews, the experience of ten people who were forming a new habit (Lally et al. 2011). Their accounts describe the same journey: at first, the effort is conscious and demanding, then the action gets easier as automaticity sets in. It is the path Murphy described with swimming and driving, and you may recognise it.

And what if you miss an evening? The landmark study looked at what a missed opportunity changes: skipping once barely affected how the habit formed (Lally et al. 2010). What makes people give up lies elsewhere. Three studies that followed the same people over time, on eating, exercise and saving, show that a small lapse makes another lapse more likely. The most consistent mechanism is a drop in how capable people feel (ten Broeke & Adriaanse 2023).

Tracking your progress

This is one of the most solid results on this page. A meta-analysis brought together 138 randomised trials (§2), with nearly 20,000 people. In each one, chance decided who was prompted to track their own progress, and who served as the comparison. Tracking your progress increases the chances of reaching the goal you set, with an effect size of 0.40, between small and medium (Harkin et al. 2016).

The effect is stronger under two conditions: when the information is physically recorded, rather than kept in your head, and when the results are reported to someone or made public.

Throughout this research, tracking concerns behaviours you can count: your weight, your steps, your cigarettes, your meals. It mainly works when it accompanies a practice. A review of studies in people who wanted to lose weight shows this clearly. In the only trial that tested weighing on its own, the effect remained uncertain; added to a more complete programme, weighing led to greater weight loss (Madigan et al. 2015).

The same time, every day

Researchers asked 95 students to take up a new study habit for six weeks. One group always practised it in the same place and at the same time; the other changed place and time on each occasion. The first group developed more automaticity, and stuck more closely to their repetition goal. The same pattern shows up in 218 users of a habit app (Stojanovic et al. 2022).

Another study gives an idea of what you can hope for. It distinguishes two stages in a habit: getting started (setting about the act) and carrying it out (going through it from start to finish). It is getting started that can become a reflex. The researchers followed 229 students for a week, on three behaviours (flossing, snacking, eating breakfast). For all three, it was the habit of getting started on the act, more than the habit of carrying it out, that best predicted what the students actually did (Gardner, Phillips & Judah 2016).

A morning practice holds up better

The 2024 meta-analysis concludes that practising in the morning seems to establish a habit better than practising in the evening. In the study it cites, on daily stretching, it took an average of 106 days for the action to become automatic in the morning, against 154 days in the evening (Singh et al. 2024).

Another study is closer to what you do. It followed 14,879 users of a meditation app for 180 days. Over the months, practice declines at every time of day, and it declines least in the morning. The higher the share of morning sessions, the better the engagement, and the longer people keep using the app. Late-evening sessions go hand in hand with the opposite (Berardi et al. 2023).

Feeling capable, and staying that way

One last result sheds light on the goal of the ritual: strengthening the feeling that you can act (§6). A meta-analysis brought together 204 randomised experiments on health behaviours such as physical activity, diet and screening. When they succeed in boosting this feeling, behaviour really does change relative to the comparison groups, with an effect size of 0.47, medium. On the intention to act, it is 0.51, also medium (Sheeran et al. 2016).

In physical activity programmes for adults, this feeling increases just after the intervention, relative to the trials’ comparison groups, with an effect size of 0.26, small. Measured later, the effect has dropped back to 0.08, almost nothing (Tang et al. 2019). So this gain fades once the programme is over: it needs to be kept up.

What we built into the app

• A 66-day benchmark. The 66-day badge marks the median from the landmark study: when you get there, you’ve practised for as long as it took half its participants for their action to become automatic. If your ritual still takes effort at that point, you’re still within the measured range. Read our milestones as markers along the way rather than as thresholds to cross.
• Your tracking, recorded and on display. Every listening session is logged in the app. Your current streak, your record and your calendar of the last thirty days appear on your progress screen. That is the first of the two conditions that strengthen the effect of tracking. And this tracking accompanies your ritual, as in the programmes where it worked best.
• A missed evening has little effect on the habit. The landmark study measured this. If your streak breaks, your record stays on display, and your calendar shows the evenings you were there. Pick it up again the next day, without judging yourself: after a lapse, people mostly give up when they feel less capable.
• The same time, every day. Bedtime comes round every evening, usually at the same time, in the same place, with the same sequence of steps, and research links this stable context with more automaticity. Our realistic goal is for starting your ritual to become natural, because the act of starting to listen is the one that can become a reflex.
• The evening first, and the morning if you prefer. We chose the evening for three reasons. The first two are in §8: Murphy sees falling asleep as the most favourable moment, and research describes the threshold of sleep as a special moment for thought. The third lies in bedtime itself, the most stable context of the day. What the morning has going for it is a practice that holds up better, and that is a real advantage. So the app offers it too. It presents waking up as the anchor that makes the habit stick, and you can schedule your reminders for the morning, during the day or in the evening.
• A ritual every day. Feeling capable is something you build up, and we are counting on this daily appointment to keep it going.

References for this section

10.1002/ejsp.674 · 10.3390/healthcare12232488 · 10.1073/pnas.2216115120 · 10.3399/bjgp12X659466 · 10.1080/13548506.2011.555774 · 10.1002/ejsp.2931 · 10.1037/bul0000025 · 10.1186/s12966-015-0267-4 · 10.3389/fpsyg.2022.883795 · 10.1111/bjhp.12189 · 10.2196/42482 · 10.1037/hea0000387 · 10.1093/abm/kay090

---

12. Attention: what you notice depends on what you look for

The essentials

You mostly notice what you are already looking for, and in our view, this mechanism is also at work in what you believe about yourself: we more readily spot what confirms what we think. In anxious adults, among several kinds of training that work on the way we see things, those that teach you to read a situation differently get the best results, with a modest effect against sham training. An affirmation ritual aims at the same target, the way you read what you experience, by a different path from these kinds of training.

You decide you’re interested in a particular car model, and suddenly you see them everywhere. There are just as many as before. What has changed is what you are looking for.

What the books say

Hill and Stone argue that sight itself is learned, that the brain does most of the work, and that we generally see what we want to see.

Doidge draws on the work of Bruce Wexler. He writes that in childhood, the brain builds images or representations that form the neural foundation of all our perceptions and all our beliefs. These representations are reinforced by repetition, and as we get older, we tend to dismiss information that contradicts our convictions.

Doidge also stresses attention itself: for the researcher Michael Merzenich, whose work he describes, lasting changes require sustained attention (§10).

Coué, for his part, asked for the opposite: saying his formula mechanically, the way you would recite a litany, without focusing on its meaning, and without the slightest effort.

On attention, then, our books pull in two directions, and we had to reconcile them.

What is measured: you see what you set

Laboratory research has put words to the phenomenon. Something catches your attention in spite of yourself only if it resembles what you are already looking for. Researchers say that this involuntary capture depends on your attentional control settings (Folk, Remington & Johnston 1992). Others made it the title of their paper, What you see is what you set, and showed the flip side: what falls outside your settings can stay invisible to you, even in plain sight (Most et al. 2005). This is called inattentional blindness: missing what is right in front of your eyes when your attention is elsewhere. Its founding demonstration dates from 1999 (Simons & Chabris 1999).

We think the same mechanism is at work in what you believe about yourself, and that is why it interests us: if you are convinced you are not up to it, you more readily notice what proves you right. A related tendency is well measured, and very widespread. A meta-analysis finds a preference for information that confirms what we already think, over information that contradicts it, with an effect size of 0.36, between small and medium (Hart et al. 2009). This preference varies with the situation. In particular, it weakens when the person has first felt validated in what they think.

What draws your attention also depends on what affects you. A meta-analysis of 243 studies, covering 9,120 people, measures how strongly positive items draw attention compared with neutral ones, with an effect size of 0.258, small. This pull is stronger when the item touches on one of the person’s own concerns than when it is positive in a generic way (Pool et al. 2015). That is an argument for sentences that sound like you rather than ready-made ones.

What is measured: a different reading of what you notice

A network meta-analysis brought together randomised trials in adults with symptoms of anxiety (65 trials, 3,897 participants) or depression (20 trials, 1,116 participants) (Fodor et al. 2020). This type of study compares several approaches with one another, including where no trial has put them head to head.

These trials test training methods that target our biases, those leanings of the mind that make us see things in a certain way. Two families interest us. Interpretation bias modification teaches you, through repetition, to read an ambiguous situation as less threatening. Attention bias modification teaches you to turn your attention away from threatening signals.

In the 65 trials on anxiety, interpretation bias modification gets the best results. Against a waiting list, meaning people waiting their turn to receive the training, symptoms fall more, with an effect size of 0.55, a medium effect. Against sham training, which looks like the real thing without its active ingredient, it is 0.30, between small and medium. Attention bias modification, for its part, only shows a benefit in an additional analysis that set aside the trials on post-traumatic stress.

An affirmation ritual aims at the same target as this interpretation training, the way you read what you experience, by a different path: sentences of your own, which you listen to every evening.

What we built into the app

• We aim at interpretation. It is the training that does best in the anxiety trials. Your affirmations offer you another reading of what you experience, written in your own words.
• Sentences that concern you. What relates to a concern of your own draws attention more than generic positivity. That is one more reason to write your sentences from what you said in the conversation (§4).
• What matters to you, first. The conversation always asks you what matters to you before a single sentence is written (§3). We see a link here with the finding above: when we have first felt validated, our preference for what proves us right is less marked.
• Attention for choosing, letting go for listening. Doidge asks for attention, Coué for letting go, and we drew on both. We put attention where you make choices: the conversation, writing your sentences, then reading them aloud when you record them. For listening, it is the tip from §5: say your sentences along with the recording for the first few evenings, then let yourself drift towards sleep. Doidge is talking about learning in general; we apply his idea to your sentences.

References for this section

10.1037/0096-1523.18.4.1030 · 10.1037/0033-295x.112.1.217 · 10.1068/p2952 · 10.1037/a0015701 · 10.1037/bul0000026 · 10.1016/S2215-0366(20)30130-9

---

13. Our commitment

The essentials

HypnoSelf is a personal development ritual that you practise every evening. Practices similar to the ones in the ritual have each been measured separately: in adults who slept poorly, listening to music every day improves sleep as they perceive it; knowingly taking a placebo helps people feel a little better; tracking your own progress helps you reach your goal. We suggest you try the ritual that brings them together. If you are going through a difficult time, talk to a doctor or a psychologist about it.

What HypnoSelf is

The sentences you listen to every evening are ones you chose yourself, then recorded in your own voice. You stay in control of the ritual: you choose what you listen to, and you can stop whenever you like. It is autosuggestion, a suggestion you give yourself, as Coué understood it: that’s what “self-hypnosis” means on this page (§2).

What you are aiming for is a relaxed state, close to the one that comes before sleep. So avoid listening while driving, or during any activity that needs your full attention: this relaxation can lower your alertness. And keep the volume low, since you are listening at bedtime.

What matters to us

If you are receiving care for a health condition, HypnoSelf complements that care. Before you start, ask your doctor’s advice if you are undergoing treatment, or if you have a condition that could be affected by deep relaxation techniques.

Our AI has a conversation with you to help you write your sentences. For any medical or psychological question, turn to a healthcare professional.

If you are going through a difficult time, one that weighs on you and lingers, talk to a doctor or a psychologist about it straight away.

What you can expect from it

Practices close to the ones in the ritual have each been measured separately. In adults who slept poorly, listening to recorded music every day improves sleep as they perceive it (§7). Knowingly taking a placebo helps people feel a little better (§2). Tracking your own progress helps you reach the goal you have set yourself (§11).

The ritual that brings them together is what we suggest you try. Going by this research, what you can expect from it is a bedtime you experience as a little calmer, sentences that become familiar to you (§10), and a practice that takes hold if you keep it up.

The ritual as a whole, the way you practise it here, has yet to be studied. Practice often comes first, and research looks into it afterwards: that was true of Coué, and of the hospital in Liège (§2).

---

14. All our references, described one by one

This section brings together every source on this page: first the books, with the edition we read, then the studies, grouped by theme in the order of the sections. Each study entry has two parts. What it shows: who took part, what the researchers had them do, or what they observed, and what they found. What we take from it for HypnoSelf: the part of the app, or the passage on this page, that it fed into.

A few words that come up often

The key pointers are explained in §2, where they are first used. You’ll find them again here, along with the other method terms you’ll come across on this page.

Randomised trial. A trial where chance decides who receives what. That’s what allows a difference to be put down to what is being tested, rather than to the profile of the people taking part.

Meta-analysis. A study that pools the trials carried out on the same question and draws an overall result from them. A systematic review searches for these trials using a method set in advance. A Cochrane review is a systematic review published by Cochrane, an independent network: these reviews are a standard reference in medicine, and are known for the rigour of their method. A network meta-analysis compares several approaches with one another, even when no trial has put them head to head. An umbrella review brings together meta-analyses.

Effect size. A number that tells you how big a difference is, whatever the number of participants. It is written d, g or SMD depending on how it is calculated, and all three are read the same way: around 0.2, it’s small; around 0.5, medium; around 0.8, large. To picture it, pick at random one person who practised and one from the comparison group: at 0.2, the first is doing better than the second about 56 times out of 100; at 0.5, 64 times; at 0.8, 71 times (§2). A minus sign most often indicates a decrease: less stress, less time to fall asleep.

Correlation (r). A number that tells you how closely two measures vary together. At 0, each one varies independently; the further you move from 0, the stronger the link. Most often, both measures are taken from the same people, in a single group. It is sometimes also used to express an effect size.

Median. The value that splits a group into two equal halves. If the median is 66 days for an action to become automatic, half the people took less than 66 days, and the other half took longer (§11).

95% confidence interval (95% CI). The range in which the true value very probably lies. The narrower it is, the more precise the result. If it includes zero, the effect may be zero.

Comparison group. The people against whom those who received the intervention are compared. People on a waiting list wait their turn to receive it. An active control group does a different activity. A placebo, or sham training, imitates the intervention without its active ingredient. Against a demanding comparison group, the measured effect is often smaller, and more convincing. For example, the same therapy scores 0.82 against a waiting list, and 0.51 against an activity that is credible but inactive (A-Tjak et al. 2015, group C).

Observational study. A study that observes people as they live their lives, whereas a trial randomly assigns who receives what. It describes links between two things; to know whether one causes the other, you need a trial. A case series reports on a set of patients treated in the same way, with no randomly assigned group to compare them with: it shows how that approach turned out for them.

The books

The seven books we talk about in §1, and the text by Émile Coué that we describe in §2, in the editions we read:

• Napoleon Hill and W. Clement Stone, Success Through a Positive Mental Attitude, Prentice-Hall, 1960. Edition we read: the French translation, Le succès par la pensée constructive, Le Jour, 1988.
• Joseph Murphy, The Power of Your Subconscious Mind, 1963. Edition we read: the French translation, La puissance de votre subconscient, new revised edition, Les Éditions de l’Homme, Montreal, 2020, with a preface by Arthur R. Pell.
• Bruce H. Lipton, The Biology of Belief. Edition we read: the French translation, Biologie des croyances, Ariane, 2017.
• Norman Doidge, The Brain That Changes Itself, 2007. Edition we read: the French translation by Éric Wessberge, Les étonnants pouvoirs de transformation du cerveau, Belfond, 2019 (first published by Belfond in 2008).
• Gregg Braden, The Spontaneous Healing of Belief, Hay House, 2008. Edition we read: the original English edition.
• Joe Dispenza, Breaking the Habit of Being Yourself, Hay House, 2012. Edition we read: the French translation by Louis Royer, Rompre avec soi-même, Ariane, 2013.
• Joe Dispenza, You Are the Placebo, 2014. Edition we read: the French translation, Le placebo, c’est vous!, Ariane, 2015.
• Émile Coué, Self Mastery Through Conscious Autosuggestion, a text already in circulation by 1920. Edition we read: the original French text, La Maîtrise de soi-même par l’autosuggestion consciente, in La Méthode Coué, Renaudot et Cie, 1989.

The studies

Next come the studies, grouped by theme, in the order of the sections. A few are placed with the theme they are closest to: Wood, Quinn and Kashy, cited in §1, and Bandura, cited in §6, appear under habit (group K).

A. Open-label placebo (§2)

An open-label placebo is a product with no active ingredient, given to someone who is told plainly what they are taking.

Fendel, J. C. et al. (2025). Effects of open-label placebos across populations and outcomes: an updated systematic review and meta-analysis of randomized controlled trials. Scientific Reports. 10.1038/s41598-025-14895-z

• What it shows. 63 randomised trials, 4,554 participants in all, patients or healthy people; in most of them, the comparison group received nothing. Across all measures, the effect size is 0.35, a modest effect, driven mainly by what people say they feel (0.39, across 55 comparisons); on objective measures, those that don’t rely on what the person reports, it is 0.09 (17 comparisons), almost nothing. It is 0.29 in healthy people, the audience HypnoSelf is designed for, and 0.47 in patients.
• What we take from it for HypnoSelf. More than any other, this result shaped our approach (§2), because it shows that a ritual can do you good even when you know how it works. It also lets us tell you plainly, on this page, where each act in your ritual comes from and what has been measured about it.

Buergler, S. et al. (2023). Imaginary pills and open-label placebos can reduce test anxiety by means of placebo mechanisms. Scientific Reports. 10.1038/s41598-023-29624-7

• What it shows. A randomised trial with 173 students who were anxious ahead of an exam: for three weeks, some knowingly took a pill with no active ingredient every day, others simply imagined taking it, and a third group, with no ritual at all, served as the comparison. Against that group, anxiety dropped in the first two groups, taken together, with an effect size of 0.71, between medium and large. The merely imagined pill did as well as the pill with no active ingredient that people actually swallowed.
• What we take from it for HypnoSelf. A study close to our format: an act repeated every day for three weeks, and anxiety that dropped just as much when the pill was merely imagined. In HypnoSelf, the daily appointment is your bedtime listening.

Schaefer, M. & Enge, S. (2024). Open-label placebos enhance test performance and reduce anxiety in learner drivers: a randomized controlled trial. Scientific Reports. 10.1038/s41598-024-56600-6

• What it shows. A randomised trial with 68 learner drivers. Those who took an open-label placebo were less anxious before the test, and fewer of them failed: 29%, compared with 53% in the group that took nothing.
• What we take from it for HypnoSelf. It finds the same principle at work before the driving test, one of those moments that matter, when the nerves kick in (§2).

B. Hypnosis: what it is, what it does (§2)

Meurisse, M. et al. (1999). Bilateral neck exploration under hypnosedation: a new standard of care in primary hyperparathyroidism? Annals of Surgery. 10.1097/00000658-199903000-00014

• What it shows. At the university hospital in Liège, patients with primary hyperparathyroidism, a disease of the parathyroid glands, needed neck surgery. Of 121 consecutive operations, 31 were done under hypnosedation, that is, with a local anaesthetic combined with hypnosis instead of a general anaesthetic, at the patients’ own request. It is a case series: the paper takes stock of these operations, their feasibility, their safety and their cost.
• What we take from it for HypnoSelf. The first half of the story told in §2: in Liège, the practice already existed in the operating theatre.

Defechereux, T. et al. (1999). Hypnoanesthesia for endocrine cervical surgery: a statement of practice. Journal of Alternative and Complementary Medicine. 10.1089/acm.1999.5.509

• What it shows. At the same university hospital, between 1994 and 1997, 218 operations were done under hypnosis, 197 on the thyroid and 21 on the parathyroids, and only 2 of them needed a switch to a general anaesthetic.
• What we take from it for HypnoSelf. In Liège, operating under hypnosis was already common practice before imaging came along to explain it (§2).

Faymonville, M.-E. et al. (2000). Neural mechanisms of antinociceptive effects of hypnosis. Anesthesiology, 92(5), 1257-1267. 10.1097/00000542-200005000-00013

• What it shows. An imaging study, in a small group of people, that tracks brain activity during hypnosis. It describes how hypnosis eases pain, notably by changing the activity of an area linked to how pain is felt (the anterior cingulate cortex).
• What we take from it for HypnoSelf. The second half of the story told in §2: the explanation came in 2000, after years of practice in the operating theatre.

Del Casale, A. et al. (2016). Hypnosis and pain perception: an Activation Likelihood Estimation (ALE) meta-analysis of functional neuroimaging studies. Journal of Physiology-Paris. 10.1016/j.jphysparis.2016.01.001

• What it shows. A meta-analysis of 8 imaging studies, 75 healthy volunteers in all, who felt pain induced in the laboratory during a hypnotic suggestion: research that seeks to understand the mechanism. When the suggestion aims to ease the pain, several areas become active, including one linked to how pain is felt (part of the cingulate cortex) and another that perceives the body’s sensations (the insula). Activity drops in a relay that passes sensations on to the rest of the brain (part of the thalamus).
• What we take from it for HypnoSelf. The word “hypnosis” is in our name. This meta-analysis helps us say what it covers in research: suggestions that change the activity of brain areas linked to pain (§2).

Jiang, H. et al. (2017). Brain Activity and Functional Connectivity Associated with Hypnosis. Cerebral Cortex. 10.1093/cercor/bhw220

• What it shows. An imaging study of 57 participants, selected from 545 healthy volunteers because they were either highly responsive or barely responsive to hypnosis, and observed under hypnosis, at rest, and while recalling a memory. Under hypnosis, activity drops in an area that monitors conflicts and errors (the dorsal anterior cingulate cortex), and communication increases between an area that helps you concentrate and direct your actions (the dorsolateral prefrontal cortex) and an area that perceives the body’s sensations (the insula). This result has yet to be replicated by another team.
• What we take from it for HypnoSelf. Along with Del Casale’s, it shows that hypnosis comes with changes in brain activity that can be seen on imaging (§2).

Rainville, P. et al. (1997). Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science. 10.1126/science.277.5328.968

• What it shows. An imaging experiment. Under hypnosis, suggestions changed how unpleasant a pain was, without changing how strong it was. Brain activity then changed in an area linked to the unpleasant side of pain (the anterior cingulate cortex), and stayed the same in the one that perceives its intensity (the somatosensory cortex).
• What we take from it for HypnoSelf. The landmark experiment in the field (§2): under hypnosis, a suggestion can have a precise, measurable effect, one that shows up even in brain activity.

De Pascalis, V. (2024). Brain Functional Correlates of Resting Hypnosis and Hypnotizability: A Review. Brain Sciences. 10.3390/brainsci14020115

• What it shows. The most recent review in the field. It brings together the studies on brain activity during hypnosis, and in people who are more and less responsive to hypnosis. It finds their results inconsistent from one study to the next, and calls for them to be replicated.
• What we take from it for HypnoSelf. §2 cites it when it talks about what the brain does under hypnosis: results vary from one imaging study to another.

Rosendahl, J., Alldredge, C. T. & Haddenhorst, A. (2024). Meta-analytic evidence on the efficacy of hypnosis for mental and somatic health issues: a 20-year perspective. Frontiers in Psychology. 10.3389/fpsyg.2023.1330238

• What it shows. An umbrella review, that is, a synthesis of meta-analyses: 49 meta-analyses and 261 studies on hypnosis in sessions with a practitioner, for mental or physical health problems. The strongest evidence concerns pain and supporting patients during medical procedures, and depending on the problem treated, the effects range from nil to very large.
• What we take from it for HypnoSelf. §2 relies on it to say what hypnosis offers in sessions with a practitioner, and to place that setting alongside your ritual, which you practise on your own.

Self-induced cognitive trance, three papers.

Gosseries, O. et al. (2020). Behavioural and brain responses in cognitive trance: a TMS-EEG case study. Clinical Neurophysiology. 10.1016/j.clinph.2019.11.011

Grégoire, C., Marie, N., Sombrun, C., Faymonville, M.-E. et al. (2022). Hypnosis, meditation, and self-induced cognitive trance to improve post-treatment oncological patients’ quality of life: study protocol. Frontiers in Psychology. 10.3389/fpsyg.2022.807741

Vanhaudenhuyse, A. et al. (2024). Phenomenology of auto-induced cognitive trance using text mining: a prospective and exploratory group study. Neuroscience of Consciousness. 10.1093/nc/niae036

• What they show. A recent field of research, led in Liège by the GIGA-Consciousness team, built around Corine Sombrun. It studies a trance you trigger yourself, which engages the body and comes from shamanic practices. The three papers are a study of a single person; the detailed plan of a trial with patients who have had cancer, a trial whose results have yet to be published; and an exploratory group study of what practitioners experience during this trance.
• What we take from it for HypnoSelf. Research on trances you trigger yourself is moving forward, and some of it is being done by a French-speaking team (§2).

C. The conversation, values, persuasion (§3)

Steele, C. M. (1988). The Psychology of Self-Affirmation: Sustaining the Integrity of the Self. Advances in Experimental Social Psychology, 261-302. 10.1016/S0065-2601(08)60229-4

• What it shows. The theoretical chapter that founded the field of self-affirmation. In it, Steele describes a mechanism: reflecting on your own values. The word brings to mind the sentences people repeat to themselves; for researchers, self-affirming means taking a moment to remind yourself what really matters to you.
• What we take from it for HypnoSelf. That meaning, reflecting on what matters to you, has its place in the conversation (§3). The evening repetition draws on other research (§10, §11).

Escobar-Soler, C. et al. (2023). Effectiveness of self-affirmation interventions in educational settings: a meta-analysis. Healthcare. 10.3390/healthcare12010003

• What it shows. 144 experimental studies, carried out with secondary school and university students. Some of them took a moment to reflect on what matters to them: that is the meaning research gives to self-affirmation (see Steele, just above). Against their comparison groups, the effect size on academic results or well-being is 0.41, between small and medium, and the authors themselves consider it weak.
• What we take from it for HypnoSelf. §3 gives this figure first, as an order of magnitude for what the reflection in the conversation can bring.

Epton, T. et al. (2015). The impact of self-affirmation on health-behavior change: a meta-analysis. Health Psychology. 10.1037/hea0000116

• What it shows. In 144 experimental tests, all participants read a message about a risk to their health, and some had first reflected on their values. Compared with those who had read the same message without this exercise, they accepted it a little more readily (effect size of 0.17), had slightly stronger intentions to change (0.14) and changed their behaviour a little more (0.32): three small effects, all in the same direction. In these tests, self-affirmation always accompanies a health message from outside, which it complements.
• What we take from it for HypnoSelf. These are the benchmark figures for self-affirmation, and §3 uses them as its yardstick when it talks about the reflection the conversation offers you.

Zhang, Y. et al. (2025). The impact of self-affirmation interventions on well-being: a meta-analysis. American Psychologist. 10.1037/amp0001591

• What it shows. 129 separate comparisons, drawn from 67 papers, in which people from the general population reflected on what matters to them: against the comparison groups, the effect size is 0.32 on self-image, 0.29 on general well-being, 0.26 on social well-being, and −0.22 on what the authors call psychological barriers, the minus sign meaning that these go down. These effects are small but real, the authors say, in the sense that chance alone cannot explain them. They last: for the drop in psychological barriers, the effect measured later even exceeds the effect measured straight away.
• What we take from it for HypnoSelf. It looks at people much like our users, the general population, and at their everyday well-being. §3 cites it to say that these effects last.

Creswell, J. D. et al. (2013). Self-affirmation improves problem-solving under stress. PLOS ONE. 10.1371/journal.pone.0062593

• What it shows. A randomised trial, with 73 participants analysed. Some began by reflecting on what mattered most to them, then everyone did a problem-solving task knowing they were being assessed. Those who had reflected on their values did better, and the benefit was concentrated among those who reported high chronic stress, that is, stress that has become long-standing.
• What we take from it for HypnoSelf. The benefit goes mainly to the most stressed people, which is rare in this field. Together with the other research on values, it supports the step where the conversation asks you what matters most to you (§3).

Falk, E. B. et al. (2015). Self-affirmation alters the brain’s response to health messages and subsequent behavior change. PNAS, 112(7), 1977-1982. 10.1073/pnas.1500247112

• What it shows. In sedentary people, reflecting on their values before receiving a health message increases activity in a brain area linked to thinking about oneself and to the value people place on things (the ventromedial prefrontal cortex). Afterwards, their physical activity increases too, measured objectively rather than self-reported.
• What we take from it for HypnoSelf. One of the few results that link a change in the brain to a real change in behaviour. In it, people reflect on their values before receiving a message, the same order as in the app, where the conversation comes before your affirmations (§3).

Cascio, C. N. et al. (2016). Self-affirmation activates brain systems associated with self-related processing and reward and is reinforced by future orientation. Social Cognitive and Affective Neuroscience, 11(4), 621-629. 10.1093/scan/nsv136

• What it shows. While their brains are being scanned, participants reflect on their core values, looking towards their future. Areas linked to thinking about oneself and to reward then become more active than when they think about their everyday activities; according to the authors, the fact that these values look towards the future strengthens this effect. The stronger this activity, the more their sedentary behaviour drops afterwards, following a physical activity programme.
• What we take from it for HypnoSelf. The idea that looking towards your future strengthens the effect helped shape the sentences that describe a path (“I’m learning to…”, “I’m becoming…”), which the conversation adds when a belief pushes back.

Baldwin, A. S. et al. (2013). Examining causal components and a mediating process underlying self-generated health arguments for exercise and smoking cessation. Health Psychology. 10.1037/a0029937

• What it shows. Two experiments in which participants were randomly assigned either to write their own arguments or to read someone else’s: the first on physical exercise, with 76 students, the second on quitting smoking, with 107 daily smokers. In the first, people rated the arguments they had written more favourably than those they had read, with an effect size of 0.63, a medium effect. The second shows what really makes the difference: the content matching your own concerns, more than having written it yourself.
• What we take from it for HypnoSelf. Your sentences start from what you have just said in the conversation, so that they match your concerns. In the second experiment, on quitting smoking, that is what mattered most.

Briñol, P., McCaslin, M. J. & Petty, R. E. (2012). Self-generated persuasion: effects of the target and direction of arguments. Journal of Personality and Social Psychology. 10.1037/a0027231

• What it shows. Four studies. When the message goes against what you already think, you persuade yourself more when you address it to yourself than to someone else. When it goes along with what you think, it is the other way round.
• What we take from it for HypnoSelf. Your affirmations are aimed precisely at beliefs you are trying to shift, like “I am capable” when you doubt it: that is the case where addressing yourself is most persuasive.

Smedslund, G. et al. (2011). Motivational interviewing for substance abuse. Cochrane Database of Systematic Reviews. 10.1002/14651858.CD008063.pub2

• What it shows. This Cochrane review looks at motivational interviewing, a way of guiding a discussion that helps the person put their own reasons for change into words: 59 studies and 13,342 participants with problematic use of substances such as alcohol or drugs. Compared with no treatment, the effect size is 0.79 right after the session, a large effect, then falls to 0.17 in the short term and 0.15 in the medium term, small effects, and to 0.06 in the long term, almost nothing, which chance could explain.
• What we take from it for HypnoSelf. What a session produces on its own fades over time. So in HypnoSelf, the conversation leads to a daily ritual, which takes over from it.

Magill, M. et al. (2017). A meta-analysis of motivational interviewing process: technical, relational, and conditional process models of change. Journal of Consulting and Clinical Psychology. 10.1037/ccp0000250

• What it shows. In a motivational interviewing session, a person can talk in favour of change, or in favour of the status quo, of what holds them back; this meta-analysis, covering 36 studies and 3,025 participants, looks at the share of each. The bigger the share directed towards change, the more the risky behaviour drops afterwards (r = −0.16, a weak link, found by observing sessions). When both kinds of talk are examined together, it is mainly status quo talk that remains linked to the outcome: the more room it takes up, the poorer the outcomes (r = 0.19, also a weak link).
• What we take from it for HypnoSelf. It led directly to a design choice: the conversation listens to what is blocking you, then brings you back to what you want to build.

Pace, B. T. et al. (2017). A multivariate meta-analysis of motivational interviewing process and outcome. Psychology of Addictive Behaviors. 10.1037/adb0000280

• What it shows. Across 19 studies and 2,614 participants, it points the same way: the more the person talks in favour of the status quo, the poorer the outcomes, whereas the amount of talk in favour of change shows no link with the outcome. Together with the previous one, it shows that what counts is the share of talk directed towards change, more than its amount.
• What we take from it for HypnoSelf. It backs up the same choice for the conversation, based on a different set of studies.

Levin, M. E. et al. (2012). The impact of treatment components suggested by the psychological flexibility model: a meta-analysis of laboratory-based component studies. Behavior Therapy. 10.1016/j.beth.2012.05.003

• What it shows. A meta-analysis of 66 laboratory studies, which test the ingredients of a single therapeutic approach one by one. The effects are larger when people actually experience the exercise, via a metaphor or hands-on practice, than when it is simply explained to them.
• What we take from it for HypnoSelf. In the conversation, you are the one who puts things into words, because an exercise you experience has more effect than an explanation.

A-Tjak, J. G. L. et al. (2015). A meta-analysis of the efficacy of acceptance and commitment therapy for clinically relevant mental and physical health problems. Psychotherapy and Psychosomatics. 10.1159/000365764

• What it shows. 39 randomised trials, 1,821 patients, on acceptance and commitment therapy, a form of psychotherapy. Against a waiting list, the effect size is 0.82, a large effect. Against a psychological placebo, an activity that is credible but inactive, it is 0.51, a medium effect.
• What we take from it for HypnoSelf. It is the example used in the box “A few words that come up often”, at the top of this section: the same therapy, two figures, depending on the comparison group.

Beck, A. T. & Haigh, E. A. P. (2014). Advances in cognitive theory and therapy: the generic cognitive model. Annual Review of Clinical Psychology. 10.1146/annurev-clinpsy-032813-153734

• What it shows. A review that sets out the current formulation of the cognitive model: our automatic interpretations of a situation influence what we feel and what we do.
• What we take from it for HypnoSelf. The research framework that overlaps with the first step of the approach Dispenza proposes, spotting your automatic patterns. We used it to structure the conversation, which helps you spot what holds you back.

D. Personalisation (§4)

Nye, A., Delgadillo, J. & Barkham, M. (2023). Efficacy of personalized psychological interventions: a systematic review and meta-analysis. Journal of Consulting and Clinical Psychology, 91(7), 389-397. 10.1037/ccp0000820

• What it shows. The review brings together 17 randomised trials and 7,617 people being treated for a disorder, who received structured psychological interventions delivered by trained therapists; 9 of these trials, with 5,134 participants, could be included in the overall calculation. Against a standardised programme, the same for everyone, the intervention tailored to the person comes out ahead with an effect size of 0.22 (95% CI: 0.05 to 0.39), an advantage the authors describe as “seemingly small”.
• What we take from it for HypnoSelf. §4 uses this figure to explain why we write your sentences with you. This small advantage was obtained against another active intervention: the comparison is demanding, and that is what gives it weight.

Li, W. et al. (2024). The efficacy of personalized psychological interventions in adolescents: a scoping review and meta-analysis. Frontiers in Psychology. 10.3389/fpsyg.2024.1470817

• What it shows. Independent researchers asked the same question about adolescents. Across 7 studies and 1,347 participants, interventions tailored to the person do better than standardised interventions, with an effect size of 0.21: a small effect.
• What we take from it for HypnoSelf. It lets §4 say that this small advantage holds across research teams and across populations.

Noar, S. M., Benac, C. N. & Harris, M. S. (2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133(4). 10.1037/0033-2909.133.4.673

• What it shows. 57 studies, 58,454 participants. Printed health-promotion material, tailored to the person, does better than a generic message, the same for everyone. The advantage is expressed here as a correlation, r = 0.074; converted into the unit used elsewhere on this page, it comes to about 0.15: a small effect.
• What we take from it for HypnoSelf. A third figure, on a completely different medium, pointing the same way as the previous two. Together, they give the general idea that §4 takes up: what is tailored to the person does a little better than what is the same for everyone.

Jabs, M. et al. (2026). Assessing the effectiveness, adherence and attrition of smartphone-based mental health interventions in university students: a systematic review and meta-analysis. Early Intervention in Psychiatry. 10.1111/eip.70207

• What it shows. A synthesis of 122 studies on mental health apps used by students. Personalisation stands out as a factor that strengthens the effect of these apps on depression.
• What we take from it for HypnoSelf. It is the only one of the four syntheses in this group that looks at apps. It supports the time we spend, in the conversation, writing your sentences with you (§4).

E. Your voice and your recordings (§5)

Fawcett, J. M. (2013). The production effect benefits performance in between-subject designs: a meta-analysis. Acta Psychologica. 10.1016/j.actpsy.2012.10.001

• What it shows. The production effect is the finding that what you say out loud is remembered better than what you read silently. This meta-analysis brings together experiments where one group reads everything out loud and another reads everything silently (what is known as a between-subjects design). For recognising a word you have already studied among new words, the effect size is 0.37 against the group that reads silently: an advantage between small and medium.
• What we take from it for HypnoSelf. This is the figure that matches your situation when you record yourself: you say all your sentences out loud, like the “out loud” group in these experiments. It supports the choice of having you say them rather than read them silently.

MacLeod, C. M. et al. (2010). The production effect: delineation of a phenomenon. Journal of Experimental Psychology: Learning, Memory, and Cognition. 10.1037/a0018785

• What it shows. Eight experiments, carried out with students on word lists. The production effect appears in lists that mix words said out loud and words read silently. Its most widely accepted explanation comes down to one word, distinctiveness: what you have said out loud stands out from the rest in memory, and standing out helps you find it again.
• What we take from it for HypnoSelf. This is the explanation of the effect that §5 takes up. For your situation, where everything is said out loud, the figure that counts is Fawcett’s (2013).

Fawcett, J. M. et al. (2022). Production improves recognition and reduces intrusions in between-subject designs: an updated meta-analysis. Canadian Journal of Experimental Psychology. 10.1037/cep0000302

• What it shows. An update of the Fawcett (2013) meta-analysis. The advantage shows up again in recognition, and saying the words out loud also reduces intrusions, those words you wrongly add when you recall a list. For recall, when you have to retrieve what you studied from memory, with nothing to prompt you, it finds no equivalent advantage.
• What we take from it for HypnoSelf. In practice, when you listen again to a sentence you said out loud, you recognise it more easily than if you had only read it (§5).

Ozubko, J. D., Hourihan, K. L. & MacLeod, C. M. (2012). Production benefits learning: the production effect endures and improves memory for text. Memory. 10.1080/09658211.2012.699070

• What it shows. In three experiments, the advantage of reading aloud holds for up to a week, with word pairs, sentences and texts.
• What we take from it for HypnoSelf. The advantage still held after a week, and you hear your sentences again every evening.

Jones, A. C. & Pyc, M. A. (2013). The production effect: costs and benefits in free recall. Journal of Experimental Psychology: Learning, Memory, and Cognition. 10.1037/a0033337

• What it shows. In free recall, you write down from memory everything you remember. There, the effect appears only when the same person says some of the words and reads the rest silently, and it comes mainly from the words read silently being remembered less well. When one group says everything out loud and another reads everything silently, the two do equally well.
• What we take from it for HypnoSelf. The gain in free recall came mainly from the contrast between words spoken and words read. In your ritual, where everything goes through your voice, the advantage is therefore measured in recognition (§5).

Bertsch, S. et al. (2007). The generation effect: a meta-analytic review. Memory & Cognition. 10.3758/bf03193441

• What it shows. The generation effect is the finding that a word you come up with yourself from a clue is remembered better than a word you simply read. Across 86 studies and 445 effect sizes, it is 0.40 against simple reading: an effect between small and medium.
• What we take from it for HypnoSelf. This related effect points the same way: what you produce yourself is remembered better than what you receive. §5 draws a parallel with your sentences, which grow out of what you put into words in the conversation.

Wakeham-Lewis, R. M., Ozubko, J. & Fawcett, J. M. (2022). Characterizing production: the production effect is eliminated for unusual voices unless they are frequent at study. Memory. 10.1080/09658211.2022.2115075

• What it shows. In a series of experiments, participants read words out loud, either in their usual voice or imitating a character’s voice. In their usual voice, words said out loud were remembered better than words read silently. In the character voice, this advantage disappeared, and it came back only when participants read more words in that voice than in their own.
• What we take from it for HypnoSelf. You record your sentences in your everyday voice, the one in which the advantage of speaking out loud appeared in these experiments.

Forrin, N. D. & MacLeod, C. M. (2017). This time it’s personal: the memory benefit of hearing oneself. Memory. 10.1080/09658211.2017.1383434

• What it shows. The only study that separates the act of speaking from the identity of the voice. Run with students, on word lists, it adds a condition very close to what the app does: listening back to a recording of your own voice, made a little earlier. The result is a ranking: you remember best what you said out loud, then what you heard yourself say, then what you heard another person read.
• What we take from it for HypnoSelf. The advice in §5 rests on this ranking: for the first few evenings, say your sentences out loud along with the recording, then let go and simply listen.

Symons, C. S. & Johnson, B. T. (1997). The self-reference effect in memory: a meta-analysis. Psychological Bulletin. 10.1037/0033-2909.121.3.371

• What it shows. A meta-analysis of the self-reference effect: what you relate to yourself (does this word describe me?) is remembered better than what you process for its meaning alone, or relate to someone else. The advantage is smaller when the comparison is with thinking about someone else than when it is with processing meaning alone.
• What we take from it for HypnoSelf. The argument that sentences like yours, which are about you, are remembered better than general ones. §5 adds it to the advantage of your voice.

Kaplan, J. T., Aziz-Zadeh, L., Uddin, L. Q. & Iacoboni, M. (2008). The self across the senses: an fMRI study of self-face and self-voice recognition. Social Cognitive and Affective Neuroscience. 10.1093/scan/nsn014

• What it shows. The seminal study on recognising your own voice: in a scanner that tracks brain activity, adults heard their own voice and a friend’s. Recognising their own voice engaged an area towards the front of the brain, on the right side (the right inferior frontal gyrus). In the authors’ view, this area recognises what comes from the self, whether your face or your voice.
• What we take from it for HypnoSelf. §5 concludes from it that your brain handles your voice separately, even compared with a friend’s.

Sun, Y. et al. (2023). Brain mechanism of unfamiliar and familiar voice processing: an activation likelihood estimation meta-analysis. PeerJ. 10.7717/peerj.14976

• What it shows. The standard imaging meta-analysis on familiar voices, covering 32 studies. Recognising a familiar voice engages a wide network (much of the sides of the brain, where sounds come in, regions towards the front, and deeper structures), whereas an unfamiliar voice goes through a narrower network.
• What we take from it for HypnoSelf. What is known about how familiar voices are processed in general. It is the backdrop against which §5 places your own voice, together with the two studies that measure it directly (Kaplan, Graux).

Graux, J. et al. (2014). Is my voice just a familiar voice? An electrophysiological study. Social Cognitive and Affective Neuroscience. 10.1093/scan/nsu031

• What it shows. 15 adults heard a single vowel, spoken either in their own voice or in the voice of someone they knew well, while the electrical activity of their brain was recorded. One of their brain’s responses was weaker for their own voice. In the authors’ view, telling your own voice apart takes less of the automatic processing that comes before attention.
• What we take from it for HypnoSelf. Your brain picks out your own voice with less work than the voice of someone close to you, even before you pay attention to it. §5 uses this to pin down how your voice is handled separately.

Orepić, P. et al. (2023). Bone conduction facilitates self-other voice discrimination. Royal Society Open Science. 10.1098/rsos.221561

• What it shows. When you speak, you also hear your voice through the bones of your skull: this is bone conduction, and that part is missing from any recording. In three studies, adults had to tell their voice apart from someone else’s, and they did better when a vibration transmitted through the bones was added. In the authors’ view, it is the absence of this part that makes your voice sound strange when you listen back to it.
• What we take from it for HypnoSelf. If your voice surprises you on your first recording, it is because it lacks the part you hear through your bones, as §5 explains.

Kimura, M. & Yotsumoto, Y. (2018). Auditory traits of “own voice”. PLoS ONE. 10.1371/journal.pone.0199443

• What it shows. In three experiments, adults listened to their recorded voice, passed through different filters, and judged which one made it sound most like their own. The answer varies a great deal from one person to another. The authors compare the discomfort of hearing yourself played back to the “uncanny valley”, the unease we feel in front of a robot or a doll that is almost human: it is almost your voice, but not quite.
• What we take from it for HypnoSelf. If your recorded voice bothers you at first, bear in mind that this discomfort is common, and that researchers offer an explanation for it (§5).

Nilsson, J.-E. & Lundh, L.-G. (2016). Audio feedback with reduced self-focus as an intervention for social anxiety: an experimental study. Cognitive Behaviour Therapy. 10.1080/16506073.2015.1126633

• What it shows. A randomised experiment, in a single session, with 41 socially anxious people, that is, people who dread the way others see them. They gave a speech, then listened to the recording of it. Those who had been asked to listen “as though they were listening to a stranger” then judged their voice more favourably than those who simply listened, and the effect still held at the next speech.
• What we take from it for HypnoSelf. §5 suggests you do as these participants did: if your voice bothers you at first, listen to it as if it were a stranger’s. This tip costs nothing to try.

F. How you phrase your sentences (§6)

Kross, E. et al. (2014). Self-talk as a regulatory mechanism: How you do it matters. Journal of Personality and Social Psychology. 10.1037/a0035173

• What it shows. Seven studies, 585 participants in total, in situations of social stress: speaking in public, making a good impression at a first meeting. Those who used distanced self-talk, addressing themselves by their first name or as “you”, were judged to perform better by outside evaluators than those who talked to themselves as “I”. They felt less distress and ruminated less afterwards.
• What we take from it for HypnoSelf. Talking to yourself as “you” or by your first name helps you step back in a stressful moment. So some of your sentences use “you” and your first name (§6).

Orvell, A. et al. (2020). Does distanced self-talk facilitate emotion regulation across a range of emotionally intense experiences? Clinical Psychological Science. 10.1177/2167702620951539

• What it shows. Two experiments, run with enough participants to detect a modest effect, in which people reflected on a difficult situation, past or still to come. The benefit of distancing held whether the situation was mildly or highly emotional, including in more vulnerable people. The authors point out that its use in therapy remains to be examined.
• What we take from it for HypnoSelf. In §6, it extends Kross: the step back that distanced self-talk gives you has also been measured in highly emotional situations, beyond minor annoyances.

Moser, J. S. et al. (2017). Third-person self-talk facilitates emotion regulation without engaging cognitive control: Converging evidence from ERP and fMRI. Scientific Reports. 10.1038/s41598-017-04047-3

• What it shows. Two experiments, with distressing images and painful memories: one measures the brain’s electrical activity, the other observes the brain in a scanner. Talking to yourself by your first name, rather than as “I”, reduces activity in an area towards the front and middle of the brain that is at work when we think about ourselves (the medial prefrontal cortex). An electrical signal that accompanies the emotional reaction also drops, and the signs of effort to control oneself stay at the same level.
• What we take from it for HypnoSelf. It gives §6 its description of what happens during distanced self-talk: the brain regulates its emotions with less effort.

Webster, C. T. et al. (2022). An event-related potential investigation of distanced self-talk: replication and comparison to detached reappraisal. International Journal of Psychophysiology. 10.1016/j.ijpsycho.2022.05.003

• What it shows. The part of the previous study that measured the brain’s electrical activity was repeated exactly for the first time, with 67 participants. The effect of distancing held.
• What we take from it for HypnoSelf. Moser’s result held when the experiment was repeated exactly, which gives it more weight than an isolated result. §6 relies on it to talk about the step back that comes from talking to yourself by your first name.

Moran, T. & Eyal, T. (2022). Emotion regulation by psychological distance and level of abstraction: two meta-analyses. Personality and Social Psychology Review. 10.1177/10888683211069025

• What it shows. Across 230 study results, psychological distancing reduces felt emotion with an effect size of 0.52, against conditions without distancing: a medium effect. For self-directed emotions, such as shame, the effect drops to 0.29: a small effect.
• What we take from it for HypnoSelf. This is the broadest meta-analysis in the field, on psychological distancing in general: §6 gives its effect, medium on felt emotion, smaller on self-directed emotions.

Dolcos, S. & Albarracín, D. (2014). The inner speech of behavioral regulation: Intentions and task performance strengthen when you talk to yourself as a You. European Journal of Social Psychology, 44(6), 636-642. 10.1002/ejsp.2048

• What it shows. In three laboratory experiments, students who gave themselves advice as “you” rather than as “I” did better at solving anagrams, puzzles where you unscramble the letters of a word. “You” also strengthened the intention to solve more of them and, in another experiment, the intention to exercise. The gain came through their attitude towards the task, which became more favourable.
• What we take from it for HypnoSelf. This is the study that launched research on the “you” form. Since then, other studies have compared “you” and “I” directly, such as Magata and Tanaka’s, with mixed results: your sentences combine both forms (§6).

Magata, Y. & Tanaka, A. (2024). The effect of second-person self-talk on performance and motivation in Japanese individuals. PLOS ONE. 10.1371/journal.pone.0305251

• What it shows. The largest published test of “you” against “I”, with 411 students, which also included a sentence with no subject and a control group. The study was run in Japanese, a language in which the subject is usually left out. The four groups did equally well, on performance and on motivation alike.
• What we take from it for HypnoSelf. The largest test of “you” against “I” puts them on an equal footing. That is one of the reasons your sentences mix several forms (§6).

Wood, J. V., Perunovic, W. Q. E. & Lee, J. W. (2009). Positive self-statements: power for some, peril for others. Psychological Science. 10.1111/j.1467-9280.2009.02370.x

• What it shows. In two experiments, students with low self-esteem who repeated a ready-made positive sentence, “I’m a lovable person”, came out of it in a worse mood than those who repeated nothing. Among these students, focusing on what made the sentence true also left them in a worse mood than reflecting on what made it both true and false. Among those with high self-esteem, repetition improved mood, but only to a limited extent. The sentence was the same for everyone, imposed, and repeated during a single session.
• What we take from it for HypnoSelf. It influenced how the conversation writes your sentences with you: when a belief pushes back, it adds sentences that describe a path, such as “I’m making progress”, and it keeps the tone of self-compassion.

Kirby, J. N., Tellegen, C. L. & Steindl, S. R. (2017). A meta-analysis of compassion-based interventions: current state of knowledge and future directions. Behavior Therapy. 10.1016/j.beth.2017.06.003

• What it shows. 21 randomised trials, 1,285 participants, with programmes led by practitioners, often in groups. Against comparison groups, self-compassion increases with an effect size of 0.70, between medium and large; depression drops by 0.64, anxiety by 0.49, and well-being increases by 0.51: medium effects.
• What we take from it for HypnoSelf. The framework we chose. Self-compassion, talking to yourself with the kindness you would show someone you love, has stronger support than positive affirmation, and it is the tone the conversation gives your sentences.

Wakelin, K. E., Perman, G. & Simonds, L. M. (2022). Effectiveness of self-compassion-related interventions for reducing self-criticism: a systematic review and meta-analysis. Clinical Psychology & Psychotherapy. 10.1002/cpp.2586

• What it shows. 19 articles, 1,350 participants. Against comparison groups, self-criticism drops with an effect size of 0.51 (95% CI: 0.33 to 0.69), a medium effect. Long interventions do better than short ones, and the effect is larger against groups doing nothing than against groups doing another activity.
• What we take from it for HypnoSelf. The critical inner voice, the one HypnoSelf targets, receded in these programmes. Duration matters too, which supports a ritual kept up over weeks, like our two-month benchmark (§11).

Lindsay, E. K. & Creswell, J. D. (2014). Helping the self help others: self-affirmation increases self-compassion and pro-social behaviors. Frontiers in Psychology. 10.3389/fpsyg.2014.00421

• What it shows. Two experiments, with 52 and then 77 participants. Some wrote for three minutes about a value that mattered to them, the others about a value that was unimportant to them. The former then showed more kindness towards themselves, and it was this kindness that led them to do more to help a person in difficulty, during a staged incident.
• What we take from it for HypnoSelf. The link between the conversation about your values (§3) and self-compassion (§6).

Cross, A. & Sheffield, D. (2019). Mental contrasting for health behaviour change: a systematic review and meta-analysis of effects and moderator variables. Health Psychology Review. 10.1080/17437199.2019.1594332

• What it shows. Mental contrasting means picturing your goal, then the real obstacle: seeing yourself running three times a week, then thinking about how tired you are in the evening. This meta-analysis covers changes in health behaviour, in 1,528 participants. Against comparison groups, its effect is 0.28 at four weeks (a value adjusted by the authors), a small effect, then 0.38 at three months in the 5 studies that followed participants that long: an effect between small and medium.
• What we take from it for HypnoSelf. Along with Wood’s study, mental contrasting played a part in the sentences that describe a path, which the conversation adds when a belief pushes back, such as “I’m making progress, even if I’m not there yet” (§6).

G. The act: listening to a recording in the evening (§7)

Jespersen, K. V. et al. (2022). Listening to music for insomnia in adults. Cochrane Database of Systematic Reviews. 10.1002/14651858.CD010459.pub3

• What it shows. This Cochrane review brings together 13 randomised trials and 1,007 adults who slept poorly: some listened to recorded music every day, for 25 to 60 minutes, over a period of three days to three months; the others received no treatment, or their usual care. In the music groups, sleep as participants perceive it improves: their score on the PSQI, a questionnaire where a lower score means better sleep, is 2.79 points lower than the others’ (10 studies, 708 participants), a large effect, whose certainty is rated moderate. Their difficulty falling asleep, as they report it, also goes down, with low certainty (3 studies, 197 participants).
• What we take from it for HypnoSelf. This is the evidence that comes closest to your evening listening: recorded music, heard every day, improves sleep as people perceive it, and no adverse effects were reported. Your ritual borrows this act, with your sentences in place of the music (§7).

Feng, F. et al. (2017). Can music improve sleep quality in adults with primary insomnia? A systematic review and network meta-analysis. International Journal of Nursing Studies. 10.1016/j.ijnurstu.2017.10.011

• What it shows. This network meta-analysis brings together 20 trials and 1,339 patients with insomnia, and ranks twelve approaches by combining direct and indirect comparisons. Listening to music comes out on top. Against usual care, it brings down the poor-sleep score on the PSQI, with an effect size of −0.61, a medium effect, obtained mainly through indirect comparisons.
• What we take from it for HypnoSelf. Another team, with a different method, finds the same perceived benefit from listening to music. In §7, it backs up the Cochrane review.

Toussaint, L. et al. (2021). Effectiveness of Progressive Muscle Relaxation, Deep Breathing, and Guided Imagery in Promoting Psychological and Physiological States of Relaxation. Evidence-Based Complementary and Alternative Medicine. 10.1155/2021/5924040

• What it shows. In this randomised laboratory trial, 60 healthy students were split into four groups of 15: three listened just once to twenty minutes of recorded instructions (muscle relaxation, deep breathing or guided imagery), and the fourth served as the comparison. In the first three groups, the relaxation they feel rises more than in the comparison group, and it also shows in the body, especially with muscle relaxation and guided imagery.
• What we take from it for HypnoSelf. This trial measures it directly: a recording, listened to with no practitioner beside you, can be enough to relax you. That is the format of your evening listening.

van Straten, A. et al. (2017). Cognitive and behavioral therapies in the treatment of insomnia: a meta-analysis. Sleep Medicine Reviews. 10.1016/j.smrv.2017.02.001

• What it shows. Cognitive behavioural therapy for insomnia is a structured therapy, carried out with a professional, which works on the habits and thoughts linked to sleep. Against the comparison groups in the trials, it reaches an effect size of 0.98 on the Insomnia Severity Index, a large effect, and of 0.57 on the time it takes to fall asleep, a medium effect.
• What we take from it for HypnoSelf. This is the dedicated treatment for insomnia: if your sleep worries you, §7 advises you to turn to this therapy, while your ritual keeps you company at bedtime.

H. Falling asleep, the hypnagogic window, sleep (§8)

The hypnagogic window is the moment when you tip over from wakefulness into sleep.

Lacaux, C. et al. (2021). Sleep onset is a creative sweet spot. Science Advances. 10.1126/sciadv.abj5866

• What it shows. In the laboratory, 103 healthy participants, chosen because they fell asleep easily, were deprived of about 30% of their sleep the night before, then worked on arithmetic problems that could be solved much faster using a hidden rule. Between two sets of problems, each person took a 20-minute break, semi-reclined, eyes closed, holding an object in one hand. Those who had spent a little over a minute on average in the very first stage of sleep, stage N1, then found the rule far more often than those who stayed awake: 83% (20 out of 24) against 31% (15 out of 49), a gap that chance can hardly account for. Those who had gone further, into stage N2, rarely found it: 14% (2 out of 14).
• What we take from it for HypnoSelf. This is what §8 rests on: those who had spent a minute on the threshold of sleep then found the hidden rule in an arithmetic problem more often, and your evening listening keeps you company on the way to this moment.

Horowitz, A. H. et al. (2023). Targeted dream incubation at sleep onset increases post-sleep creative performance. Scientific Reports. 10.1038/s41598-023-31361-w

• What it shows. In a small group of participants, a theme heard during stage N1 resurfaces in the thoughts that come as they fall asleep. After this spell in N1, measured creativity is higher than after a period awake, and the more often the theme came back, the higher it is. The answers are also further removed in meaning from the most obvious associations.
• What we take from it for HypnoSelf. Its principle is close to your ritual’s: content heard while falling asleep. What is measured afterwards is creativity (§8).

Andrillon, T. & Kouider, S. (2016). Implicit memory for words heard during sleep. Neuroscience of Consciousness. 10.1093/nc/niw014

• What it shows. In 22 participants, during the transition into sleep, the brain goes on correctly sorting what it hears (real words on one side, made-up words on the other), and it keeps an implicit trace of familiarity: an unconscious trace, but a measurable one. On waking, these words are forgotten, and the sleepers confuse them with new words.
• What we take from it for HypnoSelf. It pins down what §8 means by “the brain still hears”: as you fall asleep, it sorts the sounds it receives, and the words heard at that moment are forgotten on waking.

Jourde, H. R. & Coffey, E. B. J. (2024). Auditory processing up to cortex is maintained during sleep spindles. PNAS Nexus. 10.1093/pnasnexus/pgae479

• What it shows. During sleep spindles, those brief bursts of activity typical of light sleep, the brain goes on processing sounds all the way up to the cortex, its outer layer. That is the first step in listening; understanding what you hear is another.
• What we take from it for HypnoSelf. It confirms, by measuring brain activity, that the ear stays open during light sleep (§8).

Makov, S. et al. (2017). Sleep Disrupts High-Level Speech Parsing Despite Significant Basic Auditory Processing. Journal of Neuroscience. 10.1523/JNEUROSCI.0168-17.2017

• What it shows. Whether you are awake or asleep, the brain tracks the sounds of speech in a comparable way. As for words, phrases and sentences, it tracks them when you are awake and the speech is comprehensible: sleep disrupts this tracking.
• What we take from it for HypnoSelf. That is why your ritual comes just before sleep (§8): asleep, your brain still tracks sounds, but it is when awake that it understands a sentence, and an affirmation is a sentence.

Berardi, V. et al. (2023). Time of Day Preferences and Daily Temporal Consistency for Predicting the Sustained Use of a Commercial Meditation App: Longitudinal Observational Study. Journal of Medical Internet Research. 10.2196/42482

• What it shows. Among 14,879 users of a meditation app, followed for 180 days (899,071 app uses), the probability of practising drops over the months at every time of day, and it drops least in the morning: by 63.4%, against 67.8% to 74.5% at other times. A larger share of morning sessions goes hand in hand with practice that is more regular and lasts longer; late-evening sessions, with the opposite.
• What we take from it for HypnoSelf. So the app also offers you a listening session when you wake up, and §8 advises you to try the morning if you find it hard to listen regularly in the evening.

I. Brain rhythms (§9)

Kane, N. et al. (2017). A revised glossary of terms most commonly used by clinical electroencephalographers and updated proposal for the report format of the EEG findings. Revision 2017. Clinical Neurophysiology Practice. 10.1016/j.cnp.2017.07.002

• What it shows. This glossary is the International Federation of Clinical Neurophysiology’s standard reference for electroencephalography (EEG), which records the electrical activity of the brain with electrodes placed on the scalp. It sets the boundaries of the rhythms in hertz, that is, in cycles per second: delta from 0.1 to less than 4 Hz, theta from 4 to less than 8 Hz, alpha from 8 to 13 Hz inclusive, beta from 14 to 30 Hz inclusive.
• What we take from it for HypnoSelf. The shared reference points that §9 gives for each rhythm come from this glossary. Dispenza’s boundaries are very close to them.

J. Plasticity and Hebb (§10)

Woollett, K. & Maguire, E. A. (2011). Acquiring “the Knowledge” of London’s layout drives structural brain changes. Current Biology. 10.1016/j.cub.2011.11.018

• What it shows. The brains of 79 trainee London taxi drivers were scanned at the start of their training, then three to four years later. Initially, those who would go on to qualify, those who would go on to fail, and a comparison group of people not taking this training looked alike on these measures. At the end, in those who had obtained their licence, and in them alone, grey matter (the tissue rich in neurons) had increased at the back of the hippocampus, a structure of the brain used for memory and for finding your way around.
• What we take from it for HypnoSelf. In §10, it is the example of intensive learning that reshapes the adult brain, and the measure of what that takes: those who qualified reported around 34 hours of study a week, for three to four years.

Draganski, B. et al. (2004). Changes in grey matter induced by training. Nature. 10.1038/427311a

• What it shows. Of 24 adults, 12 learn to juggle for three months, and 12, with no training, serve as the comparison group. In the jugglers, grey matter increases; three months after they stop training, it has partly gone back towards its starting level.
• What we take from it for HypnoSelf. What the brain builds in three months of juggling needs upkeep: it is one of the arguments in §10 for a regular appointment, every evening, rather than for an intensive burst.

Caporale, N. & Dan, Y. (2008). Spike timing-dependent plasticity: a Hebbian learning rule. Annual Review of Neuroscience. 10.1146/annurev.neuro.31.060407.125639

• What it shows. The form of Hebb’s principle that experiments have confirmed. A synapse, the point of contact between two neurons, changes according to the order in which the two neurons send out their electrical signal (this is called “firing”), within a window of a few tens of milliseconds.
• What we take from it for HypnoSelf. The real timescale of the mechanism: the thousandths of a second between the signals of two neurons. The box in §10 uses it to say that the popular image holds true at the scale of synapses, and that your habits play out at another scale, through other mechanisms.

Udry, J. & Barber, S. J. (2024). The illusory truth effect: A review of how repetition increases belief in misinformation. Current Opinion in Psychology. 10.1016/j.copsyc.2023.101736

• What it shows. A review: repetition increases belief in a sentence through a fluency effect. What is familiar is processed more easily, and that ease is taken as a sign of truth. This research, run in the laboratory, looks at sentences that state facts, and at misinformation.
• What we take from it for HypnoSelf. This well-established mechanism could explain part of what you will feel as you listen to your sentences every evening: the more you hear them, the more familiar they become (§10).

Hassan, A. & Barber, S. J. (2021). The effects of repetition frequency on the illusory truth effect. Cognitive Research: Principles and Implications. 10.1186/s41235-021-00301-5

• What it shows. Within a single session, the effect of repetition on belief grows quickly at first, then less and less, following what is known as a logarithmic curve. The biggest gain comes from the second encounter with a sentence, and each repetition after that adds a little less.
• What we take from it for HypnoSelf. Over a single session, each new encounter with a sentence adds a little less: §10 draws from this the idea of an appointment that comes round every evening.

K. Habit, consistency, tracking (§11)

Wood, W., Quinn, J. M. & Kashy, D. A. (2002). Habits in everyday life: Thought, emotion, and action. Journal of Personality and Social Psychology. 10.1037/0022-3514.83.6.1281

• What it shows. In two studies, participants noted every hour what they were doing and what they were thinking about. About four in ten of their daily behaviours were repeated almost every day, in the same context; the figure often quoted is 43%. During these habitual actions, their thoughts were on something other than what they were doing.
• What we take from it for HypnoSelf. It is the measurement that §1 sets against the 95% on our home page, a figure of speech that comes from Lipton and Dispenza: in everyday life, about four actions in ten are repeated on autopilot.

Lally, P. et al. (2010). How are habits formed: modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998-1009. 10.1002/ejsp.674

• What it shows. In all, 96 volunteers chose a simple action to do every day, such as drinking a glass of water, eating a piece of fruit or going for a walk, and rated for 12 weeks how automatic it was becoming. In the 39 people for whom the calculation was reliable, it took a median of 66 days for the action to become automatic: half took less time, the other half longer, and the times ranged from 18 to 254 days. Missing the action once barely affected how the habit formed.
• What we take from it for HypnoSelf. The origin of the 66-day badge, and the reference the whole journey rests on. It also gives §11 two things to tell you: everyone goes at their own pace, and a missed evening has little effect on the habit.

Buyalskaya, A. et al. (2023). What can machine learning teach us about habit formation? Evidence from exercise and hygiene. PNAS. 10.1073/pnas.2216115120

• What it shows. An observational study, which analyses what gyms and hospitals were already recording: 30,110 members (more than 12 million observations) and 3,124 healthcare workers (more than 40 million opportunities to wash their hands). For the gym, the habit takes about four to seven months to take hold (a median of 122 to 226 days), and for handwashing, about two weeks; the authors conclude that there is no magic number for forming a habit.
• What we take from it for HypnoSelf. Depending on the action and the person, it takes much less or much more time, and that is why we present the 66 days as a benchmark.

Lally, P. et al. (2011). Experiences of habit formation: a qualitative study. Psychology, Health & Medicine. 10.1080/13548506.2011.555774

• What it shows. Interviews with 10 people enrolled in a weight-loss programme based on forming habits. At first, they say, the effort is conscious and demanding; then the action gets easier as automaticity sets in.
• What we take from it for HypnoSelf. We draw on this experience, which §11 recounts, to write the messages that support you along the way.

Singh, B., Murphy, A., Maher, C. & Smith, A. E. (2024). Time to form a habit: a systematic review and meta-analysis of health behaviour habit formation and its determinants. Healthcare. 10.3390/healthcare12232488

• What it shows. A meta-analysis of 20 studies covering 2,601 participants. Four of these studies directly measure how long it takes for a health behaviour to become automatic. The medians reported range from 59 to 66 days; in one study on daily stretching, it took an average of 106 days in the morning and 154 days in the evening. The differences between people are wide: for most health behaviours, the authors conclude, it takes two to five months.
• What we take from it for HypnoSelf. It confirms that the 66 days are a median, and that is how §11 presents them. In the stretching study it cites, the action practised in the evening, like your ritual, took longer to become automatic than in the morning; together with Berardi’s study (group H), that is why we offer the morning in the app too.

Gardner, B., Lally, P. & Wardle, J. (2012). Making health habitual: the psychology of “habit-formation” and general practice. British Journal of General Practice. 10.3399/bjgp12X659466

• What it shows. A short article for general practitioners, written by two of the authors of Lally’s 2010 study, described above. It calls the “21 days” a myth, and traces their origin back to anecdotal observations in a 1960 book: cosmetic surgery patients who got used to their new appearance in about three weeks. It also notes that a simple action, such as drinking a glass of water, becomes automatic faster than a more elaborate routine.
• What we take from it for HypnoSelf. §11 opens with this myth of the 21 days, and keeps the median of 66 days as its benchmark. §7 also takes up its point about the simple action: in the evening, all you need to do is start your recording.

ten Broeke, P. & Adriaanse, M. A. (2023). Understanding the setback effect in everyday self-regulation. European Journal of Social Psychology. 10.1002/ejsp.2931

• What it shows. Three studies followed the same people over time, on eating, exercise and saving. A small lapse makes another lapse more likely, and the most consistent mechanism is a drop in how capable people feel.
• What we take from it for HypnoSelf. When your streak breaks, our message is to pick it up again the next day, without judging yourself, because after a lapse, people mostly give up when they feel less capable.

Harkin, B. et al. (2016). Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin. 10.1037/bul0000025

• What it shows. A meta-analysis of 138 trials, covering 19,951 participants, in which chance decided who would be prompted to track their own progress, on behaviours you can count. This tracking helps people reach their goal, with an effect size of 0.40 (95% CI: 0.32 to 0.48) against the comparison groups, an effect between small and medium. The effect is stronger when the information is physically recorded, and when the results are reported to someone or made public.
• What we take from it for HypnoSelf. It is one of the most solid results the app relies on. Every listening session is logged, and your streak appears on your progress screen: that is the first of the two conditions that strengthen the effect in these trials, keeping a record of your progress.

Madigan, C. D. et al. (2015). Is self-weighing an effective tool for weight loss: a systematic literature review and meta-analysis. International Journal of Behavioral Nutrition and Physical Activity. 10.1186/s12966-015-0267-4

• What it shows. A review of trials in people who wanted to lose weight. In the only trial that tested weighing on its own, the effect remained uncertain; added to a programme, weighing strengthened the programme’s effect.
• What we take from it for HypnoSelf. It refines the previous one by showing that tracking works when it accompanies a practice, and in the app, it accompanies your ritual.

Stojanovic, M., Grund, A. & Fries, S. (2022). Context stability in habit building increases automaticity and goal attainment. Frontiers in Psychology. 10.3389/fpsyg.2022.883795

• What it shows. A total of 95 students took up a new study habit for six weeks, 2,482 repetitions in all, some always in the same place and at the same time, the others changing place and time on each occasion. The first group developed more automaticity, and stuck more closely to their repetition goal. The same pattern shows up in 308 habits tracked by 218 users of an app.
• What we take from it for HypnoSelf. It supports a daily appointment at a set time, and it is one of the reasons we chose bedtime, which comes round every evening in the same place, with the same sequence of steps.

Gardner, B., Phillips, L. A. & Judah, G. (2016). Habitual instigation and habitual execution: definition, measurement, and effects on behaviour frequency. British Journal of Health Psychology. 10.1111/bjhp.12189

• What it shows. Students were followed for a week on three behaviours: flossing, snacking, eating breakfast. The researchers distinguish two stages in a habit: getting started (setting about the act) and carrying it out (going through it). For all three behaviours, the habit of getting started was a better predictor of what the students actually did than the habit of carrying them out.
• What we take from it for HypnoSelf. It gives us a realistic goal: that starting your ritual should become a reflex.

Sheeran, P. et al. (2016). The impact of changing attitudes, norms, and self-efficacy on health-related intentions and behavior: a meta-analysis. Health Psychology. 10.1037/hea0000387

• What it shows. 204 experiments in which chance decided who received the intervention, on health behaviours: physical activity, diet, screening. When the feeling of being capable is boosted, behaviour really does change, with an effect size of 0.47 against the comparison groups, a medium effect.
• What we take from it for HypnoSelf. Feeling more capable, which is what you come here for, has measurable consequences for what people do (§11).

Tang, M. Y. et al. (2019). Behavior change techniques associated with changes in postintervention and maintained changes in self-efficacy for physical activity: a systematic review with meta-analysis. Annals of Behavioral Medicine. 10.1093/abm/kay090

• What it shows. In physical activity programmes for adults, the feeling of being capable increases just after the intervention, relative to the comparison groups, with an effect size of 0.26, a small effect. Measured later, some time after the intervention, the effect drops back to 0.08, almost nothing, and chance could explain it.
• What we take from it for HypnoSelf. These gains hold up poorly over time, and we take this as an argument for a daily appointment (§11).

Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological Review. 10.1037/0033-295X.84.2.191

• What it shows. The theoretical paper that founded the notion of self-efficacy, the conviction that we can influence what happens to us.
• What we take from it for HypnoSelf. It provides the framework for the goal of the ritual: strengthening the feeling that you can act (§6).

L. Selective attention (§12)

Folk, C. L., Remington, R. W. & Johnston, J. C. (1992). Involuntary covert orienting is contingent on attentional control settings. Journal of Experimental Psychology: Human Perception and Performance. 10.1037/0096-1523.18.4.1030

• What it shows. Laboratory experiments on what draws our attention in spite of ourselves. Something captures it only if it resembles what we are already looking for: this involuntary capture depends on how attention is set.
• What we take from it for HypnoSelf. It explains the example that opens §12, the car model you start seeing everywhere as soon as it interests you.

Most, S. B. et al. (2005). What you see is what you set: sustained inattentional blindness and the capture of awareness. Psychological Review. 10.1037/0033-295x.112.1.217

• What it shows. The title of the paper, What you see is what you set, sums up the idea, and the paper shows its flip side: what falls outside what we are looking for can go unnoticed, even in plain sight.
• What we take from it for HypnoSelf. This phrase runs through §12, right up to its title, where it becomes what you notice depends on what you look for.

Simons, D. J. & Chabris, C. F. (1999). Gorillas in our midst: sustained inattentional blindness for dynamic events. Perception. 10.1068/p2952

• What it shows. The founding demonstration of inattentional blindness: when your attention is busy elsewhere, you can miss an unexpected event, even a clearly visible one.
• What we take from it for HypnoSelf. We use it as an image for what you believe about yourself (§12): convinced you are not up to it, you can miss what shows you otherwise, even when it is right in front of you.

Hart, W. et al. (2009). Feeling validated versus being correct: a meta-analysis of selective exposure to information. Psychological Bulletin. 10.1037/a0015701

• What it shows. A meta-analysis on the information we choose to look at. We prefer information that agrees with what we think over information that contradicts it, with an effect size of 0.36, a preference between small and medium. It weakens when the person has first felt validated in what they think.
• What we take from it for HypnoSelf. It gives §12 its figure for our preference for what confirms what we think. We also see a link here with the conversation (§3): it always asks you what matters to you before a single sentence is written, and in this meta-analysis, when people have first felt validated, their preference for what proves them right is less marked.

Pool, E. et al. (2015). Attentional bias for positive emotional stimuli: a meta-analytic investigation. Psychological Bulletin. 10.1037/bul0000026

• What it shows. A meta-analysis of 243 studies, covering 9,120 participants. Attention goes more readily to a positive item than to a neutral one, with an effect size of 0.258, a small bias. It is larger when the item touches on one of the person’s own concerns than when it is positive in a generic way, and this gap is greater than chance would explain.
• What we take from it for HypnoSelf. What relates to you draws attention more than generic positivity, and that is one more argument for writing your sentences from what you said in the conversation (§4).

Fodor, L. A. et al. (2020). Efficacy of cognitive bias modification interventions in anxiety and depressive disorders: a systematic review and network meta-analysis. The Lancet Psychiatry. 10.1016/S2215-0366(20)30130-9

• What it shows. A meta-analysis whose plan was made public in advance, which compares several training methods with one another, drawing on 85 trials: 65 in adults with anxiety, 20 in adults with depression. In the anxiety trials, training that modifies interpretation bias, the way we read an ambiguous situation, reduces symptoms more than a waiting list (effect size of 0.55, medium) and more than sham training (0.30, between small and medium). Training that modifies attention bias shows a benefit only under limited conditions (details in §12).
• What we take from it for HypnoSelf. In anxious adults, the training methods that rework interpretation do best, with a modest effect against sham training. An affirmation ritual aims at the same target by a different path: your sentences offer you another reading of what you experience (§12).

What we built into the app

Most of these references have left their mark on the app. Here we go over the main ones, from book to practice.

• The conversation, before any sentence. It comes from Braden, Dispenza and Lipton. The studies on reflecting on your values support it (Steele, Epton, Zhang). Those on motivational interviewing give it its shape: a discussion focused on what you want to build, which your daily ritual carries on (Smedslund, Magill, Pace). See §3.
• Sentences written with you. In Braden’s book, it is already up to the reader to fill in the method with their own words. Writing your sentences with you, from what you said in the conversation, is our contribution. Nye, Li, Noar and Jabs measure what tailoring to the person brings: a small advantage. See §4.
• Your voice. Coué made sure that whoever said his sentence heard it, and Dispenza suggests recording yourself. Research on memory backs this choice: what you read out loud, in your usual voice, is remembered better than what you read silently (the production effect, Fawcett; the role of the usual voice, Wakeham-Lewis), and what you hear again in your own voice is remembered better than what another voice reads to you (Forrin and MacLeod’s ranking). See §5.
• Several forms of sentence, mixed together, and a tone of kindness. “You” and your first name help you step back in a difficult moment (Kross); the studies that directly compare “you” and “I” give mixed results (Magata and Tanaka). The mix is our choice, and the books themselves also mix “I” and “you”. Self-compassion, talking to yourself kindly, is better supported (Kirby, Wakelin) than the ready-made positive sentence you repeat to yourself: in people with low self-esteem, that sentence left them in a worse mood than repeating nothing (Wood, Perunovic and Lee). Hence the tone of your sentences, and the sentences that describe a path when a belief pushes back. See §6.
• Listening at bedtime, which stops by itself. The moment comes from Murphy, and Hill and Stone, Dispenza and Coué point to the same one. Jespersen measures what listening to music every day does for sleep, as people perceive it. Lacaux and colleagues observe that those who had spent a minute on the threshold of sleep then found the hidden rule in an arithmetic problem more often than those who stayed awake. The sleep timer stops playback after 30 minutes, a setting you can change: you understand your sentences while you are awake (Makov), and after that, sleep takes over. The morning is offered too, because a morning practice holds up better (Berardi). See §7 and §8.
• Consistency, and a missed evening that has little effect on the habit. The idea comes from Hill and Stone, and from Doidge. The measurements come from Lally, for the median of 66 days and that missed evening; from Stojanovic, for the same time every day; from Harkin, for recorded tracking; and from ten Broeke, for what makes people give up after a lapse, a drop in how capable they feel. Buyalskaya and colleagues point out that the time it takes varies a great deal from one action to another. See §11.
• Another reading of what you experience, with sentences that concern you. Specific kinds of training, which work on the way we see things, have been compared in anxious adults: those that teach you to read a situation differently do best, with a modest effect against sham training (Fodor). Your sentences aim at the same target by a different path: the way you read what you experience. What touches on a concern of your own also draws attention more than generic positivity (Pool). See §12.
• A ritual practised in full knowledge of what it is. The question comes from Dispenza’s You Are the Placebo. Your sentences are spoken plainly, with nothing hidden, and you are the one who chose them. Fendel and colleagues brought the trials together: a placebo taken knowingly helps people feel a little better. See §2 and §7.
• A personal development ritual that accompanies your everyday life. Coué already placed his method alongside medical treatment (§2). If you are receiving care for a health condition, the ritual complements that care, and if you are going through a difficult time, talk to a doctor or a psychologist about it. See §13.

References for this section

10.1038/s41598-025-14895-z · 10.1038/s41598-023-29624-7 · 10.1038/s41598-024-56600-6 · 10.1097/00000658-199903000-00014 · 10.1089/acm.1999.5.509 · 10.1097/00000542-200005000-00013 · 10.1016/j.jphysparis.2016.01.001 · 10.1093/cercor/bhw220 · 10.1126/science.277.5328.968 · 10.3390/brainsci14020115 · 10.3389/fpsyg.2023.1330238 · 10.1016/j.clinph.2019.11.011 · 10.3389/fpsyg.2022.807741 · 10.1093/nc/niae036 · 10.1016/S0065-2601(08)60229-4 · 10.3390/healthcare12010003 · 10.1037/hea0000116 · 10.1037/amp0001591 · 10.1371/journal.pone.0062593 · 10.1073/pnas.1500247112 · 10.1093/scan/nsv136 · 10.1037/a0029937 · 10.1037/a0027231 · 10.1002/14651858.CD008063.pub2 · 10.1037/ccp0000250 · 10.1037/adb0000280 · 10.1016/j.beth.2012.05.003 · 10.1159/000365764 · 10.1146/annurev-clinpsy-032813-153734 · 10.1037/ccp0000820 · 10.3389/fpsyg.2024.1470817 · 10.1037/0033-2909.133.4.673 · 10.1111/eip.70207 · 10.1016/j.actpsy.2012.10.001 · 10.1037/a0018785 · 10.1037/cep0000302 · 10.1080/09658211.2012.699070 · 10.1037/a0033337 · 10.3758/bf03193441 · 10.1080/09658211.2022.2115075 · 10.1080/09658211.2017.1383434 · 10.1037/0033-2909.121.3.371 · 10.1093/scan/nsn014 · 10.7717/peerj.14976 · 10.1093/scan/nsu031 · 10.1098/rsos.221561 · 10.1371/journal.pone.0199443 · 10.1080/16506073.2015.1126633 · 10.1037/a0035173 · 10.1177/2167702620951539 · 10.1038/s41598-017-04047-3 · 10.1016/j.ijpsycho.2022.05.003 · 10.1177/10888683211069025 · 10.1002/ejsp.2048 · 10.1371/journal.pone.0305251 · 10.1111/j.1467-9280.2009.02370.x · 10.1016/j.beth.2017.06.003 · 10.1002/cpp.2586 · 10.3389/fpsyg.2014.00421 · 10.1080/17437199.2019.1594332 · 10.1002/14651858.CD010459.pub3 · 10.1016/j.ijnurstu.2017.10.011 · 10.1155/2021/5924040 · 10.1016/j.smrv.2017.02.001 · 10.1126/sciadv.abj5866 · 10.1038/s41598-023-31361-w · 10.1093/nc/niw014 · 10.1093/pnasnexus/pgae479 · 10.1523/JNEUROSCI.0168-17.2017 · 10.2196/42482 · 10.1016/j.cnp.2017.07.002 · 10.1016/j.cub.2011.11.018 · 10.1038/427311a · 10.1146/annurev.neuro.31.060407.125639 · 10.1016/j.copsyc.2023.101736 · 10.1186/s41235-021-00301-5 · 10.1037/0022-3514.83.6.1281 · 10.1002/ejsp.674 · 10.1073/pnas.2216115120 · 10.1080/13548506.2011.555774 · 10.3390/healthcare12232488 · 10.3399/bjgp12X659466 · 10.1002/ejsp.2931 · 10.1037/bul0000025 · 10.1186/s12966-015-0267-4 · 10.3389/fpsyg.2022.883795 · 10.1111/bjhp.12189 · 10.1037/hea0000387 · 10.1093/abm/kay090 · 10.1037/0033-295X.84.2.191 · 10.1037/0096-1523.18.4.1030 · 10.1037/0033-295x.112.1.217 · 10.1068/p2952 · 10.1037/a0015701 · 10.1037/bul0000026 · 10.1016/S2215-0366(20)30130-9