Voices as company: how do children and adults with intellectual disability experience voices?
Imaginary friends, self-talk, inner speech and the brain.

Earlier this year, I had a chance to assist caregivers in their support and guidance duties for adults with an intellectual disability. During one of the “stay-at-home” days, we talked about our hobbies and passions. To showcase their hobbies, several guys brought their remote-controlled cars and organised a mini F1 race. The house was buzzing with excitement. One resident, who was among the cheering team, went upstairs to get his tall and very impressive toy – Robin de Robot. The robot was brought down to see this rather important day event, the robot was addressed, the robot even spoke to me and the caregiver (the resident made pauses for the Robin de Robot to introduce himself and share his thoughts).
When we went out to the playground, Robin de Robot came with us. The resident kept up an engaging conversation with his friend, prioritised it over the swings, and overall seemed to have a wonderful time showing Robin de Robot around.
Considering the fact that the resident doesn’t only talk to Robin de Robot himself, but is also certain that Robin talks to other people, I was rather puzzled by a question: is the voice of the robot really heard, or is it a convincing internal rehearsal of what the robot would actually say? Can it be that the resident simply has a hard time telling this internal speech apart from reality?
I’m going to explore what is known about both possibilities, dive into how children hear their imaginary friends and analyse what happens when the voices aren’t outgrown.
What is known about voice-hearing before the age of 7
Imaginary friends are common. The developmental psychologist Marjorie Taylor and her colleagues followed 100 children from preschool into school age, and by the age of 7, 65% of them had had an imaginary companion at some point.1 Tracy Gleason, who studies children’s relationships, sees a rehearsal in this play: with an imaginary companion, a child can try out what might happen in real relationships and feel “all manner of positive and negative emotions in a risk-free way”.2
This rehearsal often comes with a voice. When Charles Fernyhough and his colleagues asked more than 1,400 adults about imaginary companions, 69% of those who had ever had one remembered hearing its voice at least once.3
But what exactly does a young child hear when an imaginary friend speaks?
To understand what kind of voice this might be, it helps to look at how children learn to think in words. Before they can think silently, young children think out loud, keeping up a running commentary on what they are doing, a habit psychologists call private speech. This talking to oneself is most common around the age of 5. After that, as the psychologists Ben Alderson-Day and Charles Fernyhough put it, it “goes underground”: it fades first into whispers and mutters, and finally into the silent inner speech we use as adults.4
Until the age of 6 or 7, children might not yet independently understand that a mind can hold a silent inner dialogue.5 I suspect this also goes hand in hand with whether anyone has ever explained inner speech to the child. A preschooler who describes their thoughts as “voices” may simply have no other word for them.6 When a child recalls exactly what their mother told them, in their mother’s own voice, they may well call it a voice.
Adults do this too: hearing other people in one’s inner speech is common. In a questionnaire answered by 377 university students, about 59% of the answers to statements such as “I hear other people’s actual voices in my head, saying things that they actually once said to me” confirmed this experience.7
Most of us know that these remembered voices come from our own mind. Why, then, would such a rehearsal ever feel like someone else speaking? A small study of healthy adults who hear voices offers a clue. When they imagined a voice on purpose, a motor planning area in the brain fired before the hearing areas. When a voice came by itself, the two fired at the same moment.8 An imagined voice carries a signal that we started it. The voice that arrives without this signal consequently might be experienced as external.
For the voices of young children, this leaves two possibilities:
- The voice comes on its own. A child’s brain may not yet reliably mark the speech it produces as its own, so a remembered phrase, perhaps in mom’s voice, arrives as if from outside. The child is not in charge of it, which makes it close to a hallucination.
- The child is in charge. They deliberately recall or imagine exactly what mom said, in mom’s voice, and having no word yet for inner speech, they call that vivid recollection “a voice”.
By the age of 7 or 8, the picture seems clearer: in a pilot study in Groningen, children of that age could distinguish their voices from fantasy and from real voices.9
What changes when a child becomes a teenager
Voices don’t simply appear in adolescence. In the province of Groningen, Agna Bartels-Velthuis and her colleagues asked 3,870 children aged 7 and 8 about voices, and 9% had heard one in the past year.9 Across studies, hearing voices is about as common in teenagers as it is in children.10
There’s a strong connection between the maturing of the brain and its susceptibility to continuous voice-generation in adulthood. Many voices that last a lifetime begin in adolescence: in Utrecht, Iris Sommer and her colleagues studied 103 healthy adults who heard voices often, and on average their voices had begun at the age of 14.11 At the same time, most childhood voices fade as the brain matures: in a Groningen study published in 2016, researchers followed 293 children from the age of 7 or 8 to 18 or 19, and of those who had heard voices at the start, only 6.2% still heard them at the end.12
The large European IMAGEN study followed how the brain develops through the teenage years. Raka Maitra, Charlotte Horne and their colleagues at King’s College London, working with research teams across the IMAGEN consortium, scanned the brains of 1,252 teenagers in the UK, Germany, France and Ireland at the age of 14 and again at 19, and at 19 asked them about unusual, psychosis-like experiences.13
Brain folding, the pattern of ridges and grooves on the brain’s surface, was the focus of the study. Most of these folds form before birth and in the first two years of life, and then stay fairly stable. Teenagers who often heard voices, received unusual messages or felt persecuted differed in how a few language-related areas on the left side of the brain were folded, among them the superior temporal sulcus, an area that processes voices.13,8
As the brain matured between 14 and 19, these teenagers also developed more folds in the parietal cortex on the right side of the brain. The authors suggest this may be the brain compensating: this area helps us monitor what we produce ourselves, the same ability that tells us a voice is our own.13 The finding describes the group as a whole, so it is still unclear whether this extra folding helps a voice fade, or whether teenagers who don’t develop it are the ones whose voices persist.
The age at which voices begin may also influence one’s perception of the voices, and how one coexists with them in the future. The Utrecht study asked healthy adults who had lived with voices for decades about when their voices first appeared and what they were like. On average, the voices began at the age of 14, and 71% of the adults participating in the study had never heard anything negative from the voices.11 A systematic review of the literature by David Baumeister and colleagues found the same pattern more broadly: the voices of healthy voice-hearers tended to begin earlier than the voices of people with a psychotic disorder, and in every comparison, they were also less negative.14
Why some teenagers outgrow their voices and others don’t
In Maastricht, Sandra Escher and Marius Romme followed 80 children and teenagers who heard voices, and within three years 60% of them no longer did.15
What helps a voice fade is less certain. In the Maastricht study, voices that came only in a particular place or at a particular time faded more easily than voices that were always there.15 In the Generation R study, which has followed thousands of Rotterdam children since before they were born, voices that were more severe, or came with experiences in other senses, lasted longer, while higher self-esteem helped them fade.16 In Groningen, voices that persisted into adulthood were linked to traumatic events.12 Still, none of these factors lets doctors predict which child’s voices will stay.16
How a teenager lives with their voices seems to depend a great deal on the people around them. In a study of 74 young people from 11 countries, other people’s anxiety, stigma and isolation made the voices more distressing and even changed their character.17 In the Maastricht study, whether a child needed care depended “to a large part” on how the child and their parents saw the voices, “rather than the perception itself”, and children who received professional help were no more likely to stop hearing voices than those who did not.15
In a review of research on hallucinations in children and teenagers, an international group of researchers led by Renaud Jardri concluded that for most children who hear voices, “normalization, support, and reassurance may be sufficient”, ideally with the parents involved.6 A review from UMC Utrecht adds that even distressing voices do not always mean a psychotic disorder, which “warrants restraint in prescribing antipsychotic medication”, and that there is no need to screen all children for voices.18 Doctors are still advised to ask about suicidal thoughts, especially when the voices are mean.18,19
What adults with an intellectual disability might hear
So what does the resident of the supported living home hear when Robin de Robot speaks? There are three possibilities, and from the outside, it is hard to know which applies.
The first possibility is a conscious rehearsal. The resident may be doing what young children do: imagining how a robot would sound, and what it might say.
Dutch disability care looks at a person’s emotional development, not only at their intellect, building on the work of the psychiatrist Anton Došen.20
Some fantasy thinking therefore belongs to a low level of emotional development, in which fantasy and reality have never been separated.21
The German physician Tanja Sappok describes a 23-year-old woman with Down syndrome who held long conversations with an imaginary partner, a pop star she believed she was married to. She had been referred with suspected schizophrenia, but her emotional development matched that of a child aged 4 to 7, an age at which intense fantasy is typical, and antipsychotics were not indicated.22
The second possibility is self-talk that never went underground. At the Adult Down Syndrome Center near Chicago, Dennis McGuire and Brian Chicoine found that about four in five of their adult patients continue to talk aloud to themselves. Some keep imaginary companions for decades. These companions, they write, “offer support, comfort, and companionship”, and dolls, action figures and stuffed animals “are often assigned unique and creative roles”.23 They also met many adults who had been given antipsychotic medication “simply for talking out loud to themselves, and unfortunately this still happens today”.23
The third possibility is a voice in the full sense. Psychosis (opens in a new tab) is more common in people with an intellectual disability than in others.24 It is also harder to recognise, as a diagnosis rests largely on what a person can tell about their experiences. The European guideline on psychiatric diagnosis in adults with an intellectual disability adds that an adult who speaks “to an object such as a tree or a table” should not be diagnosed with schizophrenia on that basis alone.25
The character of the true auditory hallucinations of people with an intellectual disability hasn’t been fully studied yet. It’s not excluded that a voice helps a person with an intellectual disability in more ways than simply keeping them company: for example by offering clues and insights, and by pointing out what their consciousness might have missed. That would explain why the resident described in this article kept Robin de Robot in such high regard, almost as a totem, conversed with him, and brought him along when out-of-ordinary events occurred. Such ideas are, of course, grounded in nothing but observations. Nonetheless, voices seem to be more readily accepted when a person’s developmental path hasn’t imprinted a coherent narration of how thinking must work, and of where their thoughts come from.
To conclude, it may be how we perceive a voice that makes it what it is: how vivid it is, how much we feel in charge of it, and how the people around us respond to it. Tanya Luhrmann and her colleagues suggest that a voice may be “less like a misattributed thought and more like a very vivid thought which, because it is felt as vivid, is experienced as sensory.”26



