Cognitive-Developmental Mechanisms in Hallucinations

Abstract

Hallucinations figure prominently in a range of psychiatric disorders but, to date, their developmental origins are not well understood. The aim of the present article is to explore how ideas from mainstream developmental psychology can enhance understanding of how hallucinations develop in different modalities across the lifecourse. Hallucinations vary in their clinical significance depending on at what point they occur in the lifetime of the individual. Key cognitive-developmental processes include engaging with imaginary entities, exposure to adverse events, executive functioning, social cognition, and language development. The presentation of hallucinations in certain developmental conditions suggest that atypical developmental trajectories can also play a key role in shaping hallucination prevalence and phenomenology. In considering prospects for future research at this interface, we propose that two-way benefits may result from further close integration between developmental and psychiatric approaches to hallucinations.

Introduction

Hallucinations in various sensory modalities feature in a wide range of psychiatric disorders and medical conditions¹ but, to date, very little is understood about their development across the human lifecourse. Hallucinations in childhood and young adulthood have complex and varied clinical significance,² and can be understood to exist on a continuum from typical to psychopathological experience.³ Although cognitive and neural mechanisms underlying hallucinations in adulthood have been a focus of extensive research,^4,5 little effort has been expended on understanding the cognitive-developmental mechanisms that might underlie hallucinations in earlier life, and particularly in situating them in the context of research in mainstream developmental psychology. With the present narrative review, we (as a working group of the International Consortium on Hallucination Research [ICHR]) aim to establish a deeper understanding of hallucinatory experiences in young people, and to develop fresh perspectives on how relevant cognitive and neural processes operate in typical development, including in everyday, non-pathological experiences such as fantasy play and interacting with imaginary companions.

Attempting such a synthesis seems particularly timely given that many of the domains of cognitive function that are mostly heavily implicated in hallucinations (eg, executive functioning, metacognition, theory of mind [ToM], multimodal integration) are the subject of active research programmes in developmental psychology, suggesting fruitful nodes of interaction among cognitive science fields. A developmental approach can also promote caution in assuming that models and theories that are useful in explaining adult experience will be equally valuable in developmental groups.⁶ The adult brain is often an inaccurate guide to the developing brain,⁷ and neurodevelopmental models of psychotic phenomena often only focus on relatively late-maturing processes and structures (such as the link between prefrontal cortical development and reality monitoring⁸). In contrast, models that originate in developmental research may be particularly well-placed to unpick some of the paradoxes about the occurrence of hallucinations in the early stages of the lifecourse.² Across the neural, cognitive, personal, and sociocultural levels of explanation,⁵ factors that influence the occurrence of hallucinations in adulthood may not have equivalent effects earlier in life. For similar reasons, the clinical significance of hallucinations in adulthood cannot be assumed to be identical to that of experiences occurring in childhood and adolescence.^2,3

In this article, we outline some possible benefits of a closer focus on cognitive-developmental mechanisms in research on hallucinations, drawing connections with the mainstream developmental psychology literature and suggesting opportunities to make progress in model-building around new findings emerging in these areas. We build on our previous ICHR working group article² by reviewing research conducted since that review appeared in 2014, including work on hallucinatory experiences in early childhood, relations with trauma, attachment and imaginary companions, and diverse developmental trajectories such as 22q11.2 deletion syndrome. Throughout, we ask what a developmental approach can add to existing understandings emerging from research in adult psychopathology, and how we should understand the predictive significance of hallucinations for mental health later in life.

Developmental Time Course of Hallucinations

Hallucinations in children and adolescents occur in several sensory modalities. In a sample of adolescents aged 14–20 years,⁹ subclinical hallucinations were observed most frequently in the auditory and visual modalities (74% of cases). These often include hearing one’s name being called, photopsia (flashes of light appearing without any light source), acoasms (nonspecific auditory hallucinations), altered perception of auditory stimuli, and seeing shapes of unknown humans or monsters. Visual and auditory subclinical hallucinations co-occurred in 22% of cases. In the remaining modalities, most commonly reported were olfactory (in the form of sudden unpleasant smells) or bodily hallucinations (in the form of electrifying sensations, feelings of pressure or burning in the head). Several studies also highlight that multisensory hallucinations are particularly frequent in childhood and adolescence.^10,11 A longitudinal case–control study of children with and without auditory verbal hallucinations (AVH)¹² demonstrated that roughly 50% of children aged 12–13 years who reported AVH also reported hallucinations in another modality. Hallucinations in children and adolescents are sometimes difficult to distinguish from typical developmental processes,¹³ such as having an imaginary companion (see below).

Most young individuals reporting hallucinations do not experience any distress resulting from their symptoms and do not require clinical care. However, associated distress is reported by a small subgroup of individuals,^14-16 and for some the presence of hallucinations in childhood could be a risk factor for future development of mental health problems in young adulthood.^17,18 In addition, a subgroup of youth reporting hallucinations in the general population may actually be in need of care while currently not receiving it¹⁹. Overall, it is accepted that hallucinations are common and less severe in non-help-seeking youths, and could be part of typical developmental processes.²⁰ The prevalence of attenuated psychotic manifestations—including hallucinations—was examined in a representative community sample of 2916 individuals from the BEAR and BEARS-Kid studies aged 6–40 years using semi-structured clinical interviews.²¹ The prevalence of any current subclinical hallucinatory experiences decreased with age, ranging from 27.4% (8–12 years) to 6.3% (30–40 years). Similarly, a review specifically assessing the prevalence of auditory hallucinations demonstrates comparable, but slightly lower, prevalence rates (12.7% in children and 12.4% in adolescents²²). These findings are confirmed by studies in clinical samples, demonstrating a decline in prevalence of both verbal and visual hallucinations during adolescence.^9,18

Results from both community and clinical samples thus suggest that the prevalence of hallucinations significantly decreases throughout adolescence. Anecdotally, individuals who report hallucinations during adolescence do not necessarily recall having experienced such phenomena when followed up longitudinally.²³ In addition, lower clinical significance has consistently been reported in younger age groups.¹⁸ For example, an age threshold at 18 years has been identified for perceptual symptoms (including hallucinations), with higher prevalence but lower clinical significance of these manifestations below this threshold.²¹ This is consistent with previous findings highlighting lower conversion to psychosis in clinical samples meeting clinical high-risk criteria for psychosis mostly composed of children and adolescents.²⁴ Although hallucinations during adolescence may hold little clinical relevance on their own, they tend to co-occur with other debilitating mental health concerns. For example, a longitudinal case–control study demonstrated that adolescents aged 18–19 years with AVH had higher levels of depression and anxiety, and reported more traumatic experiences, than their peers without AVH.¹⁵ In addition, a study²⁵ using data from the general population-based Copenhagen Child Cohort 2000 found that the majority of children at age 11–12 years with psychotic experiences met criteria for a diagnosable mental health disorder (31.4%) or self-reported mental health complaints without a diagnosis (31.4%). The presence of hallucinations in adolescence may thus serve as a clinical marker for other co-occurring mental health problems which could warrant treatment.²⁶

The question remains of whether it is possible to predict the course of hallucinations in terms of persistence or remittance by examining risk factors. In a longitudinal cohort study of adolescent outpatients,⁹ it was found that subclinical hallucinations fully remitted in over half of their sample of adolescents, and partially remitted in two-thirds of adolescents over a period of one year. The persistence of hallucinations in nine adolescents (32.1%) was predicted by global level of functioning at baseline, but not cannabis use or psychiatric and psychological treatment. This is in line with a systematic review on the longitudinal course of hallucinatory experiences in childhood and adolescence, which found that in the majority of cases (75%) hallucinations remit over time.²⁷ A longitudinal study of a baseline case–control sample of youth with follow-up periods of 5 and 11 years¹⁵ demonstrated that the persistence rate declines throughout childhood, ranging from 18.2% in a 6-year period from ages 12–13 to 18–19 years, to 6.2% in an eleven-year period from ages 7–8 years to 18–19 years. Hallucinations that persisted over a period of 11 years were associated with more psychotic experiences, psychopathology, and traumatic events at age 18–19 years. Furthermore, a study¹⁶ found that, of 1841 adolescents aged 12 years, 123 (7%) reported hearing AVH, of which 43 adolescents (35%) continued to do so over a 12-month follow-up period. Remittance of AVH was predicted by a number of factors, including psychopathology (eg, paranoia, anxiety, and dissociative symptoms) and environmental factors (eg, cannabis, parental divorce, and school grade repetition). A recent review of the literature²⁸ examined the predictors of persistence or discontinuation of psychotic-like experiences—including hallucinations—throughout adolescence. The review highlighted highly inconsistent results across studies. Importantly, hallucinatory and delusional/unusual experiences were mostly combined in the reviewed studies, which prevented the examination of specific predictors. Risk factors for AVH persistence were examined in a series of studies and highlighted the role of concurrent depressed mood, lower self-esteem, delusional ideation, general psychopathology, non-verbal IQ scores, and adolescent onset of AVH.^29-31 Overall, it appears that hallucinations in childhood and adolescence are transient in most cases, highly dynamic, and can become persistent if the individual is exposed to a number of both environmental as well as psychopathological risk factors.

Core Cognitive-Developmental Processes in Hallucinations

In this section, we emphasize a developmental approach to processes implicated in hallucinations. Key principles here are that any individual process may vary in its significance over the life-course; that processes can usefully be described at biological, cognitive, and sociocultural levels; and that processes described at different levels of explanation may interact in significant ways.

Engaging With Imaginary Entities

Imaginary companions (ICs) are a normative play behavior seen in childhood across cultures.^32,33 Svendsen³⁴ was the first to describe the phenomenon as involving “an invisible character… having an air of reality for the child, but no apparent objective basis” (p. 988). The classification was later expanded to encompass toys which children gave a personality to while playing.³⁵ This description of ICs is congruent with the definition of hallucinations as involuntary perceptions that occur in the absence of external stimuli.³⁶ In addition to having similar psychological, biological, and perceptual properties as humans,³⁷ ICs can be experienced as being out of the child’s control,³⁸ and children can have positive, negative, and neutral interactions with their companions.³⁹ Children often report that they can “see” and “hear” an IC as an independent agent, supported by an experimental study where children have realistic perceptual experiences of imagined characters.⁴⁰

Parallels with AVH include children with ICs being more likely to report hearing words that were not present in ambiguous voice stimuli.^41,42 These studies take the Vygotskian stance that children internalize social speech exchanges to form dialogic inner speech.⁴³ This process of internalization could result in a child’s mental world being populated with voices in dialogic exchange, with children creating ICs to make interchanges more socially acceptable.^41,42,44

The existence of an IC in childhood has been found to relate in adulthood to perspective-taking, anthropomorphizing tendencies, and scene construction and imagery.^45-49 Adult report of a childhood IC has been found to be predictive of prodromal report of hallucination symptoms.⁴⁵ Using a signal-detection task, individuals in a non-clinical population with a history of ICs were shown to be more prone to AVH.⁴⁶ However, one prospective study reported that having ICs in middle school was not necessarily an early marker of mental ill-health at the end of high school.⁵⁰ It may thus be important to consider the developmental trajectory of IC creation and experience to determine where non-clinical hallucinations may convert to psychiatric symptoms. One possible factor is adverse experiences, with one study showing that the relation between childhood IC experience and adulthood hallucinations was partially mediated by childhood adversity.⁴⁵ Qualitative analyses indicate that childhood IC experience may serve a protective function, but that childhood ICs can evoke negative reactions from parents, potentially relating to clinical hallucinations.⁵¹

In terms of hallucination pathways, the development of ICs occurs alongside the emergence of mental-state as well as fantasy/reality reasoning.^32,52,53 Researchers have investigated typically developing children’s knowledge of their own IC’s reality status.⁵³ One of the hallmarks of IC creation is that children understand that their IC is not real, but are acting as if it is present.⁵⁴ Children with ICs and those with high fantasy orientation have been found to excel at the report of attributes associated with real entities (whether or not the being was in fact real), and acknowledge uncertainty around reality status more than their peers.⁵⁵ It is theorized that this is because of the interaction high-fantasy children have with invisibility, reality status, and imagination.⁵⁶ In summary, ICs and AVH appear to overlap in a range of relevant processes (eg, reality discrimination), without either experience clearly implicating the other.

Attachment and Adversity

A large body of evidence supports the existence of a causal relation between childhood trauma and hallucinations in adulthood.^57,58 Albeit much less studied, this causal relationship has also been established for childhood hallucinations. For example, a population study in 401 children up to age 13 showed that children with adverse life experiences related to their hallucinations were more likely to have hallucinations in multiple modalities, and experience distress from them.⁵⁹ Physical abuse and bullying have been associated with the occurrence of AVH in 13–16 year-old children, and with subsiding of the AVH when the abuse stopped.⁶⁰ A longitudinal study showed that particularly traumatic events played a role in the persistence of early childhood AVH into adolescence.¹⁵ Such observations raise the question of why the experience of traumatic events in childhood predisposes an individual to hallucinations. Among hypothesized mechanisms are theories based on the neurodevelopment of stress-regulating processes, attachment-related processes, and cognitive development.

These theories are not mutually exclusive; rather, each focuses on trauma-related disturbances in different aspects of childhood development. The traumagenic neurodevelopmental model⁶¹ hypothesizes that early childhood trauma results in neurodevelopmental changes to the stress system, specifically to stress and emotion-regulating systems in the HPA axis and the hippocampus, in turn increasing vulnerability to psychosis-like experiences such as hallucinations.

In relation to attachment, the Cognitive Attachment model of Voice-hearing (CAV) aims to provide a holistic account of AVH, specifically in adults with disorganized attachment patterns.⁶² It proposes that individuals become vulnerable to AVH following trauma through negative interpersonal schemas, trauma-related dissociation, and aberrant source monitoring. Studies supporting the CAV model typically focus on explaining the relation between attachment and distress from AVH, rather than on its relation with more general (ie, also non-distressing) AVH. It has been shown that adult attachment anxiety is related to AVH distress and severity, and attachment avoidance is related to AVH distress.⁶² Several studies have reported potential factors mediating the relation between adult attachment issues and distress from AVH, for example through negative interpersonal schemas. Specifically, adult attachment anxiety and avoidance linked to childhood sexual and emotional abuse were associated with negative beliefs about the self and others, and persecutory beliefs about voices and beliefs about voice malevolence, in turn leading to AVH distress.^63-65 Other potential mediating factors between trauma-related attachment issues and hallucinations are trauma-related dissociation⁶⁶ and disturbance of the development of a coherent sense of self.⁶⁷ However, given the lack of evidence for stability in attachment quality between infancy and adulthood,⁶⁸ it is important not to draw unwarranted conclusions about causal pathways between childhood attachment quality and AVH in adulthood.

The CAV model highlights that childhood traumatic experiences may lead to interruptions to cognitive development. Childhood trauma is a common factor underlying psychotic experiences as well as cognitive dysfunction (eg, frontal lobe functions such as executive functioning).⁶⁹ In addition, poor metacognitive ability has been suggested to underlie the relation between childhood trauma and psychotic symptoms.⁷⁰ One such metacognitive function is source monitoring: the process of discriminating between internally- and externally-generated experiences. The CAV model specifies that dissociation following trauma triggers AVH only in those who are vulnerable to source-monitoring difficulties.⁶⁶

A general question for future research concerns the extent to which cognitive factors act as correlates, mediators or moderators of the relation between childhood trauma and hallucinations. Several authors have identified a region-specific developmental vulnerability for stress, with hippocampal regions more impacted when trauma occurs before puberty, and prefrontal areas more affected by adversity occurring later.^71,72 One implication is that trauma will disrupt fronto-limbic system regulation of sensory cortical areas, thus leading to hallucinatory experiences, with these effects unfolding differently depending on whether exposure occurs before or after puberty (with pre-pubertal exposure leading to a greater influence of subcortical areas, and vice versa).

Executive Function and Social Cognition

Many theories of hallucination have posited disruptions of some form to source monitoring or cognitive control,⁷³ but these have rarely been studied in children and young people (perhaps due to the complexities of the paradigms often used). However, as higher-order cognitive functions requiring metacognitive strategies and processes, such skills can be broadly considered under the umbrella of executive functions, on which there is considerable developmental work. Most relevant is inhibition, which represents a potentially important mechanism for understanding hallucination from the cellular level up to the experience of intrusive and egodystonic content.⁷⁴ Basic inhibitory control processes develop relatively early in childhood compared to other executive skills such as planning and working memory, although more specific components (such as intentional cognitive inhibition) are unlikely to develop until the early teenage years. This would correspond to the emergence of difficulties with controlling intrusive thoughts and ruminative thinking in adolescence.⁷⁵

ToM or social cognition has also been identified as a relevant factor in the emergence of hallucinations.^2,76 Indeed, monitoring the source of an experience as arising from either within or outside the mind seems to depend on at least a basic understanding of mental representations. Children have typically developed a representational theory of mind by around age 4⁷⁷ and show an awareness that other people can be engaged with a stream of consciousness that is different from their own at a slightly later age.⁷⁸ Understanding that people’s minds can be occupied by a private, silent inner dialogue is not typically established until age 6 or 7 in US samples.⁷⁹ For these reasons, expecting preschool-age children to understand the difference between typical inner speech and a hallucination seems unreasonable.

Nevertheless, maturation of social cognition pathways is likely to be highly relevant for the emergence of hallucinations in the developing brain. For instance, key tracts supporting these processes in the default mode network (DMN) are some of the last to mature in adolescence⁸⁰ and momentary changes to the functional integrity of the DMN have been linked to hallucinations in young people with brief psychotic disorder.⁸¹ Vulnerability to hallucinations may also originate during fetal brain development. Unlike the plastic changes seen during the transition to psychosis in late adolescence, certain early and stable neurodevelopmental disruptions occurring in the last trimester of gestation can be detected through structural MRI later in life. Sulcation, in particular, can serve as a valuable indicator of the interplay between genetic predispositions and environmental factors during the pre- and perinatal periods.⁸² Notably, reduced cortical gyrification has been reported in schizophrenia patients with a higher developmental burden, such as those with drug-resistant hallucinations⁸³ or multimodal hallucinations.⁸⁴ This makes reduced sulcation a compelling candidate marker for hallucinatory experiences of developmental origin.

Language

Language acquisition occurs naturally in typically developing children with adequate input, predominantly through statistical learning mechanisms.⁸⁵ A critical period around puberty limits full proficiency in a first or second language thereafter.^86,87 Language delays, as seen in autism or intellectual disabilities, are associated with an increased risk of developing AVH⁸⁸ and schizophrenia.^89,90

Bilingual children may experience initial delays in language mastery but typically catch up. In adults, bilingualism has been linked to cognitive benefits,^91,92 known as the “bilingual advantage”,⁹³ though evidence is mixed.⁹⁴ Multilingualism may confer protection against psychosis by enhancing cognitive control.⁹⁵ However, migration, particularly linguistic distance from the host language, is a known risk factor for psychosis,⁹⁶ potentially due to psychosocial disempowerment⁹⁷ or social defeat.⁹⁸ AVH in bilinguals may be shaped by language acquisition history, with late L2 learners less likely to hear voices in their second language.⁹⁹ Finally, ToM development is closely linked to language acquisition, as language provides the necessary tools to represent and communicate mental states. Understanding that others have thoughts, beliefs, and intentions different from one’s own requires the ability to label and discuss these internal experiences. Studies show that children with more advanced linguistic skills develop ToM earlier, as exposure to mental-state language (eg, think, believe, know) helps them grasp the concept of different perspectives.¹⁰⁰

Evidence from deafness is also relevant as a contrast case. As 95% of deaf children have hearing parents,¹⁰¹ language delays are common in deafness and vary considerably depending on how parents and educators support children’s communicative development (with the best outcomes usually linked to early access to sign language). There is evidence that this variety in communicative history is reflected in the phenomenology of psychosis in adulthood. Atkinson et al.¹⁰² documented a wide range of multimodal hallucinations being described as “voices” by deaf adults with psychosis, with communicative experience broadly mapping on to the modality and nature of hallucinatory experiences. For instance, those who had used sign language for many years were more likely to experience visual sensations of signing and fingerspelling, while those with some experience of sound before losing their hearing reported literal auditory experiences. Notably, those with early language difficulties (and no access to signing) experienced non-verbal auditory phenomena, despite later improvements in their language proficiency. These findings combine to suggest that periods of early language development can do much to shape the later experience of hallucinations, with communication (and its difficulties) providing the through-line from early years to adulthood.

Atypical Populations as Test Cases

2‌2q11.2 Deletion Syndrome

The study of populations with atypical development is particularly relevant for a better understanding of the cognitive mechanisms involved in the development of hallucinations. In that respect, 22q11.2 microdeletion syndrome (22q11DS) represents an interesting model, since the prevalence of psychotic disorders in this rare disease—which affects around 1:2000 live births—has an estimated lifetime prevalence of ~20%.¹⁰³ Subclinical psychotic manifestations, in the form of perceptual disturbances or persecutory ideas, are present in up to 60% of adolescents and young adults.¹⁰⁴ Beyond the risk of psychosis, people affected by 22q11DS are generally characterized by a complex profile of difficulties involving somatic issues (eg, cardiac or oro-pharyngeal malformations), developmental disorders (eg, “borderline” intellectual functioning and intellectual disability in around 50% of cases), speech and language disturbances, increased prevalence of neurodevelopmental conditions (such as autism or ADHD), and/or psychiatric difficulties (eg, increased prevalence of anxiety and mood disorders).¹⁰⁵ A specific advantage of the 22q11DS model lies in the possibility of examining the developmental trajectories of young people years before the emergence of the first psychotic symptoms. However, an important limitation is that the majority of existing findings have focused on psychotic symptoms as a whole and the presence of hallucinations has rarely been considered separately. A notable exception concerns the work of Mancini et al.,¹⁰⁶ which examined thalamo-cortical connectivity in relation to the emergence of AVH in a longitudinal study involving adolescents and adults with 22q11DS. In this study, 22q11DS carriers with AVH exhibited deviant volumetric trajectories of specific thalamic nuclei, as well as an immature pattern of thalamo-cortical connectivity with the auditory cortex and speech processing regions. Interestingly, thalamo-cortical connections were shown to play a central role in filtering and attributing an appropriate salience to environmental information,¹⁰⁷ and impaired sensory gating was found both in the context of idiopathic psychosis—especially in relation to AVH^108,109—and in 22q11DS¹¹⁰. This suggests that sensory gating could play an important role in the development of AVH in this group.

When it comes to psychotic manifestations as a whole, a large body of literature has examined their neurobiological, cognitive, and clinical mechanisms in 22q11DS (for a review, see ref. Biswas and Furniss 111). Specifically, several longitudinal studies have related cognitive decline—and especially change in verbal IQ—to the onset of psychosis.¹¹² Specific cognitive domains have also been linked to the presence and/or emergence of psychotic manifestations, such as social cognition,¹¹³ memory,¹¹⁴ and executive functioning.¹¹⁵ These findings have notably been related to atypical brain maturation in 22q11DS, such as a decrease in hippocampal volume during adolescence that is more pronounced in individuals developing psychotic manifestations.^114,116 Aside from these broad cognitive markers of psychosis, a subset of studies has also highlighted relevant cognitive mechanisms potentially related to the presence and/or emergence of psychotic manifestations. For example, motor agency deficits¹¹⁷ as well as source monitoring^118,119 impairments have been related to psychotic manifestations. Interestingly, several recent findings have also revealed the influence of stressful life events and heightened reactivity to (social) stressors on the emergence of psychotic symptoms,^120,121 suggesting the existence of an “affective pathway” toward psychosis in this population,¹²² which probably interacts with the presence of cognitive alterations, notably in the field of social cognition.

Autism

The study of autism might also prove to be highly relevant for understanding the mechanisms underlying the emergence of psychosis and hallucinations. Autistic adolescents and adults are at high risk of developing psychotic manifestations and full-blown psychosis,¹²³ with two recent meta-analyses showing prevalence estimates up to 10 times higher than in the general population.^124,125 Recent accounts point to shared cognitive and sensory differences in both autism and psychosis, such as altered multisensory integration.¹²⁶

The majority of studies of psychotic manifestations in the context of autism are descriptive. Autistic individuals have been shown to have higher rates of anomalous perceptual experiences,¹²⁷ which can be described as very disturbing and frightening by young people themselves, who are able to differentiate them from sensory issues that are typically encountered in the context of autism (eg, hyper-reactivity to sensory stimulations¹²⁸). Nonetheless, a recent study applying network analysis on momentary experiences collected in daily life using Ecological Momentary Assessment showed strong commonalities between sensory issues and psychotic experiences, suggesting that similar mechanisms are involved in both phenomena.¹²⁹ To date, there are only limited mechanistic explanations to explain the frequent co-occurrence of autism and psychosis. Bayesian inference accounts may be valuable in studying this co-occurrence,¹³⁰ but evidence of a proposed differential use of priors in autism has been inconsistent.

Finally, recent evidence points to the role of adverse childhood experiences and childhood trauma in the emergence of psychosis in the context of autism,¹³¹ also highlighting the role of environmental factors.

Conclusions and Future Directions

A variety of cognitive-developmental mechanisms have been implicated in hallucinations as they appear in both childhood and adulthood. The aim of this article has been to show how engagement with the mainstream developmental psychology literature can be beneficial in understanding the varied experience of hallucinations across the lifecourse. Table 1 sets out some of the key learnings from recent research in this area along with practical implications for further research, clinical assessment, and therapy.

Table 1.

How Key Insights on Cognitive-Developmental Processes in Hallucinations Can Shape Future Research, Clinical Assessment, and Therapy

Key insight	Implications
1. Hallucinatory experiences occurring throughout the lifecourse can be illuminated by studying core cognitive processes in typical development.	Research at this nexus should continue to explore how hallucinations relate to processes such as engaging with imaginary companions,41,42,45,46 attachment changes,62-65 inhibitory control,74,75 social cognition,80,81 language processes,95,99 etc.
2. Integrating adult and childhood models of hallucination will likely require a mix of continuous and discontinuous mechanisms.	Careful longitudinal research is needed fully to understand the significance of childhood hallucinations and how they persist into adulthood, allowing researchers to tease apart which cognitive-developmental mechanisms are specific to childhood (and more precisely in what developmental window) and which contribute to later clinical significance.9,15,16,27
3. Hallucinations can have multiple causes in early life, with only some persisting in a distressing form.	Clinical assessment requires awareness of the complexities of hallucination development, such as the earlier onset reported by non-clinical voice-hearers,1 and the occurrence of non-compliant imaginary companions associated with adverse life events, which may be linked to clinical risk.132
4. A cognitive-developmental approach helps researchers to understand the differential effects of trauma at distinct periods in childhood and adolescence.	Understanding timing of trauma exposure relative to ongoing developmental processes can be valuable in understanding how emerging cognitive processes interact distinctly with neural maturation in different epochs,66,69,70 and can orient clinicians towards the optimal remediation strategies.
5. Developmental conditions such as 22q11DS103,104 and autism123-125 carry an elevated risk of psychosis. Studies assessing overlap of perceptual and cognitive mechanisms between 22q11DS and psychosis,110,113-122 and between autism and psychosis,126,129,130 have shown promise. To date, however, research attention to hallucination occurrence in these groups has been limited.	Hallucinations in 22q11DS and autism should be assessed with reference to the increased risk of psychosis in these neurodevelopmental conditions. Isolated hallucinatory experiences should not be medicated, except for intrusive cases, but a tight and regular follow-up is warranted, combined with prevention strategies (eg, metacognitive training to avoid overinterpretations of hallucinatory experiences) and awareness of psychosis warning signs. Further research should pay particular attention to hallucination phenomenology in neurodevelopmental conditions.

Many of these recent findings continue to support the idea that later psychopathology is best predicted by persistent hallucinatory experiences associated with exposure to a range of environmental and psychopathological risk factors. This suggests that a large number of other transient experiences are based on other, non-pathological developmental mechanisms. A key challenge for future research is to further elucidate which cognitive-developmental mechanisms associate with persistent (and thus clinically more significant) hallucinations and which relate to more transient experiences (see Figures 1 and 2). This in turn raises the question of how typical cognitive-developmental processes, such as the creation of imaginary companions, can act as the “fuel” or content of clinically significant hallucinations, when combined perhaps with more general Bayesian inference or self-monitoring biases.

Figure 1.

Typical developmental trajectory showing emergence of key cognitive-developmental processes. The shaded area represents the developmental epoch in which hallucination-like phenomena are likely to emerge through typical developmental processes.

Figure 2.

Trauma pathway to hallucinations, in which early adversity (eg, exposure to bullying) leads to an extension of the development epoch indicated in Figure 1, and thus persistence of hallucinatory experiences into adolescence and early adulthood.

One task for future research is what could be described as a “reverse engineering” project to map hallucination phenomenology at different time-points to different theories of cognitive development. For example, if hallucinations relate to the development of social cognition, theories of the time course of emergence of different social-cognitive capacities should constrain the developmental time course of hallucinations. Conversely, a line of future research could move in the opposite direction, from the psychopathological towards the mainstream, investigating what the developmental study of hallucinations can tell us about the emergence and specialization of perceptual and cognitive functions in typical development. For example, close analysis of the phenomenology of AVH in childhood could help to constrain theories of language functions such as inner speech. Finally, the interaction between mainstream and psychopathological perspectives may help future researchers to understand better the changing interfunctional relations between processes such as social cognition and language over the transition to adulthood.

Conflicts of Interest

None declared.