Schizophrenia Bulletin vol. 38 no. 4 pp.

683–692, 2012
doi:10.1093/schbul/sbs045
Schizophrenia Bulletin
Advance Access publication on March 23, 2012
doi:10.1093/schbul/sbs045

Auditory Hallucinations in Schizophrenia and Nonschizophrenia Populations: A

Review and Integrated Model of Cognitive Mechanisms

Flavie Waters1,2,*, Paul Allen3, André Aleman4,5, Charles Fernyhough6, Todd S. Woodward7,8, Johanna C. Badcock9,10,

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
Emma Barkus11, Louise Johns12, Filippo Varese13, Mahesh Menon14, Ans Vercammen15, and Frank Larøi16
1
Centre for Clinical Research in Neuropsychiatry, Graylands Hospital, North Metro Area Mental Health Perth, Australia; 2School of
Psychiatry and Clinical Neurosciences, The University of Western Australia, Perth, Australia; 3Department of Psychosis Studies, Institute of
Psychiatry, King’s College, London, UK; 4University Medical Center Groningen, Groningen, The Netherlands; 5Department of Psychology,
University of Groningen, Groningen, The Netherlands; 6Department of Psychology, Durham University, Durham, UK; 7Department of
Psychiatry, University of British Columbia, Vancouver, Canada; 8Department of Research, BC Mental Health and Addiction Research
Institute, Vancouver, Canada; 9Centre for Clinical Research in Neuropsychiatry, Graylands Hospital, Perth, Australia; 10School of
Psychiatry and Clinical Neurosciences, The University of Western Australia, Perth, Australia; 11School of psychology, University of
Wollongong, New South Wales, Australia; 12Department of Psychology, Institute of Psychiatry, King’s College, London UK; 13School of
Psychology, Bangor University, North Wales, UK; 14Schizophrenia Program and PET Centre, Centre for Addiction and Mental Health,
Department of Psychiatry, University of Toronto, Toronto, Canada; 15School of Psychiatry, University of New South Wales, Sydney,
Australia; 16Department of Psychology, University of Liège, Liège, Belgium
*To whom correspondence should be addressed; Centre for Clinical Research in Neuropsychiatry, Private Mail Bag No 1. Claremont, Perth
WA 6910, Australia; tel: 61-8-9347-6429, fax: 61-8-9384-5128, e-mail: [email protected]

While the majority of cognitive studies on auditory hallu- explaining a range of phenomenological characteristics
cinations (AHs) have been conducted in schizophrenia of AH across a spectrum of disorders.
(SZ), an increasing number of researchers are turning their
attention to different clinical and nonclinical populations, Key words: symptoms/psychosis/hallucinosis/research/
often using SZ findings as a model for research. Recent domain/criteria/auditory network/model
advances derived from SZ studies can therefore be utilized
to make substantial progress on AH research in other
groups. The objectives of this article were to (1) present Introduction
an up-to-date review regarding the cognitive mechanisms
Auditory hallucinations (AHs) are auditory experiences
of AHs in SZ, (2) review findings from cognitive research
that occur in the absence of a corresponding external stim-
conducted in other clinical and nonclinical groups, and (3)
ulation and which resemble a veridical perception. Strongly
integrate these recent findings into a cohesive framework.
identified with psychotic disorders such as schizophrenia
First, SZ studies show that the cognitive underpinnings of
(SZ), AHs have traditionally been investigated in SZ pop-
AHs include self-source-monitoring deficits and executive ulations. A recent shift in AH research, however, has led
and inhibitory control dysfunctions as well as distortions in to a strategic focus on other clinical and nonclinical groups
top-down mechanisms, perceptual and linguistic processes, on the basis of observations that hallucinations and
and emotional factors. Second, consistent with SZ studies, hallucination-like experiences are common in several psy-
findings in other population groups point to the role of top- chiatric and also nonpsychiatric populations (ie, they are
down processing, abnormalities in executive inhibition, and ‘‘transdiagnostic’’). That such experiences do not map
negative emotions. Finally, we put forward an integrated closely onto specific disorders has recently prompted the
model of AHs that incorporates the above findings. We sug- National Institute of Mental Health–driven Research Do-
gest that AHs arise from an interaction between abnormal main Criteria to devise new ways of classifying psychopa-
neural activation patterns that produce salient auditory sig- thology based on symptom dimensions.1 This approach
nals and top-down mechanisms that include signal detection lends itself to investigations of symptoms such as AH,
errors, executive and inhibition deficits, a tapestry of which comprise multiple phenomenological features. While
expectations and memories, and state characteristics acknowledging that the phenomenological features differ
that influence how these experiences are interpreted. Emo- somewhat between groups, transdiagnostic studies may
tional factors play a particular prominent role at all levels shed some light on the mechanisms that are specific to
of this hierarchy. Our model is distinctively powerful in AH independently of other symptoms associated with SZ.

© The
The Author
Author 2012.
2012. Published
Published by
by Oxford
Oxford University
University Press
Press on
on behalf
behalf of
of the
the Maryland
Maryland Psychiatric
Psychiatric Research
Research Center.
Center. All
All rights
rights reserved.
reserved.
For permissions,
For permissions, please
please email:
email: [email protected]
[email protected].

683
1
F.
F. Waters
Waters et
et al.
al.

Cognitive methods have been widely used since early impairments in ‘‘self-monitoring’’ processes, whose role
studies of AH. The power of cognitive approaches lies is to predict the sensory consequences of one’s own
in their ability to provide plausible and intuitive explan- actions via forward modelling/efference copy mecha-
ations for subjective (ie, nonobservable) symptoms and nisms.3,4 Such difficulties may cause mental events (par-
to generate predictions regarding associated neural ticularly inner speech) to become isolated from predictive
mechanisms that are testable using neuroscientific meth- mechanisms and misinterpreted as originating from an
odologies such as magnetic resonance imaging and elec- external source. Analogous processes are thought to exist
troencephalographs. Increasingly, researchers are for language (including inner speech), actions, and
investigating AH in different population groups, often thoughts, so that a failure in this system would lead to
applying cognitive findings and behavioral paradigms de- broad difficulties in self-recognition. Another model pro-

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
veloped for SZ research. An up-to-date review regarding poses dysfunctions in ‘‘source-monitoring,’’ which refer
the current knowledge on cognitive mechanisms of AH is to judgment processes that are used to make internal/ex-
therefore needed so that advances can be made on the ternal discriminations. These typically engage memory
basis of the existing body of evidence. and other decision-making processes that retrieve and
The first aim of this article is to present the most recent evaluate memory records in order to form a cohesive rep-
theoretical developments in cognitive research pertaining resentation of an experience. It has been proposed that
to AH. While several high-quality reviews of cognitive such source-monitoring disturbance results in an incom-
mechanisms in AH have been published in the past, plete representation of mental events and consequently
this is the first time that different phenomenological fea- a failure to identify their origins.5
tures of AH have been so comprehensively linked to their Theoretical aspects of these models have been criticized
theoretical model in one article. The second aim is to pro- on the grounds that they do not provide a comprehensive
vide an overview of findings from cognitive investigations account of the phenomenological diversity of AH.6 The
of AH in non-SZ populations and compare the findings cognitive tasks used to assess self- or source monitoring
with the SZ literature. The third aim of this article is to have also been criticized on methodological grounds.
amalgamate these findings into a new theoretical model Typically, these require participants to monitor volun-
of AH that can provide explanations for a range of tary movements online or to identify whether items in
phenomenological features in clinical and nonclinical memory originated from the self or another (‘‘agency’’
populations. tasks). However, errors may be the result of cognitive
processes other than self- or source monitoring, such
as response biases or difficulties in appraising ambiguous
Cognitive Explanations for the Phenomenological Features stimuli that could overshadow patients’ ability to recog-
of AH in SZ nize their own actions or mental events.7
Approximately 70% of people with a diagnosis of SZ re- Despite these criticisms, evidence supporting the link
port AH. It is widely accepted that AH in SZ are multi- between AH and self- and source-monitoring deficits is
dimensional and heterogenous.2 In keeping with this strong, and few studies have failed to replicate these find-
dimensional view of AH, separate explanations have ings. A recent meta-analysis showed that self- and source-
been used to account for the different phenomenological monitoring impairments were consistently reported
features of AH, each potentially representing a particular across a range of paradigms, interstimulus intervals,
circuitry of brain structures and functions. We focus on 4 and modalities in patients with SZ and particularly those
features that have been the most intensively researched: with AH.8 Such dysfunctions were thought to occur in the
(1) the failure of self-recognition, whereby the experience earlier, rather than later, stages of information process-
is perceived as alien and separate from one’s own mental ing, although such early deficits would undoubtedly im-
processes; (2) reduced sense of control over the onset, pact on higher-order processes. One problem with this
content, and frequency of AH; (3) the perceptual quality literature is that monitoring deficits also occur with other
of AH; and (4) the contribution of emotions. symptoms (eg, Frith et al4). While a possibility exists that
the cognitive underpinnings of AH may be shared across
some symptoms, it also suggests that such impairments
Failure of Self-recognition (alienation) are not sufficient for AH to occur.
One core feature of AH in SZ is that they are experienced A related approach to recognition problems has been
as somewhat separate from one’s own mental processes: provided by Signal Detection Theory (SDT) models,
the core experience is that the hallucinated percept is which state that all information recognition takes place
lacking in ‘‘self’’ attributes. In addition, and perhaps sec- in the presence of some uncertainty and that processing
ondarily, the origin of these events tends to be incorrectly relies on both pattern recognition (perceptual sensitivity)
attributed to an external agent. and biases in responding.9 This approach has been used
Several explanations have been proposed for such fail- to test the source-monitoring hypothesis, and the theory
ures of self-recognition. One explanation centers around suggests that hallucinating individuals may have relatively

684
2
 Cognitive
Cognitive Mechanisms
Mechanisms of
of Hallucinations
Hallucinations

unimpaired perceptual sensitivity but show lax decision Sense of Control

criteria about accepting an signal as real and biases in From the patients’ perspective, hallucinations are often
responding that misattribute events to a nonself source.9 described as unintentional and intrusive.14,15 This
A strength of this model is the differentiation between sig- reduced sense of control may be used by individuals to
nal detection and response biases, consistent with the idea differentiate hallucinated voices from one’s own verbal
that attributions must be differentiated from earlier pro- thoughts,16 although much variability exists in the degree
cessing stages. It is also compatible with the above models to which AH (and indeed verbal thoughts) are perceived
because self-monitoring problems might contribute to in- as controllable.17 Cognitive explanations have thus incor-
correct decisions about the source of information, partic- porated the idea that AH involve ‘‘a failure to control the
ularly for predictive mechanisms linked to self-generated contents of consciousness,’’18 which is generally assumed
actions. Recent versions of this SDT model10 posit that

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
to reflect a breakdown in one or more of the executive
perceptual hypervigilance (perhaps linked to anxiety) functions which control and regulate thought and action.
enhances biases in responding and thus produces a higher The differentiation of executive functions into separable
likelihood of errors in cognitive processing and of accept- (though correlated) components has provided a useful
ing a signal as real. The desire to reduce uncertainty under framework for considering reduced sense of control in
threat also leads to increased detection of ambiguous sig- AH.
nals, a reduction in auditory threshold, and thus halluci-
natory experiences.
As above, one problem with this theory is that SDT Inhibition. Early studies of AH tended to apply the term
impairments are not specific to AH (eg, Harvey11). In ad- ‘‘inhibition’’ fairly loosely and reported negative findings
dition, the ability to make rapid and overconfident judg- on measures of negative priming and interference (eg,
ments about the nature of perceptions is a processing Peters et al19). Since that time, there has been accumulat-
style that is commonly linked to delusions, pointing to ing evidence demonstrating a link between AH and a par-
the possibility that such underpinnings of AH may be re- ticular type of suppression, termed ‘‘intentional cognitive
lated to cognitive processes common to many positive inhibition’’ (eg, Waters et al5). These studies draw on ex-
symptoms. tensive evidence from cognitive neuroscience that inhibi-
While the above models provide explanations for ver- tion involves a family of processes, each with its own
bal and nonverbal types of AH, inner speech theories characteristic operating mechanisms. For example, dif-
have focused on providing accounts of verbal AH. ferent aspects of inhibitory processing may be differenti-
They suggest that information regarding the misattribu- ated, particularly between cognitive vs behavioral
tion of inner speech might be gained by comparing the inhibition, intentional vs automatic inhibition, and inhi-
phenomenology of inner speech in hallucinating individ- bition vs interference control. Studies that have used this
uals with ‘‘normal’’ inner speech.12 One such model model together with tasks demanding the volitional sup-
builds on the observed distinction between ‘‘expanded’’ pression of memory events and irrelevant memories
(ie, possessing an overt dialogic structure) and ‘‘con- tested the prediction that AH in SZ involve a deficit in
densed’’ (ie, abbreviated) forms of inner speech and pro- intentional cognitive inhibition.5 The degree of inhibitory
poses that AH occur during the transition from impairment was significantly correlated with the severity
condensed to expanded inner dialog, particularly during of AH. Moreover, the association was specific to AH be-
periods of high cognitive load or stress.12 Although exist- cause the number of inhibitory failures was not associ-
ing data on the phenomenology of inner speech in hallu- ated with other symptoms. Stated differently,
cinating individuals cannot yet fully address this model a particular form of prefrontal inhibitory control may al-
(see Langdon et al13), such phenomenology-based models low auditory signals to be relatively functionally auton-
are useful in pointing to the evolution and transformation omous and difficult to control effectively.
of neural information into increasingly differentiated sig-
nals that are subject to modification by factors such as Attention and Working Memory Updating. There has
emotions. been a long-standing interest in the contribution of atten-
In summary, despite different explanations for self- tion and working memory processes to AH,14 although
recognition deficits in AH, studies generally converge early studies concluded that tasks assessing the phono-
on the finding that AH are linked to monitoring deficits logical store and loops were unrelated to positive symp-
and misattributions. One key difficulty with this litera- toms (eg, David and Lucas20). Nonetheless, recent
ture is that cognitive deficits in self-/source monitoring functional imaging data show that patients with AH ex-
and SDT have been linked to other symptoms of SZ. hibit reduced activity in verbal working memory circuits,
Thus, while these deficits may still play an important albeit in the absence of deficits in working memory per-
role for AH formation and maintenance, it appears formance.21 That verbal working memory neural circuits
that such cognitive mechanisms may underpin a range are associated with AH but not behavioral performance,
of other psychotic symptoms. perhaps reflects a broader pattern of language processing

685
3
F.
F. Waters
Waters et
et al.
al.

deficits in AH.22 While cognitive evidence for attention these factors in perception has been suggested to activate
and working memory updating remains inconclusive, percepts in the absence of external stimulation. Several the-
the role of attention is clearly relevant in early auditory ories posit that a blend of distorted input from bottom-up
sensory detection mechanisms. Though not directly sensory information and aberrant top-down factors cause
linked to the sense of control, attentional processes might AH.24 Recent models25 suggest that impairments in hier-
operate in AH through the determination of resource al- archical perceptual processing, underpinned by structural
located toward processing and correction of errors dur- and functional neurological abnormalities, are a core ab-
ing information processing.10 normality in SZ. The model is also based on the fact that
processing of sensory information from the environment is
Set-Shifting. Hugdahl23 has argued that a proper un- not passive, as it is quickly combined with stored memory

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
derstanding of the neurocognitive basis of AH demands representations, other sensory experiences, and top-down
inclusion of ‘‘the ability to shift attention away from the expectations. What is experienced is thus a cascade of pro-
voices’’. He proposes that AH involve both bottom-up cesses that produces a subjective perception of reality.
internal activation of left-hemisphere speech perception Over time, this inaccurate processing leads to ‘‘hard wir-
areas and dysfunctional top-down executive control. ing’’ of networks which produce a distorted subjective ex-
On a dichotic listening paradigm, which lateralizes stim- perience of sensory input and reality distortions.
ulus input, patients with SZ and frequent AH failed to Empirical evidence for increased top-down processing
demonstrate the expected right ear advantage, indicative in SZ patients with ‘‘verbal’’ AH is mostly derived from
of a functional deficit in the left perisylvian region. SDT studies with tasks using verbal stimulus detection in
Reduced responsiveness to right ear stimulation is noisy circumstances,14,26 suggestibility and expectan-
thought to arise because left-hemisphere language cies,27,28 and semantic expectations.26 These results sup-
regions are already engaged in processing; ie, patients port the conclusion that aberrant top-down processing,
are already ‘‘tuned in’’ to the voices. In addition, patients particularly in the form of strong semantic expectations,
with AH exhibited difficulties in shifting attentional focus may contribute to the experience of AH, at least those of
to the opposite ear. The implication is that AH involve a verbal nature. In addition, the notion of suggestibility
a difficulty in the modulation of attention and in achiev- provides an interesting explanation for the reporting of
ing top-down executive control of voices and that the a phenomenological experience in the absence of a clear
inability to shift might be a negative consequence of external signal. If so, individuals with AH would be more
increased attentional focus on hallucinated voices. amenable to conditioning and effects of implicit or ex-
In summary, executive and inhibitory control dysfunc- plicit suggestion.29 However, it does not explain results
tions have been linked to AH. Intentional inhibition def- in their entirety given that differences are reported in
icits have also been linked to a reduced sense of control decision-making biases rather than sensitivity to detect
associated with AH. Theoretically, deficits in intentional biases.9
inhibition can cause mental events to be experienced as Consistent with an auditory sensory-conditioning
unintended and intrusive.5 A lack of anticipatory repre- model, one recent study indicated that hallucinating
sentation would also contribute to this reduced sense of patients acquire auditory conditioned hallucinations
control. Deficits in attention and set shifting may play an more quickly than do nonhallucinating controls, and
altogether different role, by determining expectations and these AH are more resistant to extinction.30 This effect
resources to be allocated to these unintended auditory may determine whether subjective experiences are
signals and by limiting the ability to reallocate and trans- reported as hallucinations or real percepts.31
fer attention to other adaptive information.
Emotional Quality
Perceptual Quality A range of positive and negative emotions are associated
A cardinal phenomenological feature of hallucinations is with AH, and studies now view emotions as intrinsically
their perceptual quality. A person who hallucinates linked to the content, frequency, and beliefs about AH
‘‘hears’’ sounds and voices that can be described in terms (eg, ref. 32,33). Evidence comes from multiple sources.
of parameters of loudness, pitch, and clarity. These dis- The affective consequences of AH in psychiatric patients
tinctive characteristics of AH have been explained in tend to be negative (though not always), and common
terms of top-down perceptual processing. feelings reported by patients with AH include anxiety
Top-down factors reflect the influence of internal fac- and depression.16,32 Furthermore, emotional responses
tors and stored representations on perception, which to voices are often related to beliefs and appraisals of
include prior knowledge and memories, perceptual expect- the voice rather than to the experience per se.33 For ex-
ations, and mental imagery. Given that all acts of percep- ample, beliefs that voices are malevolent tend to be asso-
tion are composed of an interaction between bottom-up ciated with negative emotions (eg, fear and depression),
input and top-down constraints, an imbalance between whereas beliefs that voices are benevolent tend to be

686
4
 Cognitive
Cognitive Mechanisms
Mechanisms of
of Hallucinations
Hallucinations

associated with positive emotions (eg, enthusiasm and Model of AH’’ section). Thus, studies must report in
respect). Increased distress in voice hearers is also related a transparent manner on the type and range of clinical
to a lack of ability to control the AH, perhaps through the experiences seen in their samples so that the specificity
influence of rumination processes on the frequency of the of findings to AH may be examined. Similarly, duration
experience.34 Finally, a host of evidence indicates the im- of illness and clinical status are important variables that
portant role of emotions as a trigger of AH, as a maintain- will influence test results. Finally, a number of other con-
ing factor,32 and in determining the need for care.35 founds can limit the interpretability of cognitive findings.
Further support for the role of emotions is derived These include the contributing effects of variability in in-
from links between dissociation and hearing voices.36 telligence, attention and working memory, effects of med-
Many studies have also reported a specific association be- ication, psychiatric comorbidities, and possible social

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
tween hallucinations and childhood trauma, prompting and psychological results of the stigma and negative expe-
recent suggestions of a ‘‘2 hit’’ model of AH in which riences. While these variables cannot always be con-
a combination of cognitive deficit and trauma lead to trolled, researchers should record these in detail and
a high risk of hallucinations in adulthood.37 However, take into consideration that multiple factors may impact
the presence of AH in nonhelp seeking community sam- on performance.
ples suggests that the experience itself is not always prob-
lematic, with evidence for a more positive emotional
Cognitive Investigations of AH in Populations Other Than
valence of AH experiences in those not seeking psychiatric
SZ
help.38
When compared with the number of cognitive studies in
SZ, fewer studies have been conducted in other popula-
Methodological Issues in Cognitive Studies of AH in SZ tion groups. Yet, the study of AH in non-SZ populations
A number of methodological lessons can be gained from is a particularly useful methodological strategy given its
a long history of AH research in SZ. First, with some potential for understanding the mechanisms of AH inde-
exceptions, cognitive tools can be limited in their ability pendently of other symptoms associated with SZ. Studies
to engage fully the theoretical constructs of interest, em- with non-SZ populations can therefore elucidate and test
phasizing the importance of striving for high construct hypotheses about the causes of AH without interference
validity when choosing tasks. Cross-disciplinary work from other symptom profiles. In this section, we review
might help to identify knowledge gaps and to bridge available cognitive findings derived from studies in different
those gaps through the integration of different method- groups.
ologies. Increased collaborations between cognitive (and The most intensively studied population includes
linguistic) sciences, electrophysiological approaches, and healthy individuals who at times experience AH. The es-
neuroimaging will benefit the development of theories. timated prevalence of AH in this group is approximately
For example, different tools and approaches maybe 15%. Phenomenologically, AH in these nonclinical pop-
used to test hypotheses derived from one discipline. ulations may be different to those in SZ, mostly in the
This rests upon collaborative efforts and possibly the domain of emotional content of hallucinations and the
pooling of data between centers. amount of control over AH.38 By contrast, nonself rec-
Second, the selection and grouping of participants is ognition, perceived location of voices, number of voices,
critical. An adequate design might compare patients loudness, and personification do not tend to differentiate
with the same diagnosis who differ on the presence/ between psychotic and healthy individuals. Cognitive
absence of state-related AH, although a group of patients studies have often focused on individuals who score
who have never experienced AH would be necessary to high on scales such as the revised Launay-Slade Halluci-
tease out state/trait factors. Few longitudinal studies of nation Scale (LSHS-R). Such individuals are thought to
AH exist, although these are useful for observing the fluc- be ‘‘prone’’ to AH, although it is important to note that
tuating nature of symptoms and underlying mechanisms the LSHS assesses a range of psychological constructs,
in the same participants. Issues of symptom assessment including different hallucination modalities, intrusive
are also of decisive importance. The presence and severity thoughts, and vivid daydreams. Although not strictly
of different AH features must be assessed using well- assessing AH, such studies can be potentially informative
validated scales2 and clearly detailed in scientific articles. given that potential confounds associated with psychiat-
In addition to AH symptoms, the assessment and report- ric research, such as medication and chronicity, can be
ing of other symptoms are important. Many studies of avoided. Cognitive investigations show that their pattern
AH, for instance have shown that cognitive deficits are of performance tends to be similar to that of SZ groups,
shared across other symptoms, often delusions, or passiv- albeit at attenuated levels. Consistent with the observa-
ity symptoms. The role of insight has received little atten- tion that self-recognition problems and complex percep-
tion, although poor insight might play an important role tual quality are phenomenological features that are
in some aspects of AH phenomenology (see ‘‘Cognitive present in both clinical and nonclinical samples, studies

687
5
F.
F. Waters
Waters et
et al.
al.

in healthy people demonstrate spontaneous biases and been reported in individuals with epilepsy presenting
misattribution39 as well as excessive top-down process- with a history of frequent AH and to (visual) hallucina-
ing.40 By contrast, findings on self-monitoring tasks tions in eye disease.51 Thus, there may be general, ie,
have been rather mixed, with some studies showing cross-modal mechanisms that underlie hallucinations re-
self-recognition difficulties,41 and others finding no rela- gardless of the diagnostic category.
tion between hallucination-proneness (incorporating all Together, the evidence points to shared impairment in
modalities) and self-recognition for actions42. inhibitory functions, emotional problems, and top-down
While the reduced sense of control is a less salient fea- mechanisms across different population groups. While
ture in nonclinical individuals compared with SZ groups, nonself recognition is a key clinical feature of AH across
studies still show broad dysfunctions on tasks of inten- all groups, evidence regarding the role of self-/source
tional inhibition.43 Other support for executive dysfunc-

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
monitoring is inconsistent, although this might be a reflec-
tions in this population comes from links between AH tion of the lack of investigations in this area. Altogether,
proneness and intrusive thoughts, ruminations, and these findings mirror the pattern of cognitive perfor-
attempts at thought suppression.44 Certain aspects of mance demonstrated in SZ. Conclusions from these
metacognitive style such as beliefs about the uncontrol- observations are limited, however, given publication
lability of thoughts have also been reported,44 although biases that ensure that only positive findings are reported
these are thought to be largely mediated by the effects of in the literature. While similarities in deficits are empha-
comorbid symptoms.45 sized, much less is known regarding how cognitive
Phenomenological studies of AH in nonclinical groups profiles differ between groups.
show that the emotional quality of AH is less negative
and intrusive than in SZ, yet findings show more negative
emotions, and dysfunctional emotional regulation strat- Cognitive Model of AH
egies, when compared with healthy individuals without Here, we combine evidence from the above into a cohesive
these symptoms.46 Executive dysfunctions and emotion model. We build on a model that was first proposed by
processing difficulties perhaps reflect the potential vul- Frith and Dolan52 and subsequently elaborated by
nerability of these individuals to experiencing psychosis. Aleman et al24 and Hugdahl.23 This model (see figure
Another approach has been to define an AH analog 1) comprises multidimensional elements, but it focuses
group in terms of susceptibility to hypnagogic and on the fact that AHs are essentially perceptions. Like
hypnopompic hallucinations. Studies have showed that all perceptions, AHs arise through an interaction between
these may also be linked to executive and inhibitory information arising from neural activations and top-
dysfunctions.47 down activity. Cognitive, imaging, electrophysiological,
Despite the informative nature of transdiagnostic stud- and phenomenological findings associated with AH
ies, there is a paucity of cognitive investigations con- can complement and support this explanation.
ducted in other clinical populations. Approximately
15% of individuals with bipolar disorder report AH, al-
though these have been rarely described phenomenolog- Model of AH in SZ
ically. Studies show that AHs in this group are linked to There are 2 types of functional brain systems that are
difficulties on a task of self-monitoring, as demonstrated needed for this interaction. One (the source) involves sa-
on a voice-distortion paradigm.48 lient auditory stimuli that provide the basic signal for
Studies of borderline personality disorder show that AH. This is thought to arise from hyperactivation in
AHs have an estimated prevalence of 20%–50%, with functional networks involving the auditory cortex that
phenomenological features that are similar to those in generate aberrant auditory signals, possibly due to a de-
SZ. Cognitive studies in this group have demonstrated viant trigger of activations in language-related areas re-
links between AH and deficits in executive functions (par- sponsible for AH.22,53 Anomalous activations might be
ticularly inhibition).43 Negative emotions might also be determined by environmental factors and/or internal
important as an etiological factor of AH in borderline (eg, emotional) conditions. One consequence of such ab-
personality disorder, with studies reporting increased in- normal neural activation includes auditory signals that
cidence of childhood trauma and emotional abuse (inde- exceed perceptual threshold, thus causing unexpectedly
pendent of the presence of paranoid delusions) and a loss intense (hypersalient) sensory information. This may
of emotional regulation, eg, Kingdon et al.49 go some way toward explaining source-monitoring diffi-
It is interesting to note that visual hallucinations in culties because it would bias such internal material to-
Parkinson’s disease (PD) are linked to both source- ward being perceived as alien and separate from one’s
monitoring and inhibitory control difficulties.50 As in internal mental processes and as arising from external
SZ, poor source-monitoring and inhibitory control influences. Specific forms of auditory signals (eg, forms
thus appear to be a risk factor for the presence of hallu- of inner speech, intrusive memories) may be particularly
cinations in PD. Deficits in executive control have also more likely to be converted into AH and conceivably

688
6
 Cognitive
Cognitive Mechanisms
Mechanisms of
of Hallucinations
Hallucinations

Deficits in signal Intentional Prior history, mental Contributions from insight and
detection = accepting inhibition deficits imagery, fantasy, past delusional beliefs:
signal as real memories expectations/hypervigilance

‘First hit’ + Emotion = hypervigilance particularly for negative trauma, depression, anxiety Depression, anxiety
Traumatic insult Modulation and biases information negative memories Paranoia, fear

Abnormal activation of auditory

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
network causing hypersalient
auditory signals

Verbal/non verbal, and Errors in processing self- and Reduced sense of control over A highly personalised auditory Beliefs about
inner/outer distinction determined non-self information. onset and frequency of the experience that is perceived to hallucinations
by differences in neural activity experience be real
along the ventral ‘what’ and
dorsal ‘where’ pathways

Fig. 1. Temporal unfolding of auditory hallucinations (AHs) in clinical and nonclinical populations (the boxes shaded in gray represent
processing that maybe more characteristic of schizophrenia and that differentiate clinical and nonclinical AH): In this model, AH arise from
an interaction between (a) signals arising from overactivation of auditory brain neural activity and (b) a range of top-down mechanisms that
produce a highly complex and multidimensional experience. These top-down mechanisms include: (1) deficits in signal detection that lead to
errors in processing; (2) intentional inhibition deficits that contribute to a diminished sense of control over this perceptual experience;
(3) a background of expectations, imagery, and memories that provide information that is personally relevant; (4) contributions from lack of
insight and delusional beliefs that provide a set of beliefs about AH; and (5) the contribution of emotions that impacts on all aspects of
processing and that ensures that emotional material is processed over neutral information. This model can be used to explain variations in
phenomenological features (bottom row, dotted lines), so that the severity or location of the cognitive deficits determine individual differences
in the extent to which AH features are present.

account for some of the verbal phenomenological prop- Finally, the meaning of AH is determined by state and
erties of the AH. trait characteristics, influencing how these experiences are
The other processes involve top-down mechanisms interpreted. In the case of SZ, the presence of reduced
that influence the form, content, and meaning of AH. insight, delusional beliefs, negative schemas/beliefs about
Different modes of attention, cognitive control capacity, oneself (e.g. low self-esteem), beliefs about the world, and
prior knowledge/experience, and emotional processes ex- negative affect all combine to produce a complex and elab-
ert influence over form and content. The sequence of pro- orate system of beliefs. For example, the perception may be
cesses might be as follows: First, deficits in signal seen as a plot from the Central Intelligence Agency (CIA) or
detection produce increased detection of ambiguous or as a message about the need to save the world.
salient signals and increased likelihood of accepting the Emotions play a particularly prominent role at all lev-
signal as real and meaningful. Second, such information els of this model (source, form, content, and meaning),
fails to be suppressed by faulty intentional inhibition also perhaps by providing the first traumatic insult
mechanisms and becomes functionally autonomous. (hit) in this ontogeny.37 Emotional events linked to
This would contribute to the failure to contain and con- trauma, dissociations, and other intense negative emo-
trol effectively the onset and frequency of these auditory tions may influence the source of AH by increasing the
signals. Over time, expectations and hypervigilance rate of firing of neural activation and aberrant auditory
would increase the likelihood of such experiences being signals. They are also likely to shape the form and content
repeated (creating a sort of ‘‘cognitive cue’’), leading of signals, due to an ‘‘automatic’’ prioritizing of emo-
to increased biases and a reduction in threshold in accept- tional processing, which ensures that emotional and per-
ing the signal as being real. The content of AH may be sonally salient material is preferentially processed over
determined by factors such as perceptual expectations, neutral information. This would produce biases toward
mental imagery, and prior experience/knowledge (eg, negative information, hypervigilance, and negative sche-
memories) that shape a perception of reality that is idi- mas that will further enhance the processing and memory
osyncratic and highly personalized. As such, voices of recall of affective material. Traumatic life-events also
family members and radio personalities, the voice of produce intrusive memories that will impact on the fre-
God, and sounds of dogs barking, can be recognized. quency of the experience and perception about

689
7
F.
F. Waters
Waters et
et al.
al.

uncontrollability. Finally, over time, this would influence spontaneous auditory neural signals and reduced neural
more broadly aspects such as beliefs and meaning attrib- suppression during covert speech in these groups. Here,
uted to AH (omnipotence etc.). we propose that, similarly to processes involved in SZ and
We noted above that space limitations prevent us from consistent with evidence to date, top-down mechanisms
reviewing other relevant phenomenological features such comprising a mix of deficits in error processing, and
as variability in verbal vs nonverbal content and the in- faulty intentional inhibition mechanisms would lead to
ner-outer space distinction. While little evidence exists re- increased attention of these aberrant perceptual signals
garding cognitive underpinnings of these AH features, and a failure to suppress such signals.
neuroimaging evidence points to different pathways Other top-down processes, however, may be the key fac-
that can provide explanations for variations in such fea- tors differentiating clinical from nonclinical hallucina-

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
tures. Verbal and nonverbal AH may stem from abnor- tions. Emotional factors may be one such differentiating
malities at different hierarchical levels along the ventral factor. The timing, or severity, of trauma and negative
‘‘what’’ neural pathway, whose role is to process sound prior experiences may occur outside a critical window,
identity in the temporal cortices and inferior frontal cor- leading to a different AH experience with more positive
tex. By contrast, inner-outer space localization may be emotional valence, and reduced hypervigilance and
due to differences in activation in the dorsal ‘‘where’’ neu- help-seeking behavior in nonpsychotic individuals. Per-
ral pathway, which projects along the planum temporale, sonal traits (eg, presence of insight) in healthy populations
frontal, and parietal cortical areas. In sum, differences in would also influence how these experiences are perceived,
neural activity provide the functional basis for verbal and such that these experiences are interpreted in the context of
nonverbal AH and for inner-outer distinctions, which a benign rationale, with an absence of an ‘‘active’’ search
therefore may provide the basic auditory material (the for a meaning as is the case in psychotic individuals.
‘‘source’’) for AH. Top-down processes, as explained Clearly, a number of issues require clarification. For
above, would further ‘‘shape’’ this signal into a perceived example, we must determine the origin of auditory sig-
reality that is personally relevant. As an example, neural nals, and the processes by which these are activated, in
activity in the planum temporale with projection to the particular, which cognitive or environmental cues con-
inferior parietal cortex may contribute to external sound tribute to spontaneous activation, and under what con-
localization, while increased activation in premotor areas ditions. Further differentiation is also needed in the
might engage ‘‘voices’’ rather than nonverbal sounds. A processes underpinning AH that distinguish psychotic,
person showing hypervigilance for a stimulus that is emo- nonpsychotic clinical, and nonclinical individuals. Un-
tional and self-relevant would be likely to perceive this fortunately, insufficient evidence currently exists regard-
signal as a critical verbal remark, ie, located outside ing the phenomenological characteristics and cognitive
the head. and biological underpinnings of AH in different condi-
In summary, we propose that AH arise through an in- tions. Emerging evidence will lead to greater understand-
teraction between hypersalient auditory signals (the ing and clearer predictions regarding the processes
source) and top-down mechanisms comprising different underlying AH in different groups.
modes of error-processing, cognitive control, prior knowl-
edge/experience, that govern the form and content of AH,
Concluding Comments
together with an influence of state characteristics (insight,
belief systems, etc.) that determine the meaning. Emo- Altogether, our knowledge of AH is slowly accumulating,
tional processing plays a prominent role, with an initial and we are now in a position to provide increasingly com-
traumatic insult creating a vulnerability for experiencing prehensive models of AH that can incorporate the broad
psychosis, and impacting at all levels of processing phenomenological variations. The above is part of a pro-
in this hierarchy. Finally, phenomenological variations cess of building a research foundation that may, in the
may be explained by individual differences in severity of short term, inform novel practical and theoretical
deficits and localization of neural activity. Clearly, there approaches toward investigations of AH. It is hoped
are many subtypes of AH2 that require different combina- that this review on the current state of knowledge will
tions of processes.54 help researchers in selecting and incorporating the
most appropriate research focus for their transdiagnostic
research on AH. In the longer term, such approaches may
Cognitive Model of AH in Non-SZ Populations facilitate the classification of patients in research and
The model proposed above can be used to explain varia- guide the best approaches for clinical diagnosis and treat-
tions in AH features and cognitive findings in popula- ment. In the meantime, investigations using novel task
tions other than SZ (figure 1). Essentially, an auditory design, careful phenomenological assessment, and
signal is an obligatory element for hallucinations to oc- cross-disciplinary research protocols are urgently needed
cur. Though this has not been yet pursued in non-SZ pop- to improve the power of cognitive approaches to assess
ulations, it should be possible to examine for increased mechanisms underlying AH. It is clear that an important

690
8
 Cognitive
Cognitive Mechanisms
Mechanisms of
of Hallucinations
Hallucinations

avenue for research is to cross borders between cognition, 12. Fernyhough C. Alien voices and inner dialogue: towards a de-
phenomenology, imaging, and neurobiological tools. velopmental account of auditory verbal hallucinations. New
Such a framework that interfaces directly with different Ideas Psychol. 2004;22:49–68.
branches of science will facilitate progress in understand- 13. Langdon R, Jones SR, Connaughton E, Fernyhough C. The
phenomenology of inner speech: comparison of schizophrenia
ing underlying etiology and maintenance of AH and in
patients with auditory verbal hallucinations and healthy con-
developing new therapeutic interventions. trols. Psychol Med. 2009;39:655–663.
14. Hoffman RE, Rapaport J, Mazure CM, Quinlan DM. Selec-
tive speech perception alterations in schizophrenic patients
Funding reporting hallucinated ‘‘voices’’. Am J Psychiatry. 1999;
FW is funded by a National Health and Medical Re- 156:393–399.

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
search Council grant. 15. Morrison AP, Baker CA. Intrusive thoughts and auditory
hallucinations: a comparative study of intrusions in psycho-
sis. Behav Res Ther. 2000;38:1097–1106.
Acknowledgments 16. Hoffman RE, Varanko M, Gilmore J, Mishara AL. Experien-
tial features used by patients with schizophrenia to differenti-
We would like to thank Richard Bentall and Daniel ate ’voices’ from ordinary verbal thought. Psychol Med.
Freeman for their input in an earlier version of this 2008;38:1167–1176.
manuscript. We also wish to sincerely apologize to the 17. Moritz S, Laroi F. Differences and similarities in the sensory
authors whose important work could not be included and cognitive signatures of voice-hearing, intrusions and
due to a journal cap in number of references. Conflicts thoughts. Schizophr Res. 2008;102:96–107.
of interest: The authors have declared that there are 18. Bullen JG, Hemsley DR. Schizophrenia: a failure to control
the contents of consciousness. Br J Clin Psychol. 1987;26:
no conflicts of interest in relation to the subject of this
25–33.
study.
19. Peters ER, Pickering AD, Kent A, et al. The relationship be-
tween cognitive inhibition and psychotic symptoms.
J Abnorm Psychol. 2000;109:386–395.
References
20. David AS, Lucas PA. Auditory-verbal hallucinations and the
1. Cuthbert BN, Insel TR. Toward new approaches to psychotic phonological loop: a cognitive neuropsychological study. Br
disorders: the NIMH Research Domain Criteria project. J Clin Psychol. 1993;32(pt 4):431–441.
Schizophr Bull. 2010;36:1061–1062. 21. Wible CG, Lee K, Molina I, et al. fMRI activity correlated
2. Stephane M. Standardized assessment of hallucinations. In: with auditory hallucinations during performance of a working
Jardri R, Pins D, Cachia A, Thomas P, eds. The Neuroscience memory task: data from the FBIRN consortium study. Schiz-
of Hallucinations.. New York, NY: Springer; In press. ophr Bull. 2009;35:47–57.
3. Feinberg I. Efference copy and corollary discharge: implica- 22. Jardri R, Pouchet A, Pins D, Thomas P. Cortical activations
tions for thinking and its disorders. Schizophr Bull. during auditory verbal hallucinations in schizophrenia: a
1978;4:636–640. coordinate-based meta-analysis. Am J Psychiatry. 2011;168:
4. Frith CD, Blakemore S, Wolpert DM. Explaining the symp- 73–81.
toms of schizophrenia: abnormalities in the awareness of ac- 23. Hugdahl K. ‘‘Hearing voices’’: auditory hallucinations as fail-
tion. Brain Res Brain Res Rev. 2000;31:357–363. ure of top-down control of bottom-up perceptual processes.
5. Waters FA, Badcock JC, Michie PT, Maybery MT. Auditory Scand J Psychol. 2009;50:553–560.
hallucinations in schizophrenia: intrusive thoughts and for- 24. Aleman A, Bocker KB, Hijman R, de Haan EH, Kahn RS.
gotten memories. Cogn Neuropsychiatry. 2006;11:65–83. Cognitive basis of hallucinations in schizophrenia: role of
6. Gallagher S. Neurocognitive models of schizophrenia: a neu- top-down information processing. Schizophr Res. 2003;64:
rophenomenological critique. Psychopathology. 2004;37:8–19. 175–185.
7. Allen PP, Johns LC, Fu CH, Broome MR, Vythelingum GN, 25. Krishnan RR, Keefe R, Kraus M. Schizophrenia is a disorder
McGuire PK. Misattribution of external speech in patients of higher order hierarchical processing. Med Hypotheses.
with hallucinations and delusions. Schizophr Res. 2004;69:
2009;72:740–744.
277–287.
26. Vercammen A, de Haan EH, Aleman A. Hearing a voice in
8. Waters F, Woodward T, Allen P, Aleman A, Sommer I. Self-
the noise: auditory hallucinations and speech perception.
recognition deficits in schizophrenia patients with auditory
hallucinations: a meta-analysis of the literature [published Psychol Med. 2008;38:1177–1184.
online ahead of print December 08, 2010]. Schizophr Bull. 27. Haddock G, Slade P, Bentall RP. Auditory hallucinations
doi: 10.1093/schbul/sbq144. and the verbal transformation effect: the role of suggestions.
9. Bentall RP, Slade PD. Reality testing and auditory hallucina- Pers Individ Dif. 1995;19:301–305.
tions: a signal detection analysis. Br J Clin Psychol. 1985; 28. Ilankovic LM, Allen PP, Engel R, et al. Attentional modu-
24(pt 3):159–169. lation of external speech attribution in patients with hallu-
10. Dodgson G, Gordon S. Avoiding false negatives: are some cinations and delusions. Neuropsychologia. 2011;49:805–
auditory hallucinations an evolved design flaw? Behav Cogn 812.
Psychother. 2009;37:325–334. 29. Bentall RP. The illusion of reality: a review and integration of
11. Harvey P. Reality Monitoring in Mania and Schizophrenia. psychological research on hallucinations. Psychol Bull. 1990;
J Nerv Ment Dis. 1985;173:67–73. 107:82–95.

691
9
F.
F. Waters
Waters et
et al.
al.

30. Kot T, Serper M. Increased susceptibility to auditory condi- ments and relation to psychosis-like experiences. Conscious
tioning in hallucinating schizophrenic patients: a preliminary Cogn. 2008;17:576–586.
investigation. J Nerv Ment Dis. 2002;190:282–288. 43. Paulik G, Badcock JC, Maybery MT. Dissociating the com-
31. Kirsch I, Wickless C, Moffitt KH. Expectancy and suggest- ponents of inhibitory control involved in predisposition to
ibility: are the effects of environmental enhancement due to hallucinations. Cogn Neuropsychiatry. 2008;13:33–46.
detection? Int J Clin Exp Hypn. 1999;47:40–45. 44. Jones SR, Fernyhough C. Rumination, reflection, intrusive
32. Freeman D, Garety PA. Connecting neurosis and psychosis: thoughts, and hallucination-proneness: towards a new model.
the direct influence of emotion on delusions and hallucina- Behav Res Ther. 2009;47:54–59.
tions. Behav Res Ther. 2003;41:923–947. 45. Varese F, Bentall RP. The metacognitive beliefs account of
33. Birchwood M, Meaden A, Trower P, Gilbert P, Plaistow J. hallucinatory experiences: a literature review and meta-
The power and omnipotence of voices: subordination and en- analysis. Clin Psychol Rev. 2011;31:850–864.
trapment by voices and significant others. Psychol Med. 46. van ’t Wout M, Aleman A, Kessels RP, Laroi F, Kahn RS.

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ at UNAM Direccion General de Bibliotecas on March 30, 2016
2000;30:337–344. Emotional processing in a non-clinical psychosis-prone sam-
34. Morrison R, O’Connor RC. The role of rumination, atten- ple. Schizophr Res. 2004;68:271–281.
tional biases and stress in psychological distress. Br J Psychol. 47. Jones S, Fernyhough C, Meads D. In a dark time: develop-
2008;99(pt 2):191–209. ment, validation, and correlates of the Durham Hypnagogic
35. Romme MAJ, Escher A. Hearing voices. Schizophr Bull. and Hypnopompic Hallucinations Questionnaire. Pers Individ
1989;15:209–216. Dif. 2009;46:30–34.
36. Offen L, Thomas G, Waller G. Dissociation as a mediator of 48. Johns L, Gregg L, Allen P, McGuire P. Impaired verbal self-
the relationship between recalled parenting and the clinical monitoring in psychosis: effects of state, trait and diagnosis.
correlates of auditory hallucinations. Br J Clin Psychol. Psychol Med. 2006;36:465–474.
2003;42(pt 3):231–241. 49. Kingdon DG, Ashcroft K, Bhandari B, et al. Schizophrenia
37. Varese F, Barkus E, Bentall RP. Dissociation mediates the re- and borderline personality disorder: similarities and differen-
lationship between childhood trauma and hallucination- ces in the experience of auditory hallucinations, paranoia, and
proneness. Psychol Med. 2011;6:1–12. childhood trauma. J Nerv Ment Dis. 2010;198:399–403.
38. Daalman K, Boks MP, Diederen KM, et al. The same or dif- 50. Barnes J, Boubert L. Executive functions are impaired in
ferent? A phenomenological comparison of auditory verbal patients with Parkinson’s disease with visual hallucinations.
hallucinations in healthy and psychotic individuals. J Clin J Neurol Neurosurg Psychiatry. 2008;79:190–192.
Psychiatry. 2011;72:320–325. 51. Graham G, Dean J, Mosimann UP, et al. Specific attentional
39. Barkus E, Stirling J, Hopkins R, McKie S, Lewis S. Cognitive impairments and complex visual hallucinations in eye disease.
and neural processes in non-clinical auditory hallucinations. Int J Geriatr Psychiatry. 2011;26:263–267.
Br J Psychiatry Suppl. 2007;51:s76–s81. 52. Frith C, Dolan RJ. Brain mechanisms associated with top-
40. Vercammen A, Aleman A. Semantic expectations can induce down processes in perception. Philos Trans R Soc Lond B
false perceptions in hallucination-prone individuals. Schizophr Biol Sci. 1997;352:1221–1230.
Bull. 2010;36:151–156. 53. Diederen KM, Neggers SF, Daalman K, et al. Deactivation
41. Johns LC, Allen P, Valli I, et al. Impaired verbal self- of the parahippocampal gyrus preceding auditory hallucina-
monitoring in individuals at high risk of psychosis. Psychol tions in schizophrenia. Am J Psychiatry. 2010;167:427–435.
Med. 2010;40:1433–1442. 54. Jones SR. Do we need multiple models of auditory verbal hal-
42. Jones SR, de-Wit L, Fernyhough C, Meins E. A new spin on lucinations? Examining the phenomenological fit of cognitive
the Wheel of Fortune: priming of action-authorship judge- and neurological models. Schizophr Bull. 2010;36:566–575.

692
10