Introduction
In the general community, and even among health professionals, the presence of auditory hallucinations (AH) is often considered synonymous with having a psychotic disorder (Parnas, Reference Parnas, Stanghellini and Fuch2013; Ford et al. Reference Ford, Morris, Hoffman, Sommer, Waters, McCarthy-Jones, Thoma, Turner, Keedy, Badcock and Cuthbert2014). This view is promoted by the diagnostic criteria for Other Specified Schizophrenia Spectrum and Other Psychotic Disorder (OSSSOPD), a diagnostic category first introduced in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (American Psychiatry Association, 2013). For related categories in the schizophrenia spectrum, such as Schizophrenia, Schizophreniform Disorder, and Schizoaffective Disorder, at least two symptoms are required from those listed under the A-criterion (comprising delusions, hallucinations, disordered thinking, disorganized or catatonic behavior, and negative symptoms). To meet the DSM-5 criteria for a diagnosis of OSSSOPD, however, the sole presence of persistent AH is sufficient. This is at odds with the fact that AH are experienced frequently in the context of other conditions, and sometimes even in the absence of any demonstrable pathology. Given that persistent AH are not always associated with a loss of contact with reality, lack of insight, or compromised dopaminergic transmission, it is questionable whether these types of AH should be treated with antipsychotic medication. Getting this issue right is also important given the stigma associated with a diagnosis of psychotic disorder and possible consequences for prognosis.
In this paper, we argue for a clearer distinction between AH as part of a psychotic disorder and those experienced in a different context, as the former may be an indication for antipsychotic drugs, while the latter may benefit from an alternative therapeutic approach. In what follows, we (i) consider the representation of hallucinations in the DSM-5 chapter on Schizophrenia Spectrum Disorders; (ii) highlight the crucial difference between psychotic disorders and psychotic symptoms; (iii) document alternative disorders with AH as a prominent feature which occur in relative isolation, i.e. in the absence of any other psychotic symptoms; (iv) discuss biological, environmental, and psychological explanations for AH that do not involve dopamine; and (v) consider whether AH are a risk factor for later psychosis. On the basis of these insights, we advocate a more stringent use of the DSM's A-criterion, requiring the presence of at least two symptoms in cases of OSSSOPD. Although the focus here is on nonverbal and verbal AH (i.e. ‘voices’), our argument also applies to hallucinations experienced in other sensory modalities (i.e. visual, somatic, etc.).
Hallucinations as a symptom domain in the DSM-5
According to the DSM-5, Schizophrenia and Schizophreniform Disorder are diagnosed in the presence of two or more symptoms listed under the A-criterion, along with signs of persistence, distress, and impairments in functioning. One of the changes made after the DSM-IV was to drop the use of Schneiderian first-rank symptoms, including voices conversing and third-person hallucinations (Esquirol, Reference Esquirol1845). Consequently, it is now possible to establish a diagnosis of Schizophrenia or Schizophreniform Disorder whenever hallucinations are present, regardless of form or content, as long as they are accompanied by a second symptom from the A-criterion, together with significant distress or impairment in functioning (APA, 2013).
The category OSSSOPD was devised for clinical presentations that are deemed ‘characteristic’ of the group of Schizophrenia Spectrum Disorders, but which do not meet all the diagnostic criteria of Schizophrenia or Schizophreniform Disorder. The diagnosis can be established in the sole presence of persistent AH, i.e. in the absence of any other features (Box 1). An important benefit of this low-threshold residual category is that it helps to alert health professionals to the presence of AH in need of treatment, even in the absence of additional symptoms required for a diagnosis of Schizophrenia. Thus, patients suffering exclusively from distressing and burdensome hallucinations (for example command hallucinations) can be treated without delay. However, this creates a new problem of overinclusion, because any individual presenting with AH may now become labelled as having a psychotic disorder, even if the primary condition is a somatic disorder (e.g. migraine or Parkinson's disease), other psychiatric disorder (e.g. dissociative disorder or personality disorder) or not demonstrable at all. The risk is that patients might be given unnecessary treatment and suffer from stigmatization. Incidentally, the risk of overinclusion is not unique to Schizophrenia Spectrum Disorders (Berrios, Reference Berrios1996), but this issue has received little attention in the literature.
Box 1 DSM-5 criteria for Other Specified Schizophrenia Spectrum and Other Psychotic Disorder (OSSSOPD)
This category applies to presentations where are prominent symptoms characteristic of schizophrenia spectrum disorders that cause clinically significant distress or impairment in social, occupational, or other important areas of functioning but do not meet the full criteria for schizophrenia, schizophreniform disorder or schizoaffective disorder. The OSSSOPD category is used in situations in which the clinician chooses to communicate that the presentation does not meet the criteria for a more specific schizophrenia spectrum disorder, and allows him or her to provide a specific reason for this (e.g. ‘persistent auditory hallucinations’).
Examples of presentations that can be specified using the ‘other specified’ designation include the following:
-
(1) Persistent auditory hallucinations occurring in the absence of any other features.
-
(2) Delusions with significant overlapping mood episodes: This includes persistent delusions with periods of overlapping mood episodes that are present for a substantial portion of the delusional disturbance (such that the criterion stipulating only brief mood disturbance in delusional disorder is not met).
-
(3) Attenuated psychosis syndrome: This syndrome is characterized by psychotic-like symptoms that are below a threshold for full psychosis (e.g. the symptoms are less severe and more transient, and insight is relatively maintained).
-
(4) Delusional symptoms in partners of individual with delusional disorder: In the context of a relationship, the delusional material from the dominant partner provides content for delusional beliefs by an individual who may not otherwise entirely meet criteria for delusional disorder.
Psychotic disorder v. psychotic symptom
Psychosis is a broad and rather non-specific term (David & Ajnakina, Reference David and Ajnakina2016). It has been in use since the 1900s, when physicians began attributing mental disorders (or ‘insanities’) to underlying disorders of the brain. Since then, the term is featured in virtually every textbook of psychiatry, but is seldom explicitly defined. Even Sims (Sims, Reference Sims2003), whose crystal-clear definitions of psychopathology have instructed generations of psychiatrists, has little more to say than that psychoses are ‘exceedingly hard to define although they are usually said to be characterized by severe symptoms, such as delusions and hallucinations, and by lack of insight (…); there is loss of contact with reality.’ In clinical practice, the term psychosis loosely encompasses a spectrum of diagnostic syndromes which includes schizophrenia and related disorders, the affective psychoses (i.e. major depressive disorder and bipolar disorder), substance-induced psychoses and withdrawal states, psychoses caused by a somatic disorder, and psychotic states associated with neurodegenerative disorders such as dementia, Parkinson's disease and Huntington's disease. Delirium is also sometimes included as a type of psychosis.
The term ‘psychosis’ therefore lacks a definition which provides meaning by specifying its necessary and sufficient conditions (i.e. an intensional definition). Attempts to provide such an intensional definition in terms of ‘loss of contact with reality’ or ‘lack of insight’ fail to cover the meaning of psychosis. After all, a loss of contact with reality cannot be considered sufficient for psychosis, as this feature may occur in other disorders such as autism, depressive disorder or eating disorder; nor is it a necessary feature, since individuals suffering from hallucinations and negative symptoms may still be in contact with reality. The same holds true for lack of insight (David, Reference David1990; Varga et al. Reference Varga, Magnusson, Flekkøy, Rønneberg and Opjordsmoen2006).
Instead, ‘psychosis’ owes its meaning to extensional definitions which vary somewhat between authors, but usually consist of lists of psychotic symptoms (such as those mentioned under the A-criterion above). If we ask ourselves why symptoms from that disparate group are called psychotic, we find that this serves to indicate that we consider them severe, and that we suspect them to signal a loss of contact with reality and/or a lack of insight. Moreover, historically, it suggests that we consider them to have a neurobiological cause even though the cause or mechanism may elude us. That is as close as we can currently get to an intensional definition of psychosis.
The question to be answered now, is whether all patients with one or more psychotic symptoms (such as persistent, stressful AH) should be considered as having an underlying psychotic disorder.
Persistent AH in non-psychotic disorders
It is widely recognized that AH can accompany virtually all psychiatric disorders, including borderline personality disorder and dissociative identity disorder (40%), unipolar depression and bipolar disorder (45%), anxiety disorders (14%), autism (6%), and post-traumatic stress disorder (15%) (Blom & Sommer, Reference Blom and Sommer2010; Sommer & Kahn, Reference Sommer and Kahn2014). In addition, AH can be a symptom of disorders of the nervous system, including Parkinson's disease, stroke, and migraine (Vreeburg et al. Reference Vreeburg, Leijten and Sommer2016), tumors or lesions to the temporal lobe, brainstem or thalamus (Braun et al. Reference Braun, Dumont, Duval, Hamel-Hebert and Godbout2003), sleep disorders (Fortuyn et al. Reference Fortuyn, Lappenschaar, Nienhuis, Furer, Hodiamont, Rijnders, Lammers, Renier, Buitelaar and Overeem2009), or due to the effects of alcohol and drugs. Sensory disturbances due to conditions which affect the auditory pathways (e.g. hearing loss, auditory cortex dysfunction, etc.) are also a common cause.
Hallucinations in these different groups of clinical disorders can be persistent and distressing, and be associated with a depressed mood, anxiety, and suicidal behavior (Honig et al. Reference Honig, Romme, Ensink, Escher, Pennings and Devries1998; Okulate & Jones, Reference Okulate and Jones2003; Blom & Sommer, Reference Blom and Sommer2010; Kingdon et al. Reference Kingdon, Ashcroft, Bhandari, Gleeson, Warikoo, Symons, Taylor, Lucas, Mahendra and Ghosh2010; Preti et al. Reference Preti, Sisti, Rocchi, Siddi, Cella, Masala, Petretto and Carta2014). Even though such AH can also present with significant impairments in social, occupational and other areas of functioning, they are not always accompanied by problems with reality testing, lack of insight or disturbed beliefs, and therefore cannot be considered as part of a psychotic disorder.
Hallucinations in individuals in the general population
One such example is that of AH reported by individuals in the general population. In the so-called ‘extended phenotype’ (van Os et al. Reference van Os, Linscott, Myin-Germeys, Delespaul and Krabbendam2009), prevalence estimates of hallucinations vary from 4% to 21% (Beavan et al. Reference Beavan, Read and Cartwright2011). Hallucinations in this group are often transient and sporadic in nature (Hanssen et al. Reference Hanssen, Bak, Bijl, Vollebergh and van Os2005), although AH can be recurrent and persistent in a minority (4%) (McGrath et al. Reference McGrath, Saha, Al-Hamzawi, Alonso, Bromet, Bruffaerts, Caldas-de-Almeida, Chiu, de Jonge, Fayyad and Florescu2015). In this latter group, AH can be highly distressing in approximately 50% of cases (Honig et al. Reference Honig, Romme, Ensink, Escher, Pennings and Devries1998; Daalman et al. Reference Daalman, Boks, Diederen, de Weijer, Blom, Kahn and Sommer2011; Woods et al. Reference Woods, Jones, Alderson-Day, Callard and Fernyhough2015), and may last for over a year (Peters et al. Reference Peters, Ward, Jackson, Morgan, Charalambides, McGuire, Woodruff, Jacobsen, Chadwick and Garety2016). In one study of 13 voice hearers in the general population, hallucinations had been ongoing for longer than 4 years in 89% of cases (Leudar et al. Reference Leudar, Thomas, McNally and Glinski1997). These types of hallucination present mostly without any other psychotic symptoms, and there is a marked absence of psychopathology but for schizotypal traits (Sommer et al. Reference Sommer, Daalman, Rietkerk, Diederen, Bakker, Wijkstra and Boks2010).
Explanations for AH which do not involve dopamine
Dopamine has once been hailed as the ‘wind of the psychotic fire’ (Laruelle & Abi-Dargham, Reference Laruelle and Abi-Dargham1999), and antipsychotic medications with their antagonistic effects on dopamine receptors are routinely used to treat psychotic symptoms, whether or not in combination with psychosocial approaches. Evidence is accumulating, however, that the causes of AH do not always involve abnormal dopamine functions, raising questions about the systematic usage of antipsychotic medications for all symptoms considered as psychotic.
Pharmacological evidence
Dopaminergic overactivity in the striatum (Abi-Dargham et al. Reference Abi-Dargham, Gil, Krystal, Baldwin, Seibyl, Bowers, van Dyck, Charney, Innis and Laruelle1998) is the primary neuropharmacological explanation for positive symptoms in schizophrenia disorders. In support, Positron Emission Tomography (PET) studies using an F-Dopa tracer show that many individuals with these disorders have increased presynaptic striatal dopamine production (Howes et al. Reference Howes, Bose, Turkheimer, Valli, Egerton, Stahl, Valmaggia, Allen, Murray and McGuire2011). Using the same method, a study in individuals from the general population (e.g. nonpsychotic) with frequent AH failed to show any such increases of striatal dopamine production (Howes et al. Reference Howes, Shotbolt, Bloomfield, Daalman, Demjaha, Diederen, Ibrahim, Kim, McGuire, Kahn and Sommer2013). This perhaps suggests that individuals who participated in the latter study may not have been benefited from antipsychotic treatment in the treatment of their hallucinations. In support, other PET studies indicate that targeting dopamine (D2) receptors with antipsychotic medication is only beneficial for individuals with increased striatal dopamine production, and not for those with normal production levels (Demjaha et al. Reference Demjaha, Murray, McGuire, Kapur and Howes2012; Ćurčić-Blake et al. Reference Ćurčić-Blake, Bais, Sibeijn-Kuiper, Pijnenborg, Knegtering, Liemburg and Aleman2017).
Consequently, it is perhaps not surprising that recent evidence is pointing to neurotransmitter systems other than dopamine in the mediation of AH (Deamer & Wilkinson, Reference Deamer and Wilkinson2015). In regards to the glutamatergic system, for example, studies show elevated levels of glutamate and glutamine metabolites in the temporal and frontal brain regions of individuals diagnosed with schizophrenia with frequent and severe AH (Hayward et al. Reference Hayward, Overton, Dorey and Denney2009; Hugdahl et al. Reference Hugdahl, Craven, Nygård, Løberg, Berle, Johnsen, Kroken, Specht, Andreassen and Ersland2015; Ćurčić-Blake et al. Reference Ćurčić-Blake, Bais, Sibeijn-Kuiper, Pijnenborg, Knegtering, Liemburg and Aleman2017). In addition, studies with ketamine show that drugs which alter glutamatergic neurotransmission are also capable of producing AH (Corlett et al. Reference Corlett, Honey, Krystal and Fletcher2011). Such a role for the glutamatergic system cannot be linked directly to the dopamine hypothesis of psychotic disorders. Despite attempts to couple both views by considering a role for aberrant NMDA-R plasticity (Thomas et al. Reference Thomas, Hayward, Peters, van der Gaag, Bentall, Jenner, Strauss, Sommer, Johns and Varese2014), the glutamate system appears to form an alternative mediating pathway for AH, consistent with the heterogeneity of conditions in which such hallucinations present.
To complicate things further, hallucinations in different sensory modalities - even when experienced in the context of the same disorder – may be mediated by different types of neurotransmitter. In the case of visual hallucinations, for example, 5HT-2a receptors appear to play a pivotal role given that serotonin agonists such as LSD and psilocybin can produce vivid hallucinations, which can subsequently be blocked with serotonin (5HT-2A) antagonists such as ketanserin and ritanserin (Vollenweider, Reference Vollenweider1998). This evidence points to different neurotransmitter pathways for hallucinations in different sensory modalities, again forcing the conclusion that the dopamine hypothesis of schizophrenia does not provide an explanation for all types of hallucination.
Environmental and psychological explanations
Other evidence that dopamine is not always directly involved in AH stems from studies on environmental and psychological factors. In vulnerable and older age groups, social isolation and withdrawal can be powerful causes for AH and other types of hallucination through the reduction of ordinary levels of sensory input. It has been speculated that the psychological effects of loneliness and social seclusion may prompt compensatory hypersensitivity of the perceptual network, with any hallucinations being projected outwards to meet the person's communicative needs (Hoffman, Reference Hoffman2007). Psychological trauma is yet another potential cause for hallucinations. Some people can trace their symptom's onset back to traumatic events. This association has been found in many individuals diagnosed with schizophrenia who report abuse in childhood (Read et al. Reference Read, van Os, Morrison and Ross2005), and to an even greater extent in individuals with borderline personality disorder and post-traumatic stress disorder experiencing AH. In the general community, traumatic events also have a substantive power to trigger hallucinations. For example, up to 80% of the recently bereaved report hallucinations of their loved one within the first month of the person's death (Grimby, Reference Grimby1993). Although predominantly visual in nature, these hallucinations can also have an auditory component. In youth, stressors such as bullying and sexual trauma are strong predictors of AH (Lardinois et al. Reference Lardinois, Lataster, Mengelers, Van Os and Myin-Germeys2011).
But even outside the context of trauma, psychological factors in general appear to play a pivotal role in prompting hallucinations. Hypervigilance, for example, and the way situations are appraised, act to shape and give credence to AH (Campbell & Morrison, Reference Campbell and Morrison2007), with anxiety acting to reduce the threshold for accepting ambiguous signals as real. Everyday examples include mothers believing their babies to be crying, and physicians on duty hearing the phantom ring of their pagers. In general, all conditions which are accompanied by high expectancy for a perceptual signal are a fertile ground for hallucinations.
Even though it is sometimes possible to treat such psychosocially-induced hallucinations with antipsychotics, psychosocial interventions should be the treatment of first choice. Options could include case-management approaches, psychoeducation, improving coping mechanisms, cognitive therapies (Hayward et al. Reference Hayward, Overton, Dorey and Denney2009; Deamer & Wilkinson, Reference Deamer and Wilkinson2015), cognitive-behavioral therapy (CBT) (Thomas et al. Reference Thomas, Hayward, Peters, van der Gaag, Bentall, Jenner, Strauss, Sommer, Johns and Varese2014), trauma-related therapy (McCarthy-Jones & Longden, Reference McCarthy-Jones and Longden2015; Steel, Reference Steel2015), voice-dialoging (Corstens et al. Reference Corstens, Longden and May2012), mindfulness approaches (Strauss et al. Reference Strauss, Thomas and Hayward2015), and integrative treatments (Jenner et al. Reference Jenner, Nienhuis, Wiersma and van de Willige2004).
In all, the above body of evidence converges with findings from the literature on schizophrenia, but also with the notion that there are many pathways independent of the dopamine system which are capable of eliciting hallucinations.
Are hallucinations a risk factor for later psychosis?
The question of whether isolated AH constitute a risk factor for developing a psychotic disorder, and therefore whether individuals with these experiences should receive early therapeutic interventions, cannot be answered conclusively. At this stage, however, the existing evidence suggests that the risk is low, unless AH are accompanied by other psychotic symptoms and/or functional deficits. Longitudinal follow-up studies of individuals who first experienced AH in childhood or early adolescence confirm this (Poulton et al. Reference Poulton, Caspi, Moffitt, Cannon, Murray and Harrington2000).
While the prevalence of hallucinations in children is relatively high (i.e. 9%), these experiences are largely transient in nature (Jardri et al. Reference Jardri, Bartels-Velthuis, Debbané, Jenner, Kelleher, Dauvilliers, Plazzi, Demeulemeester, David and Rapoport2014). A study of 337 children aged 7–8 years reporting hallucination-like experiences demonstrated that these experiences ceased spontaneously within 5 years in 76% of cases (Bartels-Velthuis et al. Reference Bartels-Velthuis, van de Willige, Jenner, van Os and Wiersma2011). Rather than being associated with a later psychotic disorder, AH in children and adolescents are more frequently associated with a later depressive disorder, anxiety disorder or behavioral problems, even after adjusting for alcohol and illicit-substance abuse (Kelleher et al. Reference Kelleher, Keeley, Corcoran, Lynch, Fitzpatrick, Devlin, Molloy, Roddy, Clarke, Harley, Arseneault, Wasserman, Carli, Sarchiapone, Hoven, Wasserman and Cannon2012). Other studies concur, showing that the odds of AH alone to predict a psychotic disorder are low, unless they are accompanied by functional deficits or beliefs about malevolent intentions of the voices (Poulton et al. Reference Poulton, Caspi, Moffitt, Cannon, Murray and Harrington2000; Daalman et al. Reference Daalman, Diederen, Hoekema, van Lutterveld and Sommer2016).
Other evidence comes from studies showing that the transition rate of young individuals with subclinical symptoms to psychosis varies widely. Psychotic symptoms that include hallucinations, delusional ideation and self-disturbances confer a relatively high risk for psychotic disorders in adulthood (Krabbendam et al. Reference Krabbendam, Myin-Germeys, Hanssen, Bijl, de Graaf, Vollebergh, Bak and van Os2004), but less so when AH occur in isolation (Daalman et al. Reference Daalman, Diederen, Hoekema, van Lutterveld and Sommer2016). A number of variables appear to influence the transition to psychosis, including the persistence of symptoms over time, the onset of secondary delusions, depression, affective dysregulation and psychosocial dysfunction (Kaymaz & van Os, Reference Kaymaz and van Os2010). Conduct disorder, when co-occurring with AH, also predicts the transition to more severe forms of psychopathology (Askenazy et al. Reference Askenazy, Lestideau, Meynadier, Dor, Myquel and Lecrubier2007).
Finally, studies of individuals in the prodromal phase (and in those with a diagnosis of schizophrenia spectrum disorder) show that no single symptom can predict the development of psychotic disorder or relapse into psychosis. More accurate predictions depend on features such as other psychotic manifestations, cognitive dysfunction, depression, and poor social functioning (Yung et al. Reference Yung, Phillips, Yuen, Francey, McFarlane, Hallgren and McGorry2003).
Conclusions
Altogether, AH and other hallucinations occur at relatively high rates in many different conditions and are not pathognomonic for any given disorder (including schizophrenia) (Ford et al. Reference Ford, Morris, Hoffman, Sommer, Waters, McCarthy-Jones, Thoma, Turner, Keedy, Badcock and Cuthbert2014). AH can occur without other psychotic symptoms and may be caused by a range of conditions in which dopaminergic transmission is not compromised. In all cases, an accurate diagnosis is critically important, as persistent AH as part of a psychotic disorder may be a good indication for antipsychotic medication, whereas persistent AH in the absence of such a disorder are probably not. Although antipsychotic medication can be very effective and potentially life-saving in individuals with psychotic disorders, it is questionable whether it should be offered to all individuals with AH.
We believe the ambiguous use of the term ‘psychosis’ has caused much confusion in the minds of professionals and the lay public, and that the distinction between psychotic symptom and psychotic disorder is a particularly important one since psychotic symptoms can occur outside of a psychotic disorder. The habitual conceptualization of AH as a sure sign of schizophrenia spectrum disorder is a particularly good example of this problem, although a primary diagnosis of psychotic disorder is not justified for all individuals troubled by persistent hallucinations.
The roots of classification systems in conventional medical models require a diagnosis to rest on the presence and absence of several symptoms to determine whether the criteria for a specific syndrome or disorder may be fulfilled. In the traditional view of psychosis (as a state of distorted reality testing etcetera), the concomitant expression of at least one more psychotic symptom is critical. To that end, the DSM's A-criterion for schizophrenia is very useful, since it demands the presence of at least two psychotic symptoms. In the sole presence of persistent hallucinations, we argue that the diagnostic criteria for Other Specified Schizophrenia Spectrum And Other Psychotic Disorder is not helpful.
Alternative locations for the listing of hallucinations in textbooks of psychiatry include other syndromes and disorders where the presence of other symptoms fulfil the criteria for these conditions (e.g. PTSD, borderline personality disorder, etc). Where other symptoms or the underlying cause cannot be clearly established or ascertained, a broader category akin to the ICD's R44 (‘symptoms involving general sensations and perceptions’) may be helpful, perhaps under a new banner titled ‘Perceptual disorders’. This new domain could include a class of perceptual phenomena causing distress and dysfunction, but without any impaired reality testing (including tinnitus, Charles Bonnet Syndrome, phantom limb pain, pareidolia, and Alice in Wonderland syndrome, to mention a few examples). This may help to improve differential diagnosis, and in differentiating between individuals who may benefit from antipsychotic medication and those who may not. Altogether, we should take note that the presence of persistent AH does not equate the presence of a psychotic disorder, and should therefore not uncritically be diagnosed or treated as such.
Acknowledgement
KH was funded by an ERC Advanced Grant #693124, and a RCN grant #223273.
