The continuum model of psychosis has been extremely influential. It assumes that psychotic symptoms, such as auditory verbal hallucinations (AVH), are not limited to patients with psychosis but also occur in healthy, non-clinical individuals – suggesting similar mechanisms of origin. Recent debate surrounding this model has highlighted certain differences, as well as similarities, in the phenomenology of AVH in clinical and non-clinical populations. These findings imply that there may, in fact, be only partial overlap of the mechanism(s) involved in generating AVH in these groups. We review evidence of continuity or similarity, and dissimilarity, in cognitive, and related neural processes, underlying AVH in clinical and non-clinical samples. The results reveal some shared (intrusive cognitions, inhibitory deficits) and some distinct (aspects of source memory and cerebral lateralization) mechanisms in these groups. The evidence, therefore, supports both continuous and categorical models of positive psychotic symptoms. The review considers potential risks of uncritical acceptance of the continuum model and highlights some important methodological issues for future research.
In recent years, a dimensional approach to understanding psychotic symptoms has become firmly established, commonly referred to as the continuum model of psychosis (Allardyce et al., 2007, Johns and van Os, 2001, van Os et al., 2009 and Verdoux and van Os, 2002, for reviews). A wealth of empirical evidence now shows, for example, that auditory verbal hallucinations (AVH) are commonly reported by healthy individuals without mental illness, as well as those diagnosed with a psychotic disorder, such as schizophrenia (e.g. Sommer et al., 2010). Such findings suggest that the experience of hearing voices lies on a continuum with normality (Claridge, 1990, for a theoretical model; Choong et al., 2007, Pierre, 2010 and Stip and Letourneau, 2009, for reviews). The rate of AVH reported in the general population varies, as a result of methodological and design factors, from 5.7–21.0% in children and adolescents (e.g. Bartels-Velthuis et al., 2010 and McGee et al., 2000) to 10–15% of the adult population (Tien, 1991 and Sommer et al., 2010). The majority of these hallucinatory experiences are transitory but some persist – increasing the risk of secondary delusional ideation and transition to psychosis (see De Loore et al., 2011, Dominguez et al., 2011 and Smeets et al., 2010, for empirical evidence; van Os et al., 2009, for a review). Interestingly, Bartels-Velthuis et al. (2011) found better theory-of-mind skills in children hearing voices, a finding the authors suggested could mitigate the risk of secondary delusion formation.
The presence of a continuum of psychotic symptoms is often taken to imply fundamentally the same phenomenological experience which, though varying in severity depends upon the same cognitive and neural mechanisms in psychotic and non-psychotic populations (see Esterberg and Compton, 2009 as an example). Consequently, it is commonly stated that studying hallucinatory experiences in non-clinical samples is likely to unveil the essential cognitive and neural mechanisms underlying schizophrenic hallucinations, while avoiding confounding effects associated with medication, hospitalization and illness duration. The advantages associated with the continuum model have led some to call for a dimensional approach to the classification of psychotic disorders (Dutta et al., 2007 and Peralta and Cuesta, 2007, for reviews and discussions). Nonetheless, several authors have called for a re-evaluation of the continuum concept (see discussion by David, 2010; and related commentaries by Kaymaz and van Os, 2010 and Sommer, 2010) wherein AVH are assumed to be not inherently pathological, and yet at the core of psychosis.
One important and potentially revealing challenge to the continuum model emerges from phenomenological comparisons between clinical (psychotic) and non-clinical (healthy) AVH. Although these point to some similarities in the characteristic features of AVH in both groups (voices heard inside or outside the head, loudness, number of voices, and attribution of voices to a real or familiar person) and a continuum of disability, they also highlight significant differences, as first reported by Honig et al. (1998) and Romme and Escher (1989). In particular, recent data provided in separate studies by Daalman et al. (2011) and Lawrence et al. (2010) show that a cluster of features – the frequency, emotional valence of beliefs and content, experience of control, age of onset and preponderance of male voices – clearly distinguishes AVH heard by patients with schizophrenia from those experienced by healthy (non-psychotic) adults in the general community. These findings raise the possibility that there may be only partial overlap in AVH experiences and the underlying cognitive (and neural) mechanisms involved (Kaymaz and van Os, 2010), in these groups. Moreover, current cognitive models of AVH assume that more than one mechanism is likely to be involved in the development of hallucinations (see discussions and literature overviews by Hugdahl, 2009, Jones, 2008, Seal et al., 2004 and Waters et al., 2006a). Such models also leave scope for the possibility that only some, but not all, cognitive mechanisms underpinning AVH occur on a continuum.
The aim of this paper, therefore, was to review current evidence regarding similarities and dissimilarities in cognitive processes underlying AVH in psychosis and in healthy, non-clinical populations.
