We examined the cognitive correlates of hallucinations and delusions in 47 schizophrenia spectrum individuals. Hallucinations were significantly negatively correlated with performance on episodic memory tasks, and were not significantly associated with performance on tasks measuring fluency or concentration/attention. Although hallucinations were more strongly associated with performance on verbal than non-verbal memory tasks, the difference was not statistically significant. There was also a trend for hallucinations to be associated with poorer performance on working memory tasks, though this association was eliminated when episodic memory performance was taken into account. Delusions were not significantly associated with any of the cognitive measures.
Two important and very common symptoms of schizophrenia are delusions and hallucinations. Although both delusions and hallucinations are considered positive symptoms of schizophrenia, psychopathology researchers have often proposed very different mechanisms to explain them. For example, whereas hallucinations are often posited to be the result of disturbances in language processes and semantic memory (e.g., Hoffman and McGlashan, 1997 and Hoffman et al., 1999), delusions are often posited to be the result of cognitive biases and misattributions (e.g., Garety and Freeman, 1999 and Bentall et al., 2001).
Auditory hallucinations have often been hypothesized to be associated with disturbances in language processes and semantic memory (e.g., David, 1994, Hoffman et al., 1999 and Kerns et al., 1999). To explore further the relation between hallucinations and memory, in the present study we examined the relation between hallucinations and episodic memory, working memory, and attention/concentration (in the form of immediate auditory memory). To explore whether hallucinations are associated specifically with semantic memory, we administered measures of both verbal and non-verbal memory. A growing body of theorizing and empirical research has begun to implicate several cognitive processes in the development of delusions. For example, there is evidence linking delusions to a “jumping to conclusions” reasoning bias and to a self-serving attributional bias (e.g., Garety and Freeman, 1999 and Bentall et al., 2001). Although the present study was not designed to test hypotheses concerning the relations between delusions and cognitive biases and misattributions, we believe it can advance our understanding of delusions in at least two ways. First, to the degree that delusions turn out to have different cognitive correlates than do hallucinations, it provides support for the utility of developing separate models to explain delusions and hallucinations rather than merely searching for the cause of positive symptoms. Second, because cognitive disturbances, such as executive functioning and working memory deficits, are regularly found to be associated with schizophrenia (e.g., Park and Lee, 2003) as well as with some specific symptoms, such as formal thought disorder (e.g., Kerns and Berenbaum, 2002), it is important to explore whether such deficits also play a role in the development of delusions.
Delusion ratings ranged from one to six (M = 2.2; S.D. = 1.7). Hallucination ratings ranged from one to seven (M = 2.1; S.D. = 1.9). Hallucinations and delusions were significantly correlated, r = 0.62, P < 0.001. Information concerning other symptoms can be found in the accompanying reports.
Associations were measured using Spearman rank order correlations since many of the score distributions were skewed. The associations between the symptom and cognitive measures are presented in Table 1. Hallucinations were significantly negatively correlated with performance on the episodic memory task, and there was a trend for hallucinations to also be negatively correlated with performance on the working memory task. Hallucinations were significantly more strongly correlated with episodic memory task performance than were delusions, z = 2.52, P < 0.01, and there was a trend for hallucinations to be more strongly correlated with working memory performance than were delusions, z = 1.38, P < 0.09. There were trends for hallucinations to be more strongly correlated with episodic memory performance than with fluency performance and digit span performance, z = 1.61, P < 0.06 and z = 1.33, P < 0.10, respectively.
