S57. CHARACTERIZING COGNITIVE HETEROGENEITY IN INDIVIDUALS WITH A SCHIZOPHRENIA SPECTRUM DISORDER: A CLUSTER ANALYTIC APPROACH

BACKGROUND: Cognitive impairment is a core symptom of patients with schizophrenia, yet substantial heterogeneity in the degree of impairment exists within and between diagnoses of the schizophrenia spectrum, leading to sub-optimal understanding of the general and individual pathology of schizophrenia. By means of a cluster-analytic approach, individuals are grouped based on profiles of traits regardless of specific diagnoses. As this approach has already proven itself in decreasing cognitive heterogeneity within specific psychotic disorders, we investigated the structure of cognitive heterogeneity within a broader spectrum of psychotic disorders by including a large sample of schizophrenia patients and healthy individuals. METHODS: An agglomerative hierarchical clustering analysis was performed using cognitive data (patients n = 324, healthy controls n = 40). All patients were diagnosed with a DSM-IV-TR diagnosis of schizophrenia (SZ), schizoaffective (SZA) or schizophreniform (SZP) disorder or another psychotic disorder/psychosis not otherwise specified (PNOS). To measure cognitive function, the composite Z-score was calculated, based on six cognitive domains as defined by the Brief Assessment of Cognition in Schizophrenia (BACS). RESULTS: Four clusters emerged from the clustering analysis, gradually and significantly distinguishable in cognitive impairment. The first cluster (n=50) contained relatively intact individuals, the second cluster (n=122) consisted of individuals with minimal to mild cognitive impairment, the third cluster (n=114) contained individuals with moderate cognitive impairment and individuals with most severe cognitive impairment were placed in the fourth cluster (n=78). All six cognitive domains show similar differences between clusters. While symptom severity and educational attainment show the same classification patterns as the neurocognitive profiles do, cluster assignment was not affected by illness duration and antipsychotic intake. DISCUSSION: Results confirm that data-driven formed groups based on the clustering of neurocognitive profiles contain less cognitive heterogeneity compared to predefined diagnostic groups. Our results demonstrate that clustering patients with a schizophrenia spectrum disorder based on their neurocognitive profile decreases cognitive heterogeneity within the formed groups, creating more meaningful groupings and findings accentuate a cognitive continuum irrespective of a specific diagnosis, providing insight into the underlying pathology of clinical manifestation of schizophrenia.