Cognitive Heterogeneity across Schizophrenia and Bipolar Disorder: A Cluster Analysis of Intellectual Trajectories

• Published in: Journal of the International Neuropsychological Society 2020-10, Vol.26 (9), p.860-872
• Main Authors: Vaskinn, Anja, Haatveit, Beathe, Melle, Ingrid, Andreassen, Ole A., Ueland, Torill, Sundet, Kjetil
• Format: Article
• Language: English
• Subjects: Bipolar disorder; Brain research; Classification; Cluster analysis; Cognition & reasoning; Cognitive ability; Intelligence; Mental disorders; Regular Research; Schizophrenia
• ISSN: 1355-6177; 1469-7661
• DOI: 10.1017/S1355617720000442

Cognitive dysfunction cut across diagnostic categories and is present in both schizophrenia and bipolar disorder, although with considerable heterogeneity in both disorders. This study examined if distinct cognitive subgroups could be identified across schizophrenia and bipolar disorder based on the intellectual trajectory from the premorbid phase to after illness onset. Three hundred and ninety-eight individuals with schizophrenia (n = 223) or bipolar I disorder (n = 175) underwent clinical and neuropsychological assessment. Hierarchical and k-means cluster analyses using premorbid (National Adult Reading Test) and current IQ (Wechsler Abbreviated Scale of Intelligence) estimates were performed for each diagnostic category, and the whole sample collapsed. Resulting clusters were compared on neuropsychological, functional, and clinical variables. Healthy controls (n = 476) were included for analyses of neuropsychological performance. Cluster analyses consistently yielded three clusters: a relatively intact group (36% of whole sample), an intermediate group with mild cognitive impairment (44%), and an impaired group with global deficits (20%). The clusters were validated by multinomial logistic regression and differed significantly for neuropsychological, functional, and clinical measures. The relatively intact group (32% of the schizophrenia sample and 42% of the bipolar sample) performed below healthy controls for speeded neuropsychological tests. Three cognitive clusters were identified across schizophrenia and bipolar disorder using premorbid and current IQ estimates. Groups differed for clinical, functional, and neuropsychological variables, implying their meaningfulness. One-third of the schizophrenia sample belonged to the relatively intact group, highlighting that neuropsychological assessment is needed for the precise characterization of the individual.