Longitudinal changes in semantic categorization performance after symptomatic remission from first-episode psychosis: A 3-year follow-up study

Abstract Semantic categorization abnormalities have been observed in schizophrenia, but studies have rarely focused on the longitudinal trajectory. In this study, we consider semantic performance and the relationship with symptomatic changes during recovery from a first-episode of schizophrenia over a period of 3 years. Thirty-seven first-episode patients with schizophrenia were compared to thirty-seven matched controls in a categorization task. Patients were assessed at first episode, after clinical stabilization, and annually for the subsequent 3 years. In the task, participants indicated whether a word belonged to a given category. Each category contained words of varying degrees of semantic relatedness: typical, atypical, borderline, related-but-outside, and unrelated. Reaction times and proportion of ‘yes’ responses were analyzed. At first assessment, semantic categorization abnormalities were observed in first-episode patients. Patients assigned more semantically-dissimilar words to the categories than controls. As patients stabilized from acute states, their semantic categorization performance improved and then remained stable throughout the entire follow up period of 3 years. Interestingly, semantic performance deficits, particularly a diminished typicality effect, correlated with negative symptoms in the initial episode, but not at stabilization when symptoms subsided. No significant associations between positive and negative symptoms, or pre-defined categorization measures were identified. The data demonstrated semantic memory abnormalities in first-episode schizophrenia. However, an improvement of semantic categorization performance was observed in stabilized schizophrenia patients. Overall, the data are suggestive of a state effect in semantic abnormalities rather than a trait effect. The correlation between degree of impairment and symptoms may explain previous inconsistent findings.