The association between semantic categorization deficit and symptoms in first-episode schizophrenia

Abstract

BACKGROUND: Semantic memory was defined as a store of culturally shared general knowledge about the meaning of the world, including words and objects, as well as their categorization. Previous studies had found that the deficits are related to some variables (such as age of onset) and symptomatology (such as delusion and formal thought disorder) in chronic patients. However, the relationship might have undermined by other factors such as medication effect in chronic/mixed samples. Importantly, further exploration is essential as clinical and cognitive correlates of the deficit could help elucidate the underlying mechanisms of the disorder. We therefore aim to explore the clinical and cognitive correlates of semantic memory deficit in first-episode schizophrenia patients. METHODS: Semantic memory was assessed using the categorization task in 37 first-episode schizophrenia patients. They were requested to make a category decision (yes/no) as to whether an exemplar word belonged to a category. Exemplars consisted of five semantic relatedness conditions, namely 'typical', 'atypical', 'borderline', 'related' and 'unrelated'. Categorization performance was further operationalized using the reaction time data as typicality effect, false-relatedness effect, borderline peak, borderline shift and overall speed. Data on proportion of yes response in each semantic relatedness condition was also measured. Other than that, positive and negative symptoms, as well as cognitive deficits (executive function, verbal fluency, and sustained attention, verbal and visual memory) were assessed during first episode, after clinical stabilization, and each year for the following three years. RESULTS: Correlation analysis demonstrated that different dimensions of categorization performance are related to different symptoms and cognitive dimensions. At presentation, negative symptoms was significantly correlated with typicality effect (r=-.43, P=.008), borderline peak (r=-.44, P=.007), overall speed (r=.38, P=.02), and over inclusiveness (% of yes response in the unrelated condition) (r=.37, P=.03). Visual memory was related to borderline shift (r=-.48, P=.002) and verbal fluency with overall speed (r=-.32, P=.05). At stabilization, positive symptoms was significantly associated with over-inclusiveness (r=.44, P=.006). At year 3, while there was a strong correlation between positive symptoms and borderline shift (r=-.52, P<.001), negative symptoms was related to borderline peak (r=.33, P=.04). DISCUSSION: Interestingly, positive symptoms were unrelated to semantic memory deficit in first-episode. However, it was strongly related to borderline shift at year 3. The current study had suggested a potential relationship between symptom formation and impairments in semantic memory. Future studies could look at specific symptomatology by separate clinical measure such as formal thought disorder in patients with schizophrenia.