Semantic networks and their disruptions in psychotic disorders

Abstract

Background Loosening of association and cognitive impairments are important core features of psychotic disorders (Bleuler, 1911/1950; Elvevåg & Goldberg, 2000). Over the past decades, research has identified various anomalies in linguistic functions in patients with psychosis. Nevertheless, despite numerous research into linguistic disruptions in psychosis, it remains inconclusive as to whether and how are language disrupted in psychosis, and what is the relationship between language disruptions and the other symptoms observed in psychosis, many of which are related to language. With the development of computational analysis methods, recent studies have attempted to apply these methods into the study of semantic disruptions in psychosis. The current study aimed to investigate the semantic network, its disruptions, and the factors behind these disruptions in patients by combining a continued word association task with semantic network analysis method. Using a longitudinal design, the current study also hopes to investigate whether semantic network changes across time as patients’ mental states change.

Methodology Forty-two Cantonese-speaking patient subjects with psychotic disorders and 40 matching control subjects in Hong Kong have been included in the study. The study assessed the clinical, cognitive, functioning and semantic association profile of patients at baseline recruitment and at a 6-month follow-up. The study also collected cognitive and semantic association data from healthy control at baseline recruitment. The word association data collected using the continued word association task was analysed using an exploratory semantic network analysis method.

Results First, the current study found that semantic networks properties differed between patients and controls. Patients’ network had less positive responses, and a denser network with slightly longer shortest path length, higher clustering coefficient, and lower modularity. Second, the study found that part of semantic network properties changed across time in patients alongside improvement in symptoms and functioning. Patients’ baseline semantic network was also less positive, denser, and with higher clustering coefficient than patients’ network at 6- month.

Discussion This is the first attempt in studying the semantic network differences in Cantonese-speaking psychosis, anomalies in patients’ semantic association was found at an active psychotic state, when compared with controls. An across-time change was also noted in the study. This suggests that some of the semantic anomalies noted perhaps are state-dependent measures, which may change as the symptoms of patients change, while some other anomalies may remain as traits that remain unchanged despite the fluctuation in mental states.