English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Background Language Disorganisation is central to the conceptualization of psychosis. Disruptions in semantic processing have been observed both as a “state”, and a “trait” phenomena in psychotic disorders. Quantification of semantic abnormalities have been improved with recent advances in semantic modeling. The current study applied such computational methods on a word association task, using immediate response to cue words to explore semantic associations. We employed a longitudinal design to investigate semantic relationships during a psychotic episode compared with the same patients after remission six months later, in order to clarify the state-trait status of the semantic variables, and their relationships with clinical symptoms. We hypothesized that semantic distance would be significantly greater in patients than controls at baseline, and would decrease upon follow-up. Methods A continued word association task (WAT) was employed to elicit three associations per cue from a set of 200 cue-words. The set of cues were previously established as being representative of words in general speech, in terms of valence, concreteness and part-of-speech composition. The task was administered to 47 patients with schizophrenia spectrum disorders and 44 matched healthy control participants. Data was collected at two time points, at baseline when patients were actively psychotic and then at 6-months follow-up. In addition, extensive clinical and cognitive measures were collected at both time points. Patterns of word associations were explored using vector representations, derived from Word2Vec, that encompass semantic meaning. Semantic distance of each cue-response pairing is defined using the cosine angle of their vectors. Changes in semantic distance were further examined on their correlation with symptom change over time. Results There was a significant interaction between group and time point on semantic distance (F = 6.865, p = 0.009), where measures of the semantic distance of patients’ responses were significantly greater than healthy controls at both time-points (p &amp;lt; 0.001).There is a significant time effect: the semantic distance reduced significantly over time (p &amp;lt; 0.001). Within the patient group, a change in semantic distance was correlated with symptom change over time, specifically with general psychopathology (p =0.024), depressive (p = 0.046) and manic symptoms (p &amp;lt; 0.01). Discussion Measures of semantic distance were significantly greater in patients both at baseline during a psychotic episode, and at follow-up upon clinical remission. There is a significant but not full normalization of semantic distance upon remission. Increase in semantic distance is therefore both a state and a trait marker in psychosis. We have employed a novel technique to quantify semantic distance of a word association task using Word2Vec to generate vector representations of responses in a high-dimensional semantic space. The findings illustrate the feasibility of applying Word2Vec to a word association task to detect subtle changes in language. Subsequent research possibilities using this approach includes exploration of the semantic content of responses, by grouping similar meaning responses into conceptual clusters, and its correlation with symptom change.

T55. DETECTING SEMANTIC DISTANCE ABNORMALITIES IN PSYCHOSIS: QUANTIFICATION OF WORD ASSOCIATIONS USING SEMANTIC SPACE MODELING