Abstract
Recent work has suggested that disorganised speech might be a powerful predictor of later psychotic illness in clinical high risk subjects. To that end, several automated measures to quantify disorganisation of transcribed speech have been proposed. However, it remains unclear which measures are most strongly associated with psychosis, how different measures are related to each other and what the best strategies are to collect speech data from participants. Here, we assessed whether twelve automated Natural Language Processing markers could differentiate transcribed speech excerpts from subjects at clinical high risk for psychosis, first episode psychosis patients and healthy control subjects (total N = 54). In-line with previous work, several measures showed significant differences between groups, including semantic coherence, speech graph connectivity and a measure of whether speech was on-topic, the latter of which outperformed the related measure of tangentiality. Most NLP measures examined were only weakly related to each other, suggesting they provide complementary information. Finally, we compared the ability of transcribed speech generated using different tasks to differentiate the groups. Speech generated from picture descriptions of the Thematic Apperception Test and a story re-telling task outperformed free speech, suggesting that choice of speech generation method may be an important consideration. Overall, quantitative speech markers represent a promising direction for future clinical applications.
Highlights
Psychotic disorders typically develop at the end of adolescence or in early adulthood, following a clinical high risk (CHR-P) phase
Previous work has identified a number of clinical, cognitive, neuroimaging and peripheral blood measures that are associated with transition to psychosis in CHR-P subjects [1,2,3,4]
If measures: one potentially related to the repetitiveness of speech, motivated by prior evidence that perseverance is a component of thought disorder [20], and another of whether a participant’s speech was ‘on-topic’, which is related to tangentiality [8] and similar to measures previously employed by [8, 23]
Summary
Psychotic disorders typically develop at the end of adolescence or in early adulthood, following a clinical high risk (CHR-P) phase. Elvevåg et al used LSA to calculate the semantic coherence between adjacent words, the tangentiality of an individual’s speech, i.e. how likely it was to diverge off-topic over time, and semantic similarity between speech excerpts from different participants. Later work extended these approaches [6, 9], for example, to use new, state-of-the-art word and sentence embedding methods to obtain vectors from words and sentences, instead of LSA [9]. Speech graph connectivity was significantly reduced in patients with schizophrenia compared to healthy control subjects [11]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
