Semantic Speech Networks Linked to Formal Thought Disorder in Early Psychosis

Abstract

Background and Hypothesis Mapping a patient’s speech as a network has proved to be a useful way of understanding formal thought disorder in psychosis. However, to date, graph theory tools have not explicitly modelled the semantic content of speech, which is altered in psychosis. Study Design We developed an algorithm, “netts,” to map the semantic content of speech as a network, then applied netts to construct semantic speech networks for a general population sample (N = 436), and a clinical sample comprising patients with first episode psychosis (FEP), people at clinical high risk of psychosis (CHR-P), and healthy controls (total N = 53). Study Results Semantic speech networks from the general population were more connected than size-matched randomized networks, with fewer and larger connected components, reflecting the nonrandom nature of speech. Networks from FEP patients were smaller than from healthy participants, for a picture description task but not a story recall task. For the former task, FEP networks were also more fragmented than those from controls; showing more connected components, which tended to include fewer nodes on average. CHR-P networks showed fragmentation values in-between FEP patients and controls. A clustering analysis suggested that semantic speech networks captured novel signals not already described by existing NLP measures. Network features were also related to negative symptom scores and scores on the Thought and Language Index, although these relationships did not survive correcting for multiple comparisons. Conclusions Overall, these data suggest that semantic networks can enable deeper phenotyping of formal thought disorder in psychosis. Whilst here we focus on network fragmentation, the semantic speech networks created by Netts also contain other, rich information which could be extracted to shed further light on formal thought disorder. We are releasing Netts as an open Python package alongside this manuscript.

Automated analysis of speech as a marker of sub-clinical psychotic experiences

“…These connectedness features were found to be informative about of negative symptoms score, predict symptom severity and schizophrenia diagnosis with 91.67% accuracy and 85% accuracy 6 months in advance ( 30 ). Novel methods have recently been developed combining connectivity and semantic approaches using semantic networks which showed significant differences between first episode patients, healthy control and clinical high-risk groups ( 51 ).…”

“…Although alongside with more subtle symptomology we can assume more subtle alterations in speech and hence a larger required sample size because of smaller effect sizes, the frequency of the investigated phenomena is much higher and the barrier to get access to these populations is much lower (e.g., easier to collect their speech in online, remote settings; easier to recruit) that can lead to bigger and more representative samples. Another solution can be collecting short speech samples in standardized, prompt based-settings [several feasible methods have been proposed by (43,50,51)] or by simply recording clinical interviews. These solutions combined with the application of automated transcription of voice into text (45,46,(68)(69)(70) can reduce the cost of, and accelerate the speed of data collection.…”

Exploring the relationship between frequent cannabis use, vocal changes and language use

“…Most common approaches include the measurement of speech connectivity, which is the way that speech flows and connects to form thoughts and helps evaluate both coherence and syntactic complexity (Mota, Natália B. et al, 2017;Mota, Natália Bezerra et al, 2023;Spencer et al, 2021a) as well as measuring markers of semantic coherence (Bedi et al, 2015;. Evaluating semantic coherence is based on Latent Semantic Analysis (LSA), an approach that captures the distance between words that have been represented in a vectorized space based on their co-occurence in a big text corpus while connectivity is measured applying graph theory on speech, representing all words as nodes and the connection between words as edges (see Methods for more detailed explanation) (Bedi et al, 2015;Morgan et al, 2021;Nettekoven et al, 2023;Spencer et al, 2021b). Former studies utilized these automated markers to identify psychotic disorders, relapse and vulnerability as well as to quantify changes in speechhowever, as did studies did not take cannabis use into account, the effect of cannabis use on these observed changes and classifications remained unexplored.…”