Lower speech connectedness linked to incidence of psychosis in people at clinical high risk

Spencer, Tom; Thompson, Bethany; Oliver, Dominic; Diederen, Kelly; Demjaha, Arsime; Weinstein, Steven J.; Morgan, Sarah E.; Day, Fern; Valmaggia, Lucia; Rutigliano, Grazia; Micheli, Andrea De; Mota, Natália Bezerra; Fusar‐Poli, Paolo; McGuire, Philip

doi:10.1016/j.schres.2020.09.002

Cited by 45 publications

(53 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The maximum similarity (repetition) measure showed no statistically significant group differences. As previously reported by [30], there were significant reductions in speech graph connectivity in FEP patients, in-line with [9], [10].…”

Section: Resultssupporting

confidence: 90%

“…Z-values are given from Mann-Whitney U-tests, with the corresponding P-values in brackets. LCC, LSC, LCCr and LSCr results for the TAT have already been reported by [30]. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.…”

Section: Limitationsmentioning

confidence: 72%

“…Briefly, each unique word in a participant’s response is represented by a node, and directed edges link the words in the order in which they were spoken. We note that prior work has already applied speech graph analysis to our TAT speech excerpts [30], and found significant group differences in speech graph connectivity measures. Here, we compared speech graph connectivity to the other NLP measures described above.…”

Section: Methodsmentioning

confidence: 89%

“…As described by [30], we used the SpeechGraphs software [10] to calculate two measures of graph connectivity: the total number of nodes in the largest connected component (LCC) and the largest strongly connected component (LSC) [9], [10]. The LCC is the largest sub-graph in which all nodes are linked by at least one path.…”

Section: Methodsmentioning

confidence: 99%

“…there is a path from node i to node j , and also from node j to node i ). To control for the number of words spoken, LCC and LSC were calculated for windows of 30 words, which overlapped by 15 words, then averaged across all windows [30]. We also calculated values of LCC and LSC normalised to the equivalent measures from randomised speech graphs (obtained by randomly shuffling the words within each window), to determine how close to randomness the connectedness measures were; denoted LCCr and LSCr [30].…”

Section: Methodsmentioning

confidence: 99%

See 4 more Smart Citations

Assessing psychosis risk using quantitative markers of disorganised speech

Morgan

Diederen

Vértes

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

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 predictive of psychosis-onset, how different measures relate to each other and what the best strategies are to elicit disorganised speech from participants. Here, we assessed the ability of twelve automated Natural Language Processing markers to differentiate transcribed speech excerpts from subjects at clinical high risk for psychosis (N=25), first episode psychosis patients (N=16) and healthy control subjects (N=13; N=54 in total). In-line with previous work, several of these measures showed significant differences between groups, including semantic coherence and speech graph connectivity. We also proposed two additional measures of repetition and whether speech was on topic, the latter of which exhibited significant group differences and outperformed the prior, related measure of tangentiality. Most measures examined were only weakly related to each other, suggesting they provide complementary information and that combining different measures could provide additional power to predict the onset of psychotic illness. 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 diagnostic applications for psychosis risk.

show abstract

Section: Resultssupporting

confidence: 90%

Section: Limitationsmentioning

confidence: 72%

Section: Methodsmentioning

confidence: 89%