Hostname: page-component-54dcc4c588-nx7b4 Total loading time: 0 Render date: 2025-09-18T04:17:17.602Z Has data issue: false hasContentIssue false
  • English
  • Français

Semantic similarity analysis in speech samples of patients with first-episode bipolar disorder

Published online by Cambridge University Press:  26 August 2025

B. Arslan ,
E. Kizilay  and
E. Bora
B. Arslan*
Affiliation:
Department of Neurosciences
E. Kizilay
Affiliation:
Department of Neurosciences
E. Bora
Affiliation:
Department of Neurosciences Department of Psychiatry, Dokuz Eylul University, Izmir, Türkiye Department of Psychiatry, University of Melbourne and Melbourne Health, Melbourne, Australia
*
*Corresponding author.

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.
Introduction

Individuals with first-episode bipolar disorder (FEBD) may have language abnormalities and sub-threshold formal thought disorder (FTD). Natural language processing (NLP) methods that have been shown to assess FTD in psychosis can be used in the early stages of bipolar disorder (BD).

Objectives

The purpose of this study was to examine the differences between FEBD and healthy controls (HC) utilizing NLP methods.

Methods

Speech samples were collected from 20 FEBD and 20 HC while describing eight Thematic Apperception Test images. The manually transcribed text was then processed using word2vec to generate vectors. The semantic similarity between words was computed utilizing a moving window method to windows ranging in size from 5 to 10. Finally, the average, variance, maximum, and minimum of these similarities were calculated.

Results

All mean similarities (windows of 5 to 10) were significantly higher in FEBD (p< .001, p=0.001, p=0.002, p=0.002, p=0.003, and p=0.004, respectively). Additionally, all variances of similarities were highly significant and were increased in FEBD (all p values< .001). Regarding maximum values, except for the window of 5, all of the remaining windows were significantly higher in FEBD (all p values <.05.).

Conclusions

Our findings indicate that semantic similarity increased in the FEBD group compared to HC. Overall, NLP methods offer an easily applicable approach for assessing FTD in FEBD and discriminating between FEBD and HC.

Disclosure of Interest

None Declared

Information

Type
Abstract
Information
European Psychiatry , Volume 68 , Special Issue S1: Abstracts of the 33rd European Congress of Psychiatry , April 2025 , pp. S263
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of European Psychiatric Association
Submit a response

Comments

No Comments have been published for this article.