Document Type

Essay

Publication Date

5-14-2026

Abstract

Psychosis is difficult to identify in early stages; however, changes in speech and language offer insights and indicators of emerging symptomatology. This paper examines the potential for Natural Language Processing (NLP) and other computational linguistic techniques for improving the predicition of psychosis disorders through automated analysis of speech. Existing research shows that lingustic markers that are aligned with Formal Thought Disorder, including reduced semantic coherence, poverty of content, tangentiality, reduced syntactic complexity, and pragmatic deficits, can help to distinguish between individuals who are experiencing or at risk for psychosis from healthy controls. Traditional clinical assessments may have difficulty identifying and quantifying linguistic characteristics operating as symptom indicators, but this paper reviews existing research on NLP methods like latent semantic analysis, word embeddings, speech graph analysis, part-of-speech tagging that are able to do so. To demonstrate potential application of these methods in a clinical setting, four speech samples from previous psychosis research were evaluated through both a simulated clinician assessment and a simulated NLP analysis using ChatGPT. The two approaches showed substantial overlap, but the simulated NLP identified several additional linguistic markers and provided additional quantifiable measures of syntactic complexity and semantic coherence. Of course, the simulation has important limitations and does not actually replicate the exact capabilities of specialized NLP software or clinical assessment, but the findings do help to support the potential value of using computational language analysis in tandem with the traditional clinical methods. Utilization of NLP methods in clinical practice could contribute to earlier detection, objective assessment, and longitudinal, measurement based care for those who are at risk of or are already experiencing psychosis.

Comments

Completed as part of Professor Amity Reading's HONR 300A: Introduction to Applied Linguistics.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.