Automated analysis of free speech predicts psychosis onset in high-risk youths

By Gillinder Bedi, Facundo Carrillo, Guillermo A. Cecchi, Diego Fernández Slezak, Mariano Sigman, N. Mota, S. Ribeiro, Daniel C. Javitt, Mauro Copelli, Cheryl Corcoran. npj Schizophrenia, 2015

617 citations recorded by the source corpus.

Background/Objectives:Psychiatry lacks the objective clinical tests routinely used in other specializations. Novel computerized methods to characterize complex behaviors such as speech could be used to identify and predict psychiatric illness in individuals.AIMS:In this proof-of-principle study, our aim was to test automated speech analyses combined with Machine Learning to predict later psychosis onset in youths at clinical high-risk (CHR) for psychosis.Methods:Thirty-four CHR youths (11 females) had baseline interviews and were assessed quarterly for up to 2.5 years; five transitioned to psychosis. Using automated analysis, transcripts of interviews were evaluated for semantic and syntactic features predicting later psychosis onset. Speech features were fed into a convex hull classification algorithm with leave-one-subject-out cross-validation to assess their predictive value for psychosis outcome. The canonical correlation between the speech features and prodromal symptom ratings was computed.Results:Derived speech features included a Latent Semantic Analysis measure of semantic coherence and two syntactic markers of speech complexity: maximum phrase length and use of determiners (e.g., which). These speech features predicted later psychosis development with 100% accuracy, outperforming classification from clinical interviews. Speech features were significantly correlated with prodromal symptoms.Conclusions:Findings support the utility of automated speech analysis to measure subtle, clinically relevant mental state changes in emergent psychosis. Recent developments in computer science, including natural language processing, could provide the foundation for future development of objective clinical tests for psychiatry.

Subfields: Experimental and Cognitive Psychology, Psychiatry and Mental health, Social Psychology

View Paper: https://www.semanticscholar.org/paper/7d5dd4bdf88a64cc9c88b5dc8e2c05b6c232faf4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Experimental and Cognitive Psychology

In Psychiatry and Mental health

In Social Psychology

Citations in the corpus

MamaBot: a System based on ML and NLP for supporting Women and Families during Pregnancy (2018, positive): “Inspired by a recent study [20] that predicted psychosis with 100% accuracy, the engineers behind NeuroLex 5 built their system to screen for schizophrenia by analyzing patients’ speech patterns.”
Machine learning in medicine: a practical introduction (2019, background): “Recently, an automated analysis of free-speech collected during in-person interviews resulted in the ability to predict transition to psychosis with perfect accuracy in a group of high-risk youths [9].”
Prediction of psychosis across protocols and risk cohorts using automated language analysis (2018, methodology): “…had an intra-protocol accuracy of 83% in the training dataset, and a cross-protocol accuracy of 79% when applied to transcripts from a second independent CHR cohort (test dataset) 15 , demonstrating signiﬁcant transfer of predictability, despite disparate methods of speech elicitation 8,15 .”
AI in mental health. (2020, methodology): “An early research instance used semantic coherence and syntactic complexity to predict later psychosis development with 100% accuracy ( n = 34) [22].”
New Tools for New Research in Psychiatry: A Scalable and Customizable Platform to Empower Data Driven Smartphone Research (2016, background): “There is increasing evidence of the clinical value of voice data in schizophrenia in predicting conversation from prodromal state [30], and there is also much literature on the value of correlating such data with changes in symptoms [31,32].”
Ethical and regulatory challenges of AI technologies in healthcare: A narrative review (2024, background): “Further, surprising speech pattern analysis with AI has been shown to predict psychotic occurrences and identify features of neurological diseases like Parkinson’s disease [37,38].”
Your Robot Therapist Will See You Now: Ethical Implications of Embodied Artificial Intelligence in Psychiatry, Psychology, and Psychotherapy (2018, neutral): “We have also excluded applications that may have a virtual or robotic interface but do not employ AI, such as telemedicine therapy (for further scholarship on this topic, please see [12-19]).”
Artificial intelligence and machine learning in precision and genomic medicine (2021, methodology): “AI is presently being utilized to auto-mate data retrieved from sources, summaries EHRs, or hand-written physician notes, combine health records, and store data on a cloud scale [14–19].”
Neural Net Models for Open-Domain Discourse Coherence (2016, background): “…parts of a text are linked into a coherent whole) is essential for tasks like summarization (Barzilay and McKeown, 2005), text planning (Hovy, 1988; Marcu, 1997) question-answering (Verberne et al., 2007), and even applications like psychiatric diagnosis (Elvev˚ag et al., 2007; Bedi et al., 2015).”
A Review of the Role of Artificial Intelligence in Healthcare (2023, background): “AI tools analyze speech patterns to forecast psychotic occurrences, and recognize and screen the features of neurological diseases such as Parkinson’s disease [53,54].”
Automated assessment of psychiatric disorders using speech: A systematic review (2019, background): “…determining whether a patient meets criteria for unipolar depression or bipolar disorder 31 ), help detect risk for chronic psychiatric disorders, 32 psychiatric episodes, 33 or suicidal behavior 29 ; and over time learn to predict the best treatment given multimodal FIGURE 1 How machine…”
Machine Learning and Natural Language Processing in Mental Health: Systematic Review (2021, neutral): “Each community has a central node: Perlis et al [55] are cited in 7 other papers, Jackson et al [73] are cited in 4 other papers, Carrell et al [74] and Afshar et al [75] are cited in 2 other papers, and Bedi et al [39] are cited in 2 other papers.”