Online Speech Assessment of the Psychotic Spectrum: Exploring the relationship between overlapping acoustic markers of Schizotypy, Depression and Anxiety.

Abstract

Background Remote assessment of acoustic alterations in speech holds promise to increase scalability and validity in research across the psychosis spectrum. A feasible first step in establishing a procedure for online assessments is to assess acoustic alterations in psychometric schizotypy. However, to date, the complex relationship between alterations in speech related to schizotypy and those related to comorbid conditions such as symptoms of depression and anxiety has not been investigated. This study tested whether (1) depression, generalized anxiety and high psychometric schizotypy have similar voice characteristics, (2) which acoustic markers of online collected speech are the strongest predictors of psychometric schizotypy, (3) whether controlling for generalized anxiety and depression symptoms can improve the prediction of schizotypy. Methods We collected cross-sectional, online-recorded speech data from 441 participants, assessing demographics, symptoms of depression, generalized anxiety and psychometric schizotypy. Results Speech samples collected online could predict psychometric schizotypy, depression, and anxiety symptoms, however, most influential features of these models largely overlapped. The predictive power of speech marker-based models of schizotypy significantly improved after controlling for symptoms of depression and generalized anxiety (from R2 = 0.296 to R2 = 0. 436). Conclusions Acoustic features of online collected speech are predictive of psychometric schizotypy as well as generalized anxiety and depression symptoms. The acoustic characteristics of schizotypy, depression and anxiety symptoms significantly overlap. Speech models that are designed to predict schizotypy or symptoms of the schizophrenia spectrum might therefore benefit from controlling for symptoms of depression and anxiety.

Towards a scalable approach to assess speech organization across the psychosis-spectrum -online assessment in conjunction with automated transcription and extraction of speech measures

“…Testing the usability of these measures in online settings with short speech samples would complement the work presented here. Furthermore, whilst our study focused on text-based language markers, numerous findings support the use of vocal, acoustic markers to assess psychotic symptoms from speech [10,[42][43][44], which were not included here. Other findings suggest the potential of other, remotely assessable markers coming from visual cues or passive sensing data of smart devices for monitoring or assessment [10,[42][43][44][45][46][47].…”

“…Furthermore, whilst our study focused on text-based language markers, numerous findings support the use of vocal, acoustic markers to assess psychotic symptoms from speech [10,[42][43][44], which were not included here. Other findings suggest the potential of other, remotely assessable markers coming from visual cues or passive sensing data of smart devices for monitoring or assessment [10,[42][43][44][45][46][47]. Testing these markers in online settings, especially in additive nature to text-based markers would be another important step to potentially improve the power and precision of automated, online assessment of the psychotic spectrum.…”

“…For example, the performance of speech-based machine learning models in psychiatry and neurology showed a negative correlation with the sample sizes of the study [17]. Utilizing standardized prompts and online assessment can mitigate the challenge of insufficient training and sample data in scientific studies and lead to more reliable, replicable, and realistic findings [17,18].…”

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

“…Furthermore, studies have demonstrated that individuals with high schizotypy exhibit alterations in acoustic features of speech like reduced speech variability, expressiveness and atypical pitch patterns compared to those with lower schizotypy scores (20,67). In remotely collected speech samples, schizotypy scores were positively associated with acoustic features like Mel frequency cepstral coefficients (MFCC), loudness parameters, Hammerberg -index, Spectral flux, and slope measures (68)(69)(70), where changes in loudness parameters were uniquely associated with schizotypy (compared to features associated with anxiety and depression symptoms). Importantly, analytical and modelling approaches that have been used to discriminate clinical psychosis from healthy controls could be successfully applied to discriminate between low and high level of schizotypy, reaching 69-88% of accuracy (68)(69)(70).…”

“…In remotely collected speech samples, schizotypy scores were positively associated with acoustic features like Mel frequency cepstral coefficients (MFCC), loudness parameters, Hammerberg -index, Spectral flux, and slope measures (68)(69)(70), where changes in loudness parameters were uniquely associated with schizotypy (compared to features associated with anxiety and depression symptoms). Importantly, analytical and modelling approaches that have been used to discriminate clinical psychosis from healthy controls could be successfully applied to discriminate between low and high level of schizotypy, reaching 69-88% of accuracy (68)(69)(70). These findings suggest that speech analysis techniques hold promise in capturing subtle speech abnormalities associated with subclinical psychotic symptoms.…”