Abstract
The detection of changes in mental states such as those caused by psychoactive drugs relies on clinical assessments that are inherently subjective. Automated speech analysis may represent a novel method to detect objective markers, which could help improve the characterization of these mental states. In this study, we employed computer-extracted speech features from multiple domains (acoustic, semantic, and psycholinguistic) to assess mental states after controlled administration of 3,4-methylenedioxymethamphetamine (MDMA) and intranasal oxytocin. The training/validation set comprised within-participants data from 31 healthy adults who, over four sessions, were administered MDMA (0.75, 1.5 mg/kg), oxytocin (20 IU), and placebo in randomized, double-blind fashion. Participants completed two 5-min speech tasks during peak drug effects. Analyses included group-level comparisons of drug conditions and estimation of classification at the individual level within this dataset and on two independent datasets. Promising classification results were obtained to detect drug conditions, achieving cross-validated accuracies of up to 87% in training/validation and 92% in the independent datasets, suggesting that the detected patterns of speech variability are associated with drug consumption. Specifically, we found that oxytocin seems to be mostly driven by changes in emotion and prosody, which are mainly captured by acoustic features. In contrast, mental states driven by MDMA consumption appear to manifest in multiple domains of speech. Furthermore, we find that the experimental task has an effect on the speech response within these mental states, which can be attributed to presence or absence of an interaction with another individual. These results represent a proof-of-concept application of the potential of speech to provide an objective measurement of mental states elicited during intoxication.
Highlights
In recent years, psychiatry researchers have endeavored to identify alternative, objective evaluations to aid subjective clinical assessments and diagnoses [1]
We aimed to test the following hypotheses: (i) each drug condition had a unique signature in speech, spanning different domains such as acoustics and content; (ii) the higher the dose of MDMA, the greater the associated changes in speech were; (iii) participants during the monologue could express their emotions more freely given the fact that they were alone in the room during the task; (iv) the trained models would generalize well in the independent validation datasets
After generating features based on this representation, we evaluated the following classification tasks: placebo vs. each active drug condition (MDMA 0.75 mg/kg; MDMA 1.5 mg/kg, and oxytocin), and MDMA 0.75 mg/kg vs. MDMA 1.5 mg/kg for the Description and Monologue tasks individually
Summary
Psychiatry researchers have endeavored to identify alternative, objective evaluations to aid subjective clinical assessments and diagnoses [1]. Speech assessment was largely reliant on clinical observation, manual coding, or word counting methods (e.g. see [3]) These approaches, while providing important information, have limitations in objectivity and how comprehensively they can assess this nuanced, complex behavior e.g. acoustic components. As a complement to existing methods, recent rapid developments in computerized natural language processing [4] provide increasingly sophisticated automated methods to quantitatively characterize speech and investigate mental states based on the features extracted These methods are routinely used in industry for the purpose of speech recognition [5], chatbots and conversation agents [6], and recommender systems [7] among others. Whether they could aid research and practice in psychiatry is only beginning to be explored in the context of simulated psychiatric evaluations (e.g. see [1, 8, 9]) or the analysis of alternative ways of communication such as social media (e.g. see [10,11,12])
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call