Semester of Graduation

Spring 2023

Degree

Master of Arts (MA)

Department

Psychology

Document Type

Thesis

Abstract

Psychomotor retardation (PMR) is a diagnostic criterion that reflects a gross slowing of movement, speech, and thought. More so, PMR is a transdiagnostic, pervasive source of impairment and distress that can be observed across disorders, such as Major Depressive Disorder (MDD), Schizophrenia (SZ), and Bipolar Disorder (BD). In clinical settings, PMR is associated with more severe psychopathology and is used to both inform treatment and predict long-term treatment outcomes. However, despite being clinically meaningful, PMR is poorly understood and is often conceptualized as a motor disturbance that happens in isolation from, rather than concurrently with, higher-order cognitive-emotional processes – such as rumination. Although many psychological phenomena are thought to be relatively stable over time and context, there is reason to suspect they are more dynamic. Spoken language is a dynamic medium sensitive to cognitive and affective changes that can be measured and quantified through increasingly sophisticated, sensitive, and non-intrusive methods – including acoustic analysis and natural language processing (NLP). Therefore, by analyzing archival audio samples from two prior projects, the proposed study explored PMR and rumination through sentence level NLP and acoustic analyses. Our findings indicate that PMR and rumination can be predicted at a better than chance accuracy level using the acoustic properties of speech. Similar to previous findings, we found that the combination of diverse speech samples created more generalizable models. However, contrary to expectations, we did not find that the addition of ruminatory features meaningfully improved model performance. Future studies can expand on this study by exploring how, and to what degree, ruminated thought (as manifest through speech) interacts with the moment to moment acoustic properties of speech.

Date

4-4-2023

Committee Chair

Cohen, Alex

DOI

10.31390/gradschool_theses.5734

Download

Available for download on Friday, April 03, 2026

Share

COinS