Can an Unobtrusive, Multimodal Mixed-Effects Regressor Based on
Open-Ended Interviews Predict OCD Severity?
Abstract
Reliable, valid, efficient measurement of symptom severity in
internalizing disorders is critical to gauge treatment response.
Self-report and clinical interview are subjective and difficult to
standardize, impose patient burden, and lack granularity. We tested the
hypothesis that comprehensive sampling of audio and visual modalities
during open-ended interviews can reveal severity of obsessive-compulsive
disorder (OCD) and comorbid depression. Participants were six patients
with chronic, refractory OCD that were treated with deep brain
stimulation (DBS). They were recorded during open-ended interviews at
pre- and post-surgery baselines and at 3-month intervals following
activation of the DBS. Ground-truth severity was assessed by clinical
interview and self-report. Visual and auditory modalities included
facial action units, head and facial landmarks, speech behavior and
content, and voice acoustics. Using mixed-effects random forest
regression with Shapley feature reduction strongly predicted severity of
OCD, severity of comorbid depression, and total electrical energy
delivered by the DBS electrodes (ICC = 0.83, 0.87, and 0.81,
respectively). Multimodal measures of behavior outperformed ones from
single modalities. The approach could contribute to closed-loop DBS that
would automatically titrate DBS based on affect measures.