To develop reliable, valid, and efficient measures of obsessive-compulsive disorder (OCD) severity, comorbid depressionseverity, and total electrical energy delivered (TEED) by deep brain stimulation (DBS), we trained and compared random forestsregression models in a clinical trial of participants receiving DBS for refractory OCD. Six participants were recorded during open-endedinterviews at pre- and post-surgery baselines and then at 3-month intervals following DBS activation. Ground-truth severity wasassessed 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. Mixed-effects random forest regression with Shapley feature reduction stronglypredicted severity of OCD, comorbid depression, and total electrical energy delivered by the DBS electrodes (intraclass correlation,ICC, = 0.83, 0.87, and 0.81, respectively. When random effects were omitted from the regression, predictive power decreased tomoderate for severity of OCD and comorbid depression and remained comparable for total electrical energy delivered (ICC = 0.60,0.68, and 0.83, respectively). Multimodal measures of behavior outperformed ones from single modalities. Feature selection achievedlarge decreases in features and corresponding increases in prediction. The approach could contribute to closed-loop DBS that wouldautomatically titrate DBS based on affect measures

Modeling of human emotion is a challenging problem that can require multiple signals types, as well as contextual information that has been obtained over time. Considering this, in this paper we present our approach, based on physiological signals, to the Emotion Physiology and Experience Collaboration (EPIC) challenge at Affective Computing & Intelligent Interaction (ACII), 2023. In total there are four scenarios that we model: 1) across time; 2) across subjects; 3) across elicitor; and 4) across version. To tackle this challenge, we propose to use a physiological fusion-based approach to solve each scenario. Along with this, we give a detailed analysis of the evaluated physiological signals and personalized predictions for each subject are shown. Our proposed approach shows encouraging results with the lowest root mean square error achieved for scenario 4 (across version) for both valence and arousal on the challenge test set.