Automated Detection of Enhanced DBS Device Settings

Area of Science:

• Neuroscience
• Psychiatry
• Computer Science

Background:

• Deep brain stimulation (DBS) targeting the ventral striatum (VS) is a recognized therapy for severe, refractory obsessive-compulsive disorder (OCD).
• Treatment efficacy is often linked to optimal parameter settings, historically identified via subjective clinician assessments of patient mirth responses.
• The subjective nature of these assessments presents challenges in standardization and objective evaluation.

Purpose of the Study:

• To introduce and validate an objective method for measuring mirth responses during DBS for OCD.
• To assess the feasibility of using Automatic Facial Affect Recognition (AFAR) for quantifying affective changes related to DBS parameter adjustments.

Main Methods:

• A patient undergoing DBS for OCD was longitudinally assessed.
• Automatic Facial Affect Recognition (AFAR) technology was employed to analyze facial expressions.
• Statistical analyses and machine learning models (SVM, XGBoost) were used to compare pre- and post-DBS adjustment states.

Main Results:

• A significant increase in positive affect was observed after DBS parameter adjustments.
• Machine learning models successfully differentiated pre- and post-adjustment facial appearances with an F1 score of 0.76.
• These findings indicate the potential for objective quantification of treatment-related affective responses.

Conclusions:

• Objective measurement of mirth responses using AFAR is feasible for optimizing DBS in OCD treatment.
• This approach offers a standardized, data-driven method to complement subjective clinical evaluations.
• Further research can explore broader applications of AFAR in neurological and psychiatric interventions.