Smartphone-based facial landmark analysis for objective monitoring during rTMS in treatment-resistant depression: A...

Objective, scalable markers for monitoring treatment response in treatment-resistant depression remain limited. This pilot study evaluated the feasibility of repeated smartphone-based facial landmark analysis during repetitive transcranial magnetic stimulation (rTMS) and explored associations with clinical change.

Six patients with treatment-resistant depression underwent a standardized 20-session rTMS course. At each session, smartphone videos were recorded during five facial expressions: neutral, smile, surprised, angry, and sad. Automated facial landmark tracking was used to extract facial motion metrics. Hamilton Depression Rating Scale and Hamilton Anxiety Rating Scale scores were assessed at baseline and after treatment.

Depression and anxiety scores decreased from baseline to end-of-treatment in this uncontrolled pilot cohort. Facial analyses showed nominal expression-dependent longitudinal changes, most consistently during the sad expression. Exploratory correlations identified preliminary nominal associations between selected facial motion changes and symptom-score reductions; however, these findings did not survive correction for multiple comparisons.

Repeated smartphone-based facial landmark analysis during rTMS was technically and procedurally feasible and generated longitudinal facial motion metrics. Larger controlled studies are needed to determine whether these changes reflect treatment-related effects, nonspecific longitudinal factors, or clinically meaningful behavioral signals.