BACKGROUND: Simpler and more feasible light therapy protocols, and objective indicators for assessing its effectiveness is lacking. We aimed to evaluate the efficacy of light therapy on subthreshold depression (SD) among college students and explore facial expressions as an objective biomarker across different treatment groups. METHODS: From September 13, 2021, to January 4, 2022, college students with SD were recruited from a university in Hubei Province, randomly assigned to Bright Light Therapy (BLT) group (10,000 lx), Dim Light Therapy (DLT) group (200 lx), or Waiting List Control (WLC) group (no intervention). Self-reported questionnaire and facial expressions were assessed for all participants before and after intervention. Repeated measures ANOVA and logistic regression were conducted to compare baseline and post-intervention differences among three groups. RESULTS: 135 participants were enrolled and 121 participants completed the study. Depression symptom and sleep quality scores significantly decreased in both BLT and DLT groups (P <
0.002), while no significant changes were observed in WLC group. BLT (OR, 4.50
95 % CI, 1.11-18.27
P = 0.035) and DLT group (OR, 4.17
95 % CI, 1.04-16.79
P = 0.045) had higher efficacy rates than WLC group. For facial expressions, DLT group showed significant increases in two happy-related facial action units (AU) including AU14 values (positive, negative and neutral stimuli) and AU26 values (neutral and negative stimuli). BLT group showed a significant decrease in fear-related AU20 values under negative stimuli (P <
0.002). CONCLUSION: Light therapy improves depressive symptoms and sleep quality in individuals with SD, and facial expressions can serve as an objective biomarker to support its effectiveness.