Home-based transcranial direct current stimulation (tDCS) in major depressive disorder: Enhanced network synchronization with active relative to sham and deep learning-based predictors of remission

Abstract

Aim

To investigate neural oscillatory networks in major depressive disorder (MDD), effects of home-based transcranial direct current stimulation (tDCS) treatment, and predictors of treatment remission.

Methods

In a randomized controlled trial, EEG data were acquired from 21 MDD participants (16 women, mean age 36.63 ± 9.71 years) with moderate to severe depressive episodes (mean HAMD score 18.42 ± 1.80). Participants were randomized to active (n = 11) or sham tDCS (n = 8). Home-based tDCS treatment was administered for 10 weeks, with 5 sessions per week for 3 weeks, then 3 sessions per week for 7 weeks. Active tDCS was 2 mA, and sham tDCS was 0 mA with brief ramp-up/down periods. Clinical remission was defined as HAMD score ≤ 7. Resting-state EEG data were collected at baseline and at the 10-week end of treatment using a portable 4-channel EEG device. EEG band power and functional connectivity (phase locking value, PLV) were analyzed. Deep learning identified predictors of treatment remission from baseline PLV features.

Results

The active tDCS group showed higher gamma PLV in frontal and temporal regions compared to the sham group. Positive correlations between changes in delta, theta, alpha, and beta PLV and depression improvement were observed in the active group. Combining PLV features from theta, alpha, and beta achieved the highest treatment remission prediction accuracy: 71.94 % (sensitivity 52.88 %, specificity 83.06 %).

Conclusions

Synchronized brain activity in gamma PLV may be a mechanism of active tDCS. Baseline resting-state EEG could predict treatment remission. Home-based EEG measures are feasible and useful predictors of clinical outcomes.