Longitudinal coupling between depressive symptoms and divided attention during transcranial magnetic stimulation in treatment-resistant depression

Background – Cognitive dysfunction is a core feature of major depressive disorder. During transcranial magnetic stimulation (TMS) for treatment-resistant depression (TRD), the temporal relationship between symptom changes and cognitive performance remains unclear. We examined depressive symptoms and cognitive trajectories during TMS and tested whether time-varying cognitive change was associated with symptom trajectories. Methods – This retrospective naturalistic study included 89 patients with TRD who completed 20 sessions of left dorsolateral prefrontal cortex TMS (high-frequency repetitive TMS or intermittent theta-burst stimulation). Depression severity (Hamilton Depression Rating Scale [HAMD-17]) and computerized attentional control measures from the Test of Attentional Performance were assessed at baseline, day 10, and day 20. Generalized estimating equations examined time effects and time-adjusted symptom-cognition associations. Exploratory analyses evaluated whether early cognitive change was associated with subsequent symptom improvement. Multiple comparisons were controlled. Results – Depressive symptoms decreased significantly (p < 0.001), with response and remission rates of 65.2% and 51.7%. Several attentional indices showed significant time effects; however, after false discovery rate correction, only visual divided-attention commission errors were independently associated with depressive severity beyond time effects. Person-mean centering indicated that this association was driven by within-person change (B = 0.209, p < 0.001), not between-person differences. Greater early reduction in visual divided-attention commission errors was associated with larger overall symptom improvement (B = 0.172, p = 0.016). Conclusion – Within-person changes in divided attention tracked depressive symptom trajectories during TMS. Attentional control measures may therefore help track symptom-cognition dynamics during TMS in real-world clinical settings.