Comparative efficacy of interventions for transdiagnostic anhedonia: a network meta-analysis

Anhedonia is a transdiagnostic disturbance of reward processing, yet the comparative efficacy and certainty of available interventions remain unclear. We searched PubMed, Embase, Web of Science, and MEDLINE for controlled trials reporting quantitative anhedonia outcomes. The protocol was registered in PROSPERO under CRD42025630784. Random-effects frequentist network meta-analyses were conducted overall and by population, including major depressive disorder, schizophrenia, and healthy individuals with trait anhedonia. Randomized trials were assessed with the Cochrane Risk of Bias 2 (RoB 2) tool, and the single non-randomized trial was assessed with the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool. Certainty of evidence was assessed using the Confidence in Network Meta-Analysis framework. Sixty reports representing 61 independent trials, 5794 participants, and 80 contrasts were included. The primary network comprised 35 trials, 39 contrasts, and 17 treatment nodes. In the overall transdiagnostic network, estimates versus the pooled control node favoured vortioxetine 10 mg/day, left dorsolateral prefrontal cortex-targeted transcranial magnetic stimulation (left DLPFC-targeted TMS), behavioural activation, and meditation. The corresponding standardized mean differences were -0.90, -0.55, -0.52, and -0.49. Heterogeneity was substantial at I² = 71.90%, but the design-by-treatment test did not indicate global inconsistency. Left DLPFC-targeted TMS showed the most consistent evidence pattern across the overall and major depressive disorder networks, supported by five direct trials in the overall network and stable leave-one-trial-out estimates. These findings clarify the comparative evidence across pharmacological, neuromodulatory, and psychological approaches and provide a structured basis for future diagnosis-specific trials and treatment development.