Efficacy and neural mechanisms of transcranial direct current stimulation for social function in children and adolescents with autism spectrum disorder: A systematic review and meta-analysis.

Objectives

To evaluate the intervention effects of transcranial direct current stimulation (tDCS) on social functioning and neurophysiological indicators in Autism Spectrum Disorder (ASD), and to explore neuro-behavioral consistency at the individual and cross-study levels.

Methods

Following the PRISMA guidelines, we conducted a systematic search of PubMed, Web of Science, Embase, Scopus, and the Cochrane Library from inception until January 7, 2026. We excluded studies involving exclusively adult cohorts and included randomized controlled trials (RCTs) reporting both social behavioral and neurophysiological outcomes. Risk of bias was assessed with Risk of Bias 2 (RoB 2). A random-effects model was employed to calculate the standardized mean difference (SMD), along with tests for heterogeneity and subgroup analyses. A multidimensional matrix was constructed to assess neuro-behavioral consistency.

Results

Thirteen RCTs were included (N = 357, ages 4-21 years). tDCS significantly improved social functioning (SMD = -0.51, I2 = 0%, p = 0.0002) and induced significant overall neurofunctional modulations (SMD = 1.03, I2 = 71.9%, p = 0.004). Subgroup analyses stratified by measurement modalities revealed distinct profiles across modalities: the functional near-infrared spectroscopy (fNIRS) subgroup showed the largest pooled effect size in neurophysiological responses (SMD = 1.32, 95% CI [0.31, 2.33], I2 = 63.5%); the electroencephalography (EEG) subgroup also exhibited a large pooled effect (SMD = 1.16, 95% CI [0.07, 2.25], I2 = 85.8%); whereas the event-related potential (ERP) subgroup demonstrated a modest and non-significant effect (SMD = 0.37, 95% CI [-0.23, 0.97]) with the lowest heterogeneity (I2 = 0%). Consistency analysis revealed a cross-level dissociation: cross-study regression showed no significant linear correlation between neurobiological and behavioral effect sizes (r = 0.19, p = 0.760), while the multidimensional consistency matrix at the individual-study level showed that, across the four consistency dimensions, two of five studies were fully consistent, two showed partial consistency, and one was inconsistent.

Conclusion

The left dorsolateral prefrontal tDCS protocol shows potential in improving social functioning in children and adolescents with ASD. Translational research on its neural mechanisms requires standardized measurement paradigms, with ERP showing high cross-study consistency in the limited available data (k = 2), suggesting it may warrant further exploration as a potential biomarker, though this finding remains preliminary. Future large-scale clinical trials should expand sample representativeness, with attention to the potential impact of gender differences and different neurodevelopmental windows on intervention efficacy.