MVPA-based fMRI neurofeedback shows moderate neural effects but inconsistent behavioral outcomes
This systematic review and meta-analysis examined 29 studies evaluating MVPA-based fMRI neurofeedback across multiple domains including emotion regulation, fear conditioning, associative and perceptual learning, attention, craving, semantic neurofeedback, and motor rehabilitation. The global meta-analysis revealed a moderate, statistically significant effect of MVPA-based neurofeedback on neural outcomes. In the emotion regulation subgroup specifically, a moderate effect on neural outcomes was also observed. Behavioral outcomes were less consistent than neural outcomes across studies, with associative learning and craving reduction studies showing unclear behavioral outcomes, motor rehabilitation and semantic neurofeedback studies lacking behavioral assessments, and only a few emotion regulation studies reporting behavioral improvements. Safety and tolerability data were not reported. Key limitations include considerable variability in protocol designs and methodological aspects across studies, the need for standardized methodologies and clearer theoretical frameworks, the need to clarify terminology, and the need to address MVPA-specific methodological considerations including preprocessing, motion correction, and classifier selection. Funding and conflicts of interest were not reported. The evidence represents associations from a systematic review, and practice relevance was not specified. Given the methodological heterogeneity and inconsistent behavioral translation, this approach remains primarily a research tool requiring further standardization before clinical implementation.