Effects of EEG neurofeedback and training interventions on golf putting performance: a systematic review and meta-analysis

Background Electroencephalography neurofeedback (EEG-NFT) is used to modulate brain function to enhance golf putting performance. However, the effectiveness of EEG-NFT and its relationship with non-EEG training interventions in improving putting performance remain unclear. Methods Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and the International Prospective Register of Systematic Reviews (PROSPERO) registration (CRD420251068586), this study systematically searched multiple databases up to 8 June 2025. Randomized controlled trials (RCTs) investigating EEG-NFT and other training interventions for golf putting were included. The meta-analyses employed random-effects models, with subgroup analyses conducted to examine moderators. Results A total of 12 RCTs involving 424 participants were included. The overall analysis revealed a significant positive effect on putting performance (standardized mean difference, SMD = 1.33; 95% CI: 0.36–2.29; p = 0.007), though with substantial heterogeneity ( I 2 = 94%). Subgroup analyses demonstrated that skill level significantly moderated the intervention effects ( p = 0.0003), with novices showing the largest improvement ( SMD = 2.10). Crucially, intervention efficacy was strongly dependent on both the anatomical specificity and frequency band targeting: central region-sensorimotor rhythm (Cz-SMR)-NFT demonstrated robust positive effects ( SMD = 1.12, 95% CI: 0.59–1.66), while inhibitory protocols targeting Mu rhythms ( SMD = −1.84) and frontal midline theta ( SMD = −1.92) showed significant negative effects. Non-EEG training interventions also exhibited positive effects on putting performance, with skill practice-based interventions showing comparable efficacy to effective EEG-NFT protocols in novice golfers. Conclusion EEG-NFT can enhance putting performance, but success requires precise protocol selection aligned with skill level and neural targets. Cz-SMR enhancement appears to be reliable, whereas non-specific suppression may be detrimental. Future research should establish standardized protocols and conduct larger RCTs to validate these findings. Systematic review registration https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=1068586 , identifier (CRD420251068586).