Bidirectional EEG Neurofeedback of Sensorimotor Beta Oscillations Reveals Asymmetric Modulation and Strategy-Dependent Learning.

EEG-based neurofeedback (ENF) may be utilised as an intervention for improvement of motor performance in Parkinson's movement impairment. This exploratory study aimed to determine the ability of healthy participants to bidirectionally modulate sensorimotor beta power using ENF in real time and examine learning trends. The protocol trained participants to increase and decrease the amplitude of their individual beta power over C3 using a double-blind sham-controlled and within-subjects design. Fifteen healthy participants completed the study (mean age = 38.9, SD = 14.9, age range = 20-66; nine female and six male). Thirteen participants successfully decreased beta activity, three participants successfully increased it and two participants demonstrated bidirectional control. A Wilcoxon signed-rank test showed that successful EEG control during real ENF was significantly higher than sham ENF when participants received the sham condition first (Z = 3.80, p < 0.001). When participants received the real condition first, the difference in success was not statistically significant (Z = 1.19, p = 0.233), indicating a transfer learning effect from real to sham sessions. ENF performance varied across ENF training based on modulation direction and cognitive variables, including mental strategy, engagement and fatigue. This study showed that initial ENF performance depends critically on how performance is defined and measured. Learning trends between real and sham conditions further suggest that early ENF performance primarily reflects strategy-dependent, effort-based control rather than stable neurophysiological plasticity. These findings highlight the need for a more precise framework for ENF evaluation, one that separates transient modulation, strategic learning and durable physiological adaptation.