Cross-Subject Assistance: Inter- and Intra-Subject Maximal Correlation for Enhancing the Performance of SSVEP-Based BCIs

Objective: The current state-of-the-art methods significantly improve the detection performance of steady-state visual evoked potentials (SSVEPs) by using individual calibration data. However, time-consuming sessions limit number training trials and may give rise to fatigue, which weakens effectiveness For addressing this issue, study proposes a novel inter- intra-subject maximal correlation (IISMC) method enhance robustness SSVEP recognition via employing similarity variability. Through efficient transfer learning, similar experience under same task is shared across subjects. Methods: IISMC extracts subject-specific information task-related from oneself other subjects performing maximizing correlation. Multiple weak classifiers are built several existing then integrated construct strong average weighting. Finally, powerful fusion predictor obtained for target recognition. Results: proposed framework validated on benchmark data set 35 subjects, experimental results demonstrate that obtains better than state art component analysis (TRCA). Significance: has great potential developing high-speed BCIs.