Matlab code: "Wong, C. M., et al. (2020). Inter-and intra-subject transfer reduces calibration effort for high-speed SSVEP-based BCIs. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 28(10), 2123-2135.."
Dear author, in Figure 5 of your paper, you presented the number of source subjects has little effect on the performance when it >5.
But when I reproduced your code and set the number of source subjects to 9, I got a worse result below. . In my experiments, I just randomly selected 10 subjects and repeated the experiment in different subjects. I wonder if there is any mistake in my experiments.
Dear author, I found that the effect of intra-transfer seems minimal or even negative. In line 235 in fun_stcca.m, we can set to use only intra-transfer, inter-transfer, or both, and I got the accuracy results below. Is it right?
Dr. Wong, why do you use sort in line 8 of fun_calculate_ssvep_template.m and line 12 of fun_stcca.m? When I change this setting to 1-40 to keep the original order of stimulus frequency. The result is much lower (0.7385 VS. 0.4156 for 0.5s SSVEP).