Algorithm 1 Individual analysis

input: individual weight vector of participants output: the ranked channels of participants separately 1.The initial weight vectors are obtained as weight vector ‘W’; 2.The nonzero absolute value elements in the weight vector ‘W’ are arranged in descending order, and the interval of top 50%/top 25% is defined; 3.In the weight vector ‘W’, the number of absolute elements that fall within the 3 intervals (full/top 50%/top 25% intervals) are counted respectively along the channel dimension, and the weight count vector can be recognized as [counting values of 64 channels × 3 kinds of intervals]; 4.The weight count vector in step (3) is ranked in descending order, and the ranked channels of three counting intervals can be acquired, structured as [64 descending ordered channels × 3 kinds of intervals]; 5.The descending ordered channels of three different counting intervals are evaluated.