I think the following explanation should help:
If flag is 0 (default), then zscore scales X using the sample standard deviation...
If flag is 1, then zscore scales X using the population standard deviation
zscore(X,flag,dim) standardizes X along dimension dim
Also, if you would like to use zscore functionality without the Statistics and Machine Learning Toolbox, the following answers discussion might help: https://www.mathworks.com/matlabcentral/newsreader/view_thread/275744
Also the "bsxfun" function can help you combine the two functionalities with bsxfun(@rdivide, bsxfun(@minus, Data, meanvalue), standard deviation value)
mean can be calculated using the "mean" function and standard deviation using "std"