<big sigh>. As a rule of thumb, if you gather more multi-modal data, you are going to need more samples (patients) rather than less to be confident of results. Importantly you need some sort of credible belief that for the measurements you propose to do you have some hope of the analysis working
