How to solve this GLM problem

I have a dataset with 58 EEG channels per patient. I would like to solve this equation:

X1*b1 + X2*b2 + c = Y

X1 is a 57x1 vector X2 is a 57x1 vector Y is a 57x1 vector

I can easily solve this for each patient individually, which will give me the b-values and T- en p-statistics.

However, I would like to obtain statistically useful information on the entire population of 42 patients. I don't think a multivariate GLM will work as both the patients as the channels are observations, not variables.

Looking forward to hearing from you!