I'm confused with your description. Are you talking about voxel-wise time series for each subject? Is this a subject- or population-level analysis? With the former, how do gender and age come into play? With the latter, you would have to elaborate it more. What is exactly your model?