> When I bring the output of these GLMs to higher level modeling like 3dMVM or 3dLME, will the variance or
> uncertainty of these parameter estimates be taken into account or are all parameter estimates at that point
> weighted equally?
No, 3dMVM or 3dLME only takes the effect estimates (or the beta values) to the group level, so the uncertainty of those estimates is not considered in the analysis. If you can formulate the group analysis as one or more contrasts, then you may use 3dMEMA, which takes into account the reliability information at the individual subject level.
Gang