Shantanu,
If you mean the slope of the fitted baseline for each run by 't^1 coefficient', then it is just a voxel-dependent number measuring the drift effect. In other words, the baseline is modeled in the following linear fashion:
c(0)*t^0 + c(1)*t^1
c(0) and c(1) are estimated based on the default TR implied. Therefore there is no 'time scaling' issue involved about c(1) unless you are talking about something else.
Gang