I'm having a dimensionality mismatch problem. The preprocessed resting-state NYU scans correspond to a 49 x 58 x 47 x 172 matrix where 172 is the number of time steps. However, when I look at the extracted time series from "NYU 1000 ROI extracted time courses" directory, the brain atlas is a  53 x 64 x 46 matrix and the time series have 176 time steps. What accounts for this difference? I'm doing some analysis on the preprocessed resting state data and would like to extract time series myself using the functional parcellation atlas but I can't because the atlas is the wrong dimension. 

Thanks for you help,

Erica