* Fitting the regression line for cerebellum group and SST cell type was using only 1 datapoint, and producing erroneous results. Now, the requirement for fitting the regression coefficient is have a minimum of 6 datapoints.
* Added some debug statement for logging, which helps in checking how well the regression ran.
* Changed how compute_region_volumes function in atlas_densities.densities.utils is calculating the "id_volume" column. Earlier it was generating nan values which leads to errors downstream in the pipeline.

…ties