Dear John and SPM designers/users,
I have 2 questions regarding how the batch for 2nd level analysis actually works:
1. I've run a VBM analysis using factorial design. I selected at the end of batch "Global normalisation - Normalisation - proportional" since I would like to model GLM on GM regional volume relative to total GM volume (my inputs are modulated GM images). I learned that this setting will adjust individual map voxel values by dividing them by the sum (or mean?) of the individual image, is this the case? Could you please clarify how the individual-level re-scaling is actually done, in terms of what are the scaling factors (sum or mean or others) etc?
2. In the batch I also select absolute threshold masking. Could you please clarify if the individual scaling factor (mean or sum or others of the individual map) is from the original input image, or the image thresholded by this absolute threshold masking?
I asked these because I would need to get the rescaled voxel values and do some post-hoc analysis, but this cannot be done directly since the rescaling is run inside the batch and no extra rescaled maps are generated. I would need to calculate it outside SPM so the clarifications will be very helpful. Thank you very much!!