Hi AFNI experts,

I have two related questions:

(1) I try to understand how the voxel values in BRIK files are internally scaled. To be more specific, after I generate the impulse response function output file, called “IRF_Cong+orig” for now, with the option –iresp in 3dDeconvolve, I understand that each voxel value at each time point is float but stored as short integer in the “IRF_Cong+orig.BRIK” file. I would like to use Matlab to do some manipulation. To do this, I need to convert the short integer back to float. What is the scaling factor? Does AFNI use the same scaling factor for all sub-bricks with float values in 3dDeconvolve, such as the coefficients and t statistics?

(2) After the Matlab manipulation, what will be the easiest way to generate the HEAD file? For example, based on the “IRF_Cong+orig.BRIK”, I generate a file called “time_dealy+orig.BRIK”. Since the “IRF_Cong+orig” and the newly created “time_delay+orig” would share the same voxel size, image location and data axes orientations, what will be the way to extract these information (without manually entering them in “to3d” again) from “IRF_Cong+orig.HEAD” and apply them to “time_delay+orig”.

Thank you.

-- David