Brain Imaging & Analysis Center

AMPLE generates normalized activation maps by converting statistical map values within each region of interest (ROI) to percentage of the peak values in that ROI. Any map can be normalized (e.g. t-map, correlation, etc); each user-defined ROI is normalized independently.

AMPLE normalization in fScan can be performed in two ways, with each using different commands. The -ample command converts an existing map by copying each ROI in the input data set to a new output data set. In the process, it identifies the peak values (positive or negative) in the ROI and then converts all voxels to percentage of that peak value.

The other way to use the AMPLE algorithm in fScan is to have it generate normalized AMPLE activation maps repeatedly during a fMRI scan and calculate how many voxels are at or above each percentage level (in 10% contour bands), The goal of this approach is to characterize the relative spatial distribution of the pattern of activated voxels, and determine whether that pattern stabilizes as a function of scan time. Stable AMPLE activation patterns (as seen by plateau counts of voxels at each AMPLE percentage level) have been shown to be highly reproducible and thus provide a way of evaluating confidence in fMRI results (Voyvodic, 2006).

See Also:
fScan Home, fScan Manual, Topics, Manual Help