Inconsistency and uncertainty of the human visual area loci following surface-based registration: Probability and Entropy Maps.

Abstract

Here we created two different multisubject maps (16 subjects) to characterize interindividual variability in the positions of human visual areas (V1, dorsal and ventral parts of V2/3, V3A, V3B, V7, LOc, MT+, and hV4 [or V4v and V8]), which were localized using fMRI and coregistered using a surface-based method. The first is a probability map representing the degree of alignment inconsistency for each area, in which each point in space is associated with the probability affiliated with a given area. The second, a novel map termed an entropy map in which each point is associated with Shannon entropy computed from the probabilities, represents the degree of uncertainty regarding the area that resides there, and is maximal when all areas are equally probable. The overall average probability and entropy values were about 0.27 and 1.15 bits, respectively, with dependencies on the visual areas. The probability and entropy maps generated here will benefit any application which requires predictions of areas that are most likely present at an anatomical point and know the uncertainty associated with such predictions.