Abstract
The uncinate fasciculus (UF) is a cortico-cortico pathway that links the temporal and frontal lobes. In non-human primates transection of the UF causes impairment in learning visuo-visuo and visuo-motor associations, while object discrimination remains intact. Here, we used diffusion tensor imaging (DTI) and behavioral experiments to examine the role of the UF in mediating visual associative learning in humans. Specifically, we sought to establish whether there was any correlation between learning performance and the structural properties of the UF in 18 participants who were trained to learn associations between specific sets of indoor and outdoor visual scenes and individuals (six female, caucasian faces, oriented in three possible viewpoints). In each trial, a scene was presented along with two faces (Target/Distractor) and participants were asked to match the scene to the target individual. Visual and auditory feedback was provided after each trial. In total, participants learned 36 associations over 1080 trials. To measure the structural and architectural properties of the UF, we collected diffusion-weighted images (120 directions). After correcting the DTI data for artifacts, the data were split into two halves of 60 directions each. For each half of the data a tensor volume was reconstructed and standard DTI measures were computed. From one tensor volume, streamlines depicting the bilateral uncinate fasciculi (UF), and inferior longitudinal fasciculi (ILF) were extracted using deterministic tractography. The streamlines from each pathway were used as a mask to compute the mean fractional anisotropy (FA) and mean diffusivity (MD) within the tract from the independent tensor volume. Correlation analyses between the learning rate in the behavioral task and tract measures revealed a strong relationship between learning and structural properties of the UF, but not other tracts. These findings highlight a role for the human UF in visual associative learning similar to that reported in monkey.
Meeting abstract presented at VSS 2012