Patients with V1 damage exhibit increased orientation decoding in hMT+, but only if pulvinar is intact

Abstract

Orientation selectivity is a core property of V1 in mammals. Patients with V1 damage can relearn orientation discrimination at trained, blind-field locations. Here, we investigated a potentially key role for pathways bypassing V1, which directly transmit information to downstream visual cortical areas, generating orientation selectivity in these areas. We studied 3 stroke patients (33-63 y/o, all females): one with a large right V1 lesion, a second with a right V2/V3 lesion that spared V1, and a third with lesions affecting both right V1 and pulvinar. Participants were scanned with BOLD fMRI. They viewed small (2.5 deg radius), oriented (45 or 135 deg) gratings in the periphery (7.1-11.2 deg eccentricity) while performing a demanding task at fixation. Stimuli were presented either deep within the blind field or in a mirror-symmetric location in the intact hemifield. We also performed retinotopic mapping, scanning with an MT localizer, and T1-weighted structural scans. In the patient with extensive V1 damage, the ipsilesional hMT+ was visually responsive and able to decode orientation. However, in the patient with a V2/V3 lesion and the patient with right V1 plus pulvinar damage, the ipsilesional hMT+ was visually responsive, but orientation decoding failed to reach significance. Healthy controls exhibited significant orientation decoding in V1 but not in hMT+. Our findings suggest that after V1 damage, strong orientation selectivity may emerge in hMT+ from circuits bypassing V1, including via the pulvinar. When V1 is spared, even when V2/V3 are damaged, these alternative circuits do not generate robust orientation-selective BOLD signals in hMT+. Ongoing work is assessing the functional implications of these findings for perception and rehabilitation potential.