Abstract
Purpose: The population receptive field approach (pRF) 1 provides an estimate of the region of visual space that best excites each fMRI voxel. Recently there has been increasing interest in using the pRF method to examine cortical reorganization resulting from vision loss2. Here, we tested the validity of the pRF method in four normally sighted subjects under conditions that simulated a retinal lesion. Methods: We measured BOLD responses to bars drifting in random directions stimulating the central 15 deg of the visual field1 or with the central 3 deg masked, simulating a foveal scotoma. We analyzed the data in two ways. In the 'full visual stimulus' condition we used the full unmasked stimulus as the input into the pRF model, as is currently the standard procedure. In the 'retinal lesion stimulus' condition we used the masked stimulus as input into the pRF model. Results: Using the 'full visual stimulus' analysis method, voxels with pRFs that fell within or near the edge of the scotoma were shifted peripherally, even in voxels with relatively low fit errors (there was no systematic bias in receptive field size). However, this 'apparent rapid re-organization' completely disappeared when analyzing the same data using the ‘retinal lesion stimulus’ method. A model simulation similarly found that the peripheral shifts observed with the 'full visual stimulus' analysis can be explained purely in terms of biases in the pRF fits, without any underlying cortical reorganization. Conclusion: It is possible to accurately measure pRFs under condition of vision loss, if and only if pre-cortical visual losses are taken into account. 1Dumoulin SO, Wandell BA. NeuroImage. 39(2): 647-660 (2008). 2Baseler HA et al. Nature Neuroscience. 14: 649–655 (2011).
Meeting abstract presented at VSS 2012