Abstract
Multifocal VEPs can be used to investigate retinotopically organized visual areas. Source localization works best using the smallest patch sizes that give reliable signals. We investigated how the scalp signal varies with stimulus patch size.
The stimulus was a dartboard pattern consisting of 12 spokes and 4 concentric cortically scaled rings for a total of 48 patches. The rings extended from 1 to 6.5 deg in eccentricity. Each patch contained 4 squares in the radial direction, and from 1 to 7 squares per spoke in the tangential direction. Patches were modulated using a binary M-sequence and the scalp potential topographies (64 to 96 recording electrodes) were cross correlated to get 2nd order kernel responses.
Amazingly, reliable responses were obtained for stimuli that activate approximately 2 by 8 mm of cortical surface in V1. Moreover, when averaged across all stimulus locations signal strength increased linearly with patch size. Some individual patches exhibited deviations from linearity that were consistent with known cortical topography. For example, if the activated cortical area were at the apex of a sulcus, as patch size grows the activated area would extend around the sulcus. As it grows around the sulcus equal and opposite currents cancel with little or no change in signal strength at the recording electrode.
In further experiments we tested spatial summation directly by independently stimulating subdivisions of patches and comparing the summed response to the undivided patch response. For patches in the inner and outer rings the linear summation of the sub-regions matched the response to the large region. However, in the two middle rings subdivisions exhibited faster response latencies.
Future VEP source localization studies using small areas of activation could potentially provide a fine 3D retinotopic map of the cortical surface.
Supported by EY015825, T32EY 07043