Abstract
Exposure to a new environment elicits changes that keep vision functioning well. Adaptation occurs at various time scales, with some effects arising after only a few milliseconds and others growing slowly over hours or days. Long-term adaptation has been measured in prior behavioral work. However, its neural bases are as of yet unknown. We filled this gap, by measuring long-term adaptation with event-related potentials (ERP). We induced a form of contrast adaptation by placing subjects in a vertically-deprived environment. Subjects wore virtual reality goggles that showed video from a head-mounted camera. The images were filtered to reduce vertical energy by 85%. Subjects performed everyday activities while viewing the filtered world for 4 hours. Before and after adaptation, we measured ERP responses to vertical and horizontal circular 10 deg grating patches at 15% contrast. Stimuli randomly appeared in upper or lower visual fields and were interleaved with filtered "top-up" movie clips to maintain adaptation. Peaks in the ERP were observed at around 100 ms following stimulus presentation at posterior electrodes. To measure response amplitude, we averaged ERP voltages across 16 ms surrounding the peak. Prior to adaptation, response amplitudes were equal to both orientations. Immediately following adaptation, the responses were significantly different (p < .05). This effect of adaptation decayed during the 1-hour test period. It was also larger in the lower visual field than in the upper visual field. The simplest account of these results is that adaptation caused neurons tuned to vertical and horizontal to change their gain. Our data suggest that deprivation of vertical produced larger reductions in horizontally-tuned neurons' gain and smaller increases in vertically-tuned neurons', likely as a result of contrast normalization across orientations. Our results represent some of the first electrophysiological measurements of long-term adaptation in cortex.
Meeting abstract presented at VSS 2016