The coding of perceived eye position.

In order to reveal the mechanism of eye position coding, we measured the effect of eye position on assessing the position of a light viewed in a dark environment using a sound as a reference point, before and after adapting to prolonged periods of eccentric viewing (11 degrees right or left). During testing, eye position was varied over +/-22 degrees . For each test position, a PEST procedure was used to align the perceived position of a light (with no visual reference points) and a fixed sound source. The perceived position of the light was veridical when looking straight ahead but large and consistent errors were found with eccentric viewing indicating that only about 77% of the eye eccentricity was taken into account, an eye position gain of 0.77, consistent with previous reports. The error was altered by prolonged eccentric viewing. There were two components to this effect: a direction bias and a gain change. The direction bias was symmetric: in a change rather like that induced by prism adaptation, the perceived straight ahead moved in the direction of the previous eccentric viewing. The gain change was asymmetric: after looking to the left, the gain of eye position increased to closer to one (0.77-0.81), in other words the subjects became more accurate. After looking right, the gain of eye position decreased (0.77-0.73): subjects consequently became less accurate. We model these systematic changes in terms of an asymmetric coding system.