Spatial integration and sensitivity changes in the human rod visual system.

1. The factor by which increment threshold rises with increasing background intensity is less if the target is small than if it is large. The difference is usually attributed to a reduction in the area over which visual signals are integrated as the visual system light adapts. Recently, however, it has been argued that the difference in slope may instead be caused by an increase in the gain of the local response function with light adaptation. 2. To test this hypothesis in the rod-driven visual system, we compared monoptic, small and large target increment thresholds, and dichoptic, large target brightness matches, measured as a function of background intensity in a typical, complete achromat, who has no cone vision. 3. The dichoptic brightness matches were made using a large target of a similar intensity to the threshold intensity of the small target. If local intensity is important, the large target brightness matching curve should be more similar to the shallow, small target threshold curve. But, if changes in spatial integration are important, the brightness matching curve should be similar to the steeper, large target threshold curve. 4. The slope of the large (1.85 deg) target increment threshold functions measured with either 200 or 50 ms test flashes were steeper than those of the small (10 min of arc) target functions by 0.10 (on logarithmic co-ordinates) or about 15%. 5. The logarithmic slopes of the dichoptic brightness curves were also slightly steeper than the small target increment functions. This is contrary to the local response (only) hypothesis, which predicts that the brightness curve should have the same slope as the small target function because the luminance of the targets in the two cases is the same. 6. We conclude that there must be a change in spatial integration in the rod visual system during light adaptation, over and above that due to local gain changes.