Abstract
The threshold for contrast detection improves when a target is presented between two collinear flankers (Polat & Sagi, Vision Res 1993). Facilitation is found when the target is presented simultaneously with the flankers or with a delay (Tanaka & Sagi, Vision Res 1998). Here we test the temporal properties of lateral facilitation by manipulating asynchrony of both onsets and offsets of the target and flankers. In the experiments, the contrast threshold for a Gabor target was measured in isolation or in the presence of two collinear Gabor masks using a 2AFC staircase method. Two sets of experiments were carried out: (1) the mask (60 ms) was presented before (Forward Masking: FM, isi=60 ms), simultaneously with (SM) or after (Backward Masking: BM, isi=60 ms) the target (60 ms); (2) the mask was presented for 500 ms while the target was presented for 60 ms at the onset of (SM+BM), during (FM+SM+BM) or at the end of (FM+SM) the mask stimulus. High contrast crosses marked the target's presentation time. The results showed threshold reduction, 0.15–0.35 log units, (1) for simultaneous target and mask presentations (SM) with all durations tested, 60–320 ms, (2) when the mask preceded the target (FM), but (3) not when the target offset preceded the masks offset (BM, SM+BM, FM+SM+BM), where some suppression at 3l was observed instead. This pattern of results implicates a long integration time for lateral excitatory effects. The masking of facilitation by sustained masks (SM+BM, FM+SM+BM), in the absence of changes in the stimulus uncertainty level, argues against an uncertainty reduction based explanation for the increased sensitivity. The results point to the existence of a neuronal network with slow excitatory processes that are activated by low contrast inputs (target) and fast inhibitory processes that are activated by high contrast inputs (flankers). The presence of the flankers after the target's offset reinforces the inhibitory effects.