Abstract
During task-irrelevant perceptual learning (TIPL), sensitivity to stimulus features improves through exposure despite a lack of directed attention to them (Seitz and Watanabe, 2005). The properties of TIPL have been actively studied. For motion direction stimuli, Watanabe et al. (2002) suggested that TIPL occurs at a lower level in motion processing because learning was found for the local but not global motion direction of a dynamic dot display. However, the type of motion processing cells that provide a substrate to this kind of directional learning is still unknown. Here, we investigated if this learning is specific to the contrast polarity of the motion signals of the task-irrelevant stimulus. The procedure was similar to that employed in Seitz and Watanabe (2003). We developed a random dot motion stimulus, based on Wehrhahn and Rapf (1992), to target motion cells selective to contrast polarity by ensuring motion information arises only from signal dot onsets, and not their offsets. We measured the change in discrimination performance of human subjects in 4 non-cardinal directions for both positive and negative contrast polarities, after being exposed repetitively to a motion stimulus of a single contrast polarity while performing an attentionally-demanding RSVP task. Results show that learning does not transfer to the unexposed contrast polarity, suggesting that TIPL for motion stimuli can occur at or before the stage of directional V1 simple cells. These data support the prediction of the 3D FORMOTION model that a short-range directional filter generates directional V1 cells sensitive to contrast polarity, before projecting to MT via a long-range directional filter that pools across opposite contrast polarities (e.g., Chey, Grossberg, and Mingolla, 1997; Grossberg, Mingolla, and Viswanathan, 2001).
Supported in part by the National Science Foundation (SBE-0354378 and BCS-0549036) and the National Institutes of Health (R21 EY017737).