Abstract
We have previously documented that reliable reductions in contrast detection thresholds (as a function of practice) are accompanied by reliable increases in false alarm rates. These increases in false alarm rates occur with and without the presence of trial-by-trial feedback, and also occur when observers practice the contrast detection task without the requirement of making explicit present/absent judgments. These patterns suggest a liberal shift in response criterion in perceptual learning. However, the lack of effects due to feedback call into question the extent to which these shifts reflect decisional (“top-down”) influences in a form of learning that has typically been interpreted as reflecting changes at low levels of a feed-forward system. The present study examines the relative responsiveness of cortical regions (including V1–V2) to varying levels of contrast, before and after practice with a contrast detection task. A twelve-day protocol was used. On day 1, a multifocal visual evoked potential (mfVEP, Slotnick et al. 1999) procedure was used to localize neural regions responsive to contrast onsets at known retinal locations. On days 2–11, observers practiced the contrast detection task, using a 2I-2AFC staircase procedure, with stimuli presented at a single, invariant location throughout practice. On day 12, the mfVEP procedure was used, with variations in contrast at both trained and untrained locations, and in the presence of valid and invalid cues for spatial attention. The design allows examination of the responsiveness of low levels of the visual system, at trained and untrained locations, and in the presence and absence of directed spatial attention.