Abstract
Goal. Perceptual learning is a long lasting improvement in performance following training. Attention has been shown to enhance performance and facilitate transfer across locations. Here we examine whether attention also: (1) enables learning when it would otherwise be absent; (2) facilitates transfer across tasks; (3) facilitates transfer across features. Methods. We used 2AFC orientation and spatial frequency comparison tasks, for which the standard stimulus was the same. On each trial, two Gabor patches appeared–a standard and a comparison. Observers compared either their orientation (and reported which one was more clockwise) or their spatial-frequency (and reported which one was of higher frequency). During Pre- and Post-tests (Days 1 and 5, respectively), observers were tested on both tasks with a standard Gabor (4-cpd oriented 30° or 120°) using a neutral cue. During Orientation Training (Days 2-4, 800 trials per day), observers compared the orientation of two 4-cpd stimuli, one which was always oriented 30° (standard stimulus). Half the participants trained with a central-neutral cue (Neutral-group) and the other half trained with peripheral cues adjacent to the two upcoming stimuli (Attention-group). Results. For the Neutral group there was no learning effect in either task. In contrast, the Attention group showed learning overall: For the trained orientation task, there was learning for the trained stimulus (30°), but not for the orthogonal untrained stimulus (120°). For the untrained spatial-frequency task, learning emerged for both the trained and untrained stimulus orientations. Conclusions. This study shows that training with a neutral cue does not yield learning, whereas training with exogenous attention: (1) enables perceptual learning when otherwise absent; (2) facilitates transfer across tasks; and (3) facilitates transfer across features for the untrained, but not the trained, task. These findings further our understanding of the benefits of attention–it enables perceptual learning when otherwise absent.
Meeting abstract presented at VSS 2014