Abstract
In order to find a target, one needs to know what the target is. The process of internalizing the target template is critical for visual perception. For example, by previewing a target object just once, this one-shot seeing has been shown to enable sustained improvement in perceiving subsequent objects, an effect known as insight (Rubin, Nakayama & Shapley, 1997) or eureka (Ahissar & Hochstein, 1997). Here, we report a new type of template enhancement effect that lasts longer than (transient) visual priming, but has a shorter lifetime than (sustained) insight/eureka. Subject performed a shape discrimination task (square vs. diamond) with two target durations: 200 ms (easy) and 16.7 ms (hard). We compared discrimination performance in mixed and single blocks. The mixed blocks consisted of 10 hard trials, 20 alternating easy and hard trials, and 10 hard trials. The single blocks consisted of 40 hard trials. The mixed and single blocks alternated with at least 15 s interval between blocks; a total of 28 blocks were run. We found that, performance was much better for the 2nd 10 hard trials than for the 1st 10 hard trials (83% vs. 69% correct), but only for the mixed blocks, demonstrating that mixing easy trials greatly enhances performance in the hard trials. In addition, this effect spilled over to enhance subsequent hard trials, with little drop in performance, resulting in comparable performance between the 2nd and 3rd 10 hard trials in the mixed blocks. However, this effect did not carry over to the single blocks, resulting in comparable performance between the 1st 10 hard trials in the mixed and single blocks. These results reveal a fast route of visual learning, whereby the visual system exploits target templates to enhance visual perception in a temporally restrictive manner. We call this template effect seeing to see.
Meeting abstract presented at VSS 2015