Abstract
We obtained cued partial reports of arrays of 9 letters, 3 per row, to trace out iconic decay (Sperling, 1960). Letters were enlarged in successively lower rows below fixation to form a size gradient with more eccentric letters being larger. In ‘flat’, disparity was adjusted to cancel the depth impression caused by the size gradient. In ‘depth’, disparity reinforced the size gradient, so that the larger (bottom) letters appeared closer to the observer and the smaller (top) ones further away, as on an inclined plane. One row was cued at random in each display with an arrow. Letter and cue displays were 50 ms in duration. ISIs were -100 (cue before), 100, 300, or 700 ms (cue after). Results: reports were less accurate in ‘depth’ than in ‘flat’, at all ISI’s. We speculate that shifting attention across depth planes is slower than shifting within a depth plane, so more letters are lost in ‘depth’ than in ‘flat’ before transfer to working memory for report. In contrast, when the relevant depth plane is known in advance, so an attention shift is not required, Xu & Nakayama (JoV, 2009) found that depth improves visual memory by 5%.
Meeting abstract presented at VSS 2012