Abstract
Previous studies point to the existence of two qualitatively different enumeration processes in vision: a fast and efficient process specialized for small numbers of items (4 items), counting. Enumerating visual items requires the integration of the items into a single quantity. Previous findings have suggested temporal and qualitative differences between the integration of visual information within and across the hemifields, with within-hemifield integration preceding (Large & al., 2008) or at least being more efficient (Pillow & Rubin, 2002; Humphreys & al., 2000) than across-hemifield integration. Accordingly, one might expect enumeration to be more efficient when the items are presented in one hemifield only. Against this, the present study reveals that enumerating is more accurate when the visual items are split between the left and right visual fields than when they are all presented within the same hemifield. This finding is consistent with the notion of independent resources in the left and right hemispheres (Luck & al., 1989) and with recent data that have shown that parallel processing by the two hemispheres can expand the capacity of visual processing (Alvarez & Cavanagh, 2005; Delvenne, 2005; Scalf & al., 2007).