Abstract
It is known that humans can rapidly and accurately estimate small sets of items, a process referred to as subitizing. Similarly, shape recognition is rapid and accurate for simple geometric shapes. The precise but limited nature of both processes suggests that they might be mediated by common mechanisms. Here we investigate the role of spatial configuration and luminance polarity of elements on number estimation. Stimuli consisted of a small number of Gaussian blobs (3 to 6) positioned in different spatial configurations, either on the vertices of simple geometric shapes (triangle, square, pentagon, hexagon) or random. The elements were either the same or different in luminance polarity. We measured event-related potentials (ERPs), reaction times (RTs) and accuracy while participants performed a number estimation task. We found (a) faster RTs and higher accuracy when the blobs were located on simple shape configurations compared to random conditions and, when the number of elements and shape vertices were the same; (b) faster RTs with same than different luminance polarity but only for larger numbers of elements and no effect of polarity on accuracy; (c) an N2 component, around 250-400 ms over lateral occipital electrodes, that increased in amplitude for shape configurations compared to randomly placed elements; (d) an N2 mean amplitude that gradually increased with the number of elements; (e) an inhibitory effect over left mid-frontal electrode sites (termed LMFE) measured between 400-700 ms whose mean amplitude was inversely related with the number of elements; specifically, LMFE appeared to separate low (3, 4) and high (5, 6) number of elements irrespective of their shape configuration. Altogether, the performance and the graded ERP pattern of responses in the N2 range and LMFE suggest that rapid number estimation can benefit from simple shape configurations, involving a fast shape template matching process.
Meeting abstract presented at VSS 2017