Abstract
When we look at a complex scene we are able to perceive a limited set of objects simultaneously. Such ability involves at least two stages of processing. Early individuation processes tag a limited number of elements by binding basic features to object locations and produce approximate representations of the indexed objects. Subsequent mechanisms, likely dependent on Visual Working Memory (VWM), encode in greater details the individuated objects, ultimately leading to their complete representation. Using a four-dot masking procedure the present electrophysiological study investigated whether the functioning of early and late stages of multiple object processing requires awareness. We measured N2pc, a neural marker of individuation, and CDA, a marker of VWM, while participants saw a variable number (from 0 to 3) of uniquely colored target-dots displayed among distractors. Participants’ task was to report target numerosity. On target-present trials, one target and one distractor were surrounded by four dots. The four-dot mask could offset together with the stimuli (common-offset) or not (delayed-offset). Results showed that participants were less accurate in delayed-offset trials than in common-offset trials, indicating that targets were successfully masked. ERP results showed that the amplitudes of both N2pc and CDA decreased in delayed offset trials. However, while the N2pc amplitude increased as function of numerosity in both delayed and common-offset trials, such modulation in the CDA was found only for the common-offset condition. Further analyses on the N2pc for the delayed-offset condition showed a numerosity-related modulation in both correct and incorrect trials. These results indicate that although awareness overall affects the functioning of multiple object analysis, multiple target individuation can operate in conditions of reduced awareness. In contrast, detailed encoding procedures on multiple targets are significantly less effective with reduced awareness. This in turn suggests that awareness is progressively required to build a full representation of multiple objects.
Meeting abstract presented at VSS 2012