A Multi-level Account of Selective Attention

Abstract

To achieve behavioural goals, relevant sensory stimuli must be processed more quickly and reliably than irrelevant distracters. The ability to prioritize relevant over irrelevant stimuli is usually referred to as selective information processing, or selective attention. Over the last 50–60 years, there has been an ongoing debate about the point along the sensory–response processing stream at which selective attention operates: are relevant and irrelevant inputs segregated early in processing based on low-level featural differences, or does this segregation occur late in processing after the meaning of each stimulus has been computed? As with nearly all dichotomies in psychology, the emerging consensus is that neither extreme is correct. Instead, depending on task demands, the mechanisms of selective attention can flexibly operate on the quality of low-level sensory representations as well as on later stages of semantic analysis and decision-making.