چكيده به لاتين
How does the brain integrate various features together into a unified representation? Two competing theories dominate the literature on feature binding: feature integration (FIT) and binding-by-synchrony (BBS). FIT emphasizes spatial attention in associative regions including the prefrontal cortex (PFC), whereas BBS emphasizes short-range synchrony in the visual cortex. However, neither theory fully explains how the brain binds features from different dimensions together, and the ‘feature binding problem’ remains unresolved. To tackle this problem, we analyzed electroencephalography (EEG) data obtained from 34 adults during performaing of a task which requires the rapid binding of spatial and temporal features in working memory. Critically, 14 subjects had discrete PFC lesions, permitting interpretation of causality. Using cutting-edge methods, including data-driven analyses of cross-frequency coupling and frequency band-limited functional connectivity, we show that: • Feature binding is linked to anterior delta-theta (2-7 Hz) activity, anterior-posterior delta-theta synchrony, and posterior cross-frequency coupling to gamma fluctuations. • Delta-theta synchrony supports interactions between PFC, temporal, and parieto-occipital regions, thus revealing an oscillatory mechanism sub-serving FIT. • Parieto-occipital cross-frequency coupling to gamma fluctuations, a mechanism in line with BBS, depends on PFC. • Behavior and both anterior and posterior mechanisms also depend on PFC independent of feature binding, corroborating a domain-general control role for PFC. We propose a novel mechanistic explanation of feature binding based on neural oscillations and top-down PFC control. By defining how interactions occur within and between regions of the brain involved in both attention and vision, we show how oscillatory synchrony underlies feature binding across the whole brain. We further provide the first neurological demonstration that neural signatures of feature binding depend on PFC. PFC directs posterior signatures of feature binding via low-frequency oscillations, supporting optimal behavior. Our study provides a comprehensive but succinct report of the neural basis of feature binding.
