Kwon Soyoung, Watanabe Masataka, Fischer Elvira, Bartels Andreas
Vision and Cognition Lab, Centre for Integrative Neuroscience, University of Tübingen, 72076 Tübingen, Germany; Department of Psychology, University of Tübingen, 72076 Tübingen, Germany; Max Planck Institute for Biological Cybernetics, 72076 Tübingen, Germany; International Max Planck Research School for Cognitive and Systems Neuroscience, 72076 Tübingen, Germany.
Max Planck Institute for Biological Cybernetics, 72076 Tübingen, Germany; Department of Systems Innovation, University of Tokyo, 113-0033 Tokyo, Japan.
Neuroimage. 2017 Jan 1;144(Pt A):217-226. doi: 10.1016/j.neuroimage.2016.10.014. Epub 2016 Oct 11.
Attention allows our brain to focus its limited resources on a given task. It does so by selective modulation of neural activity and of functional connectivity (FC) across brain-wide networks. While there is extensive literature on activity changes, surprisingly few studies examined brain-wide FC modulations that can be cleanly attributed to attention compared to matched visual processing. In contrast to prior approaches, we used an ultra-long trial design that avoided transients from trial onsets, included slow fluctuations (<0.1Hz) that carry important information on FC, and allowed for frequency-segregated analyses. We found that FC derived from long blocks had a nearly two-fold higher gain compared to FC derived from traditional (short) block designs. Second, attention enhanced intrinsic (negative or positive) correlations across networks, such as between the default-mode network (DMN), the dorsal attention network (DAN), and the visual system (VIS). In contrast attention de-correlated the intrinsically correlated visual regions. Third, the de-correlation within VIS was driven primarily by high frequencies, whereas the increase in DAN-VIS predominantly by low frequencies. These results pinpoint two fundamentally distinct effects of attention on connectivity. Information flow increases between distinct large-scale networks, and de-correlation within sensory cortex indicates decreased redundancy.
注意力使我们的大脑能够将其有限的资源集中于特定任务。它通过选择性地调节神经活动以及全脑网络中的功能连接（FC）来实现这一点。虽然有大量关于活动变化的文献，但令人惊讶的是，与匹配的视觉处理相比，很少有研究考察可明确归因于注意力的全脑FC调制。与先前的方法不同，我们采用了超长试验设计，避免了试验开始时的瞬态，纳入了携带FC重要信息的缓慢波动（<0.1Hz），并允许进行频率分离分析。我们发现，与传统（短）块设计得出的FC相比，长块得出的FC增益几乎高出两倍。其次，注意力增强了网络之间的内在（负或正）相关性，例如默认模式网络（DMN）、背侧注意力网络（DAN）和视觉系统（VIS）之间的相关性。相比之下，注意力使内在相关的视觉区域去相关。第三，VIS内的去相关主要由高频驱动，而DAN-VIS之间的增加主要由低频驱动。这些结果指出了注意力对连接性的两种根本不同的影响。不同大规模网络之间的信息流增加，而感觉皮层内的去相关表明冗余减少。
