Roe Anna W, Ts'o Daniel Y
Department of Psychology, Vanderbilt University, Nashville, USA; Zhejiang University Interdisciplinary Institute of Neuroscience and Technology, Zhejiang University, Hangzhou 310027, China.
Department of Neurosurgery, SUNY-Upstate Medical University, Syracuse, NY, USA.
Cortex. 2015 Nov;72:168-178. doi: 10.1016/j.cortex.2015.07.007. Epub 2015 Jul 22.
The computation of texture and shape involves integration of features of various orientations. Orientation networks within V1 tend to involve cells which share similar orientation selectivity. However, emergent properties in V2 require the integration of multiple orientations. We now show that, unlike interactions within V1, V1-V2 orientation interactions are much less synchronized and are not necessarily orientation dependent. We find V1-V2 orientation networks are of two types: a more tightly synchronized, orientation-preserving network and a less synchronized orientation-diverse network. We suggest that such diversity of V1-V2 interactions underlies the spatial and functional integration required for computation of higher order contour and shape in V2.
纹理和形状的计算涉及各种方向特征的整合。初级视觉皮层(V1)内的方向网络往往包含具有相似方向选择性的细胞。然而,V2中的涌现特性需要多个方向的整合。我们现在表明,与V1内的相互作用不同,V1-V2方向相互作用的同步性要低得多,而且不一定依赖于方向。我们发现V1-V2方向网络有两种类型:一种是同步性更高、保持方向的网络,另一种是同步性较低、方向多样的网络。我们认为,V1-V2相互作用的这种多样性是V2中高阶轮廓和形状计算所需的空间和功能整合的基础。