Discovery And Research Of Arc Recognition Neurons In The Primary Visual Cortex Of Macaques

Vision is of great importance to primate.How different kinds of visual stimuli are encoded in different stages of visual pathway is always worth exploring.Visual shape recognition in primary depends on a multistage pathway from primary visual cortex to inferotemporal cortex.Visual stimuli are first perceived by retina,and then relayed by lateral geniculate nucleus(LGN)to primary visual cortex(V1 area),where most of neurons reactive to linear stimuli such as edge or bar.In V1,simple cells and complex cells encode linear stimuli in different orientations,and another kind of neuron called supercomplex cell can encode not only the orientation of linear stimuli but also the length.Contrast to linear stimuli,contours of higher complexity such as curvature,angle or cross,are generally thought to berepresented only in higher extrastriate cortex like V2 or V4 area.Among the complex visual stimuli,curvature is widely distributed in natural environment,and is important to creatures.Plenty of electrophysiological researches about neurons encoding curvature in V4 have been done so far,but this kind of neurons have not been found in V1.Though neurons reactive to curvature stimuli in V1 were reported in a few of previous researches,their specificity for curvature was not confirmed for shortage of neuron number and lack of systemic control experiments.Integrated study of their response features and mechanism has not been done as well.There were also researchers holding the view that the curvature-reactive neurons found in V1 were supercomplex cells reactive to short bar in essence.In our research,we adopted long-term two-photon calcium imaging technology in two awake monkeys:AAV mediated calcium indicator GCaMP5 was transfected into V1 neurons;after the calcium indicator fully expressed,responses to visual stimuli of V1 neurons in layer 2 was observed and recorded by two-photon calcium imaging.With the help of this advanced technology,a subset of V1 neurons specifically reactive to curvature stimuli were discovered for the first time.Then by performing a series of elaborate control experiments,we excluded the possibility that the responses to curvature resulted from the following cells in V1:1)supercomplex cells,2)cells with orthorhombic optimum orientations in central and peripheric receptive field,3)cells encoding spatial frequency;and confirmed the specificity of reactivity to curvature in these cells.Then we verified these curvature detecting cells as complex cells which had general response bandwidth for image position or size,and recognized curvatures by their contour features;rather than simple cells which were sensitive to image position or size,and encoded curvatures via their curve-shaped receptive fields.We also discovered that the curve direction preference of a certain curvature detecting cell coordinated with the line orientation preference of the functional column the cell located in,which revealed functional relationship between a curvature detecting cell and its surrounding V1 cells.In the end,two possible circuit models about how these curvature detecting cells receive and pr.ocess signal in visual perceptive pathway were proposed.One is simple-curve-cell model:several LGN cells whose receptive fields aligned in curve shape project into a V1 simple curve cell,and then several linear simple cells with same direction preference project to one complex curve cell.The other is line-to-curve model:several linear complex cells with adjacent direction preference project to one complex curve cell.We also proposed that the line-to-curve model is more possible and discussed the research direction in the future.In this research,the discovery of neurons encoding shape independently in V1 is an important breakthrough to traditional shape representation theory.It expanded our understanding of primary visual cortex' s complexity and function,and provided new perspective for future theoretical research in related fields.