The Hyper-Column Structure Research Of Primary Visual Cortex Of Bionic Brain

Primary visual cortex (Ⅵ) is the first cortical regions, when the visual information project to the brain cortex. Information processing on the primary visual cortex has very important significance for the perception and understanding of the outside world of follow-up cortex. Generally agreed that hyper-column Structure is the basic unit of the primary visual cortex. So it is a key process to study the hyper-column for understanding and simulating the primary visual cortex.According to the research on the physiology, this paper summarized and analyzed the neural connections between Lateral geniculate (LGN) and V1, the local neural interconnection of V1, and the neural connections between V1 and V2. This paper also Summarized and analyzed the structure relationship of function columns. The information input from LGN is the fundamental source of the V1 response. The input from other cortical regions only can maintain the response of V1. Layer 6 of V1 can project back to LGN, and this project can modify the response of LGN, and thus make this neurons have some orientation characteristics. According to the forward projection, the local interconnection of V1 can be divided into three layer, they were the layer of most simple cells (layer 4C), the layer of simple cells (layer 3B/4B), and the layer of complex cells (layer 2/3A). V1 also have two layers, layer 5 and layer 6, which mainly modify the response of V1. These two layers played an important role in the formation of V1 local loop, and the feedback interconnection can maintain the processing in V1. Different information from V1 project to different regions of V2, and this is good for extracting and processing visual information respectively. The function columns of V1 including the ocular dominance columns, orientation columns, the color column and so on, these columns connect to each other both on the position and functional characteristics. Some color-sensitive cells also sensitive to orientation, and some binocular cells also sensitive to orientation and direction.In this paper, we designed the neural network structure, and simplified several secondary neurotransmitters in the connection. We divided the neural Transmission network of V1 into two loops, the bottom loop (layer 4C, layer 3B/4B and layer 6) and the surface loop (layer 3B/4B, layer 2/3A and layer 5). As the most important part of the bottom loop, layer 6 receive not only the project from layer3B, but also 4C and LGN. The projection from 4C is the main input source of layer 6, but the small 3B projection determine whether the layer 6 neurons feedback to layer 4C and LGN, in other words, it plays the role of a control signal. The role of layer 5 in the surface loop is similar to the layer 6 in the bottom loop, and layer 2/3 plays the role of a control signal. In the design of the V1 columnar structure, we mainly analyzed the response characteristics of neurons at different layers, and the response effect between them, and designed response models of simple cell and complex cell. We established the response model of double-opponent neurons in the color channel, calculated the response of neurons, and analyzed the effect of the local layer connection and the value of feedback in the model.Through the establishment of the position category of neurons and the loop category of the neural loop about projection, we found the importance of the correspondence of neurons location in the visual system for the steady and correct information projection, and found the isomorphic relationship between different Neurotransmitter pathways. This will help us studying the functions of the V1 through the creation of a single super-column structure. These basic rules can be used in non-biological visual system, so can be used in the bionic brain.By experimental research, we found the imprecise detection of simple cells to external things, that is the function is not perfect, this is the result of single neurons' simple logic. After the interconnection between neurons and project to complex cells, the response of neurons becomes more accurate. This indicates that the interaction of the large number of neurons in the nervous system can increase the response of the whole system. These results note that the study of biological visual system is able to be used in the field of biomimetic, and it is ability to simulate the performance of biological systems.