Natural image reconstruction on the basis of local field potential signals of pigeon optic tectum neurons.

Reconstruction of visual input through a neuron response helps to understand the information processing mechanism of the visual system. This paper uses the amplitude and phase characteristics of the local field potential signal in the pigeon optic tectum area to reconstruct the visual input from the neuron response data by means of local information accumulation using a linear inverse filter and a back propagation neural network algorithm. The reconstructed results show that the correlation between three reconstructed images and their corresponding stimulus images (tree branches, birds, and eyeglasses) was 0.8461±0.1135 for optimal values of number of channels, response duration, time from stimulus onset, and frequency band. This method of reconstructing the natural image from the pigeon optic tectum area neuron response signal can be applied to coding mechanism analysis of brightness and structural information in the visual system and to feedback from implantable visual prostheses.