基于多元线性逆滤波器的视顶盖神经元集群亮度信息解码研究

Decoding Visual Luminance Information from a Population of Optic Tectum Based on Multivariate Linear Inverse Filters

通过解码鸽子视顶盖(the optic tectum,OT)神经元集群亮度信息的方法,研究了OT区神经元集群编码亮度信息的神经机制。首先设计了具有瞬态闪变特性的亮度视觉刺激模式,采用微电极阵列记录了多通道的锋电位(spike)发放序列;然后提取了神经元集群的spike发放率特征,构造多元线性逆滤波器解码视觉刺激亮度;最后采用互相关和信息论的方法对解码结果进行了分析。结果表明:采用的多元线性逆滤波器有效解码了OT区神经元集群编码的亮度信息。通过对解码参数变化与解码精度的分析可知神经元集群数目13个、bin的宽度5 ms、刺激后15 ms、刺激后持续时间35 ms具有最高的解码精度。通过神经元集群中逆滤波器形态的分析发现,单神经元在亮度信息的编码过程中受周围神经元的动态调制。通过对解码精度与刺激闪变频率关系的分析,发现在刺激闪变频率19 Hz到53 Hz的范围内,重建质量较好,在刺激闪变频率为33.5 Hz时,重建质量达到最优。

The neural nance information from a mechanism that how visual luminance information is encoded was studied by decoding lumi- population of the tectal tectum（OT） neurons in pigeons. First, the visual luminance stimu- lus consisted of transient characteristics of flicker was designed and muhi-channel spike trains were recorded simulta- neously using a muhielectrode array. Then, spike firing rate characteristics were extracted from a neural population so that multivariate linear inverse filters could be constructed to decode visual stimulus intensity. Finally, the decoding results were analyzed by combining cross-correlation computation with information theory. The results showed that the multivariate linear inverse filters effectively decoded the luminance information which was encoded by neural popula- tion in OT. From the analysis of decoding accuracy under the decoding parameters were changed, a result was found that decoding accuracy was the best when the number of neural population reached 13, the time bin was 5 ms, the time from stimulus onset was 15 ms, and the duration of response from stimulus onset was 35 ms. Through the analy- sis of the inverse filter form in neural populations, single neuron was modulated dynamically by its surrounding neu- rons in the process of encoding the luminance information. Through the analysis of the relationship between decoding accuracy and stimulus flicker frequency, the quality of the stimulus reconstruction was better when flicker frequency between 19 Hz and 53 Hz and the performance was optimal when flicker frequency near 33.5 Hz.