獼猴初級視覺皮質區明暗偏好神經之共振活動

Abstract

過去研究指出在多物種的初級視覺皮質區中，暗偏好神經的數量較亮偏好的神經多，這可能是視覺系統處理負向對比刺激優於正向對比刺激的表現之一。在表層皮質區的遞迴神經連結，可能是造成暗偏好的其中一種機制——表層皮質中的暗主導多神經活動（multiunit activity），相較於輸入層的局部場電位（local field potential）訊號要來的持久。基於上述發現，我們假設暗偏好神經之間的共振活動強度會比亮偏好神經來得強。我們使用64通道電極陣列同時紀錄獼猴初級視覺皮質中多個不同位置的神經細胞。神經視覺受域（receptive field）透過發放－觸發平均法（spike-triggered average）與白噪音刺激（white noise）計算而成，進一步透過受域特性可以計算細胞對明暗反應的偏向性（暗偏好、亮偏好）。我們透過交叉相關圖（cross-correlogram）量化一對神經細胞共振的活動強度，發現神經共振強度與細胞間距呈負相關，也如預期發現暗偏好神經間的共振活動強度較強，尤其是輸入層與表層間的亮、暗偏好細胞的共振活動強度差異最明顯。這樣的結果提供了另一種可能的機制，從輸入層到表層，暗偏好及明偏好的投射強度不一致，導致在表層發現更強的暗偏好現象。另外在使用多光柵刺激（Hartley gratings）的情況下，我們也發現交叉相關圖的強度較使用白噪音刺激來得高，且交叉相關圖的寬度也比較寬，這些發現可能是表層神經遞迴連結的證據。這些結果支持輸入層到表層的投射不一致與表層神經的遞迴連結的機制，對於暗優勢在視覺皮質的表層較強的機制都非常關鍵。

Previous studies have shown that black-dominant neurons outnumber white-dominant neurons in the primary visual cortex in many species. The black-over-white bias may serve as the neural substrate for better and faster processing of black than white objects. One possible neural mechanism for the ample black preference is through recurrent connections in the superficial layer of V1 — the black-dominant signal is more sustained for the multiunit activity of superficial layer than for the local field potential of input layer. Based on the finding, we hypothesized that the strength of synchronous firing among black-dominant neurons should be stronger than that among white-dominant neurons. Here we used a multi-electrode matrix to simultaneously record from multiple neurons in different cortical layers of macaque V1. The receptive field measured with white noise (a binary stimuli generated by m-sequence) by reverse correlation was used to determine a cell’s response bias (dominated by positive or negative contrast stimuli). We quantified the strength of synchronous firing of a cell pair as the peak amplitude minus the baseline amplitude of the cross-correlogram. The correlation strength was negatively correlated with the distance between two cells. The correlation strength between black-dominant cells was stronger than that of white-dominant cells, especially the cell pairs from the input to the superficial layer. These results might provide evidence for another mechanism, the unequal feedforward projection from input layer, might also be critical for generating the stronger black bias in V1 superficial layer. Moreover, the strength of the synchronous firing was also modulated by the spatial property of stimulus. In comparison with white noise, correlation strength was higher and the peak width of the cross-correlogram was broader when cells were stimulated by Hartley gratings (stimuli with high spatial correlation). These results are evidence for the stronger recurrent activity evoked by Hartley gratings. Overall, our results supported both feed forward and recurrent hypotheses that largely amplify the black signal in the superficial layer of V1.