在初級視覺皮質(V1)中,空間受域(spatial receptive field)是研究色彩以及明暗反應細胞的一種重要方法。在獼猴主要視覺區V1,許多色彩反應細胞在方向(orientation)及空間頻率(spatial frequency)上有良好的專一性。根據這樣的反應特性,我們可以預期這類高專一性細胞會有多重且細長的視覺受域形態(就如同簡單細胞(simple cell)一般)。然而,過去研究卻獲致不一致結果──許多這類型細胞有較寬短的受域型態。這個研究主要想要探討視覺受域及其反應型態不一致的神經基礎為何。實驗中會使用兩種不同刺激來量測V1神經的受域型態:Hartley光柵(Hartley Gratings)和二元雜訊(binary sparse noise)。受域形態會透過反向相關(reverse correlation)計算,並套用二元Gabor函數獲得視覺受域特性。所有細胞會依照空間頻率的反應分成雙對比細胞(double-opponent cell)及單對比細胞(single-opponent cell)。我們發現雙對比細胞偏好高空間頻率,且其受域有較高的長寬比,而單對比細胞偏好低空間頻率,其受域長寬比有較低。受域形態與其反應特性之間的相關,在Hartley光柵狀況下遠大於二元雜訊狀況。相較於單對比細胞,雙對比細胞也有較低的調控比率(modulation ratio)。另外一個有趣的發現是,大部分V1色彩細胞都有著偏紅色刺激的受域。總結來說,雙對比和單對比細胞的受域形態有所差異,而由Hartley光柵所測量的視覺受域特性,與方向和空間頻率的反應特性較為一致。雙對比細胞及單對比細胞在同亮度(equiluminance)視覺環境下,可能負責處理不同面向的訊息: 雙對比細胞負責偵測邊界訊息,而單對比細胞負責處理色彩訊息。
Spatial receptive fields have been studied to understand the properties of color- and luminance-responsive neurons in primary visual cortex (V1). In macaque V1, many color-responsive neurons are highly selective for orientation and spatial frequency. One might predict that these highly selective neurons should have receptive fields with multiple and elongated sub-regions (like simple cells). However, previous studies showed that this was not the case — many sharply tuned neurons had blub-like and less elongated receptive fields. Here we measured spatial receptive fields of V1 color-responsive neurons with two different stimulus ensembles: Hartley gratings and binary sparse noise. Both stimulus ensembles consisted of equiluminance red/green colors. Receptive fields were calculated by reverse correlation and fitted with the 2-D Gabor function. Color-responsive neurons were separated into double- and single-opponent cells based on their spatial frequency tunings. We found that double-opponent cells had significantly higher aspect ratios and spatial frequencies than single-opponent cells. The agreements between receptive field properties and orientation tunings were established only in Hartley grating condition but not in sparse noise condition. Double-opponent cells were also found to have lower modulation ratios. Furthermore, most color-responsive cells had red-preferring receptive fields. In summary, double- and single-opponent cells differed in their receptive field properties. For those neurons that were well tuned for orientation and spatial frequency, the tuning properties could be well predicted by Hartley maps but not sparse-noise maps. Our results supported the idea that double-opponent cells might serve as detectors of color edge in the equiluminance scene.