Dey Ashim, Zele Andrew J, Feigl Beatrix, Adhikari Prakash
Centre for Vision and Eye Research, Queensland University of Technology (QUT), Brisbane, QLD 4059, Australia; School of Biomedical Sciences, Queensland University of Technology (QUT), Brisbane, QLD 4059, Australia.
Centre for Vision and Eye Research, Queensland University of Technology (QUT), Brisbane, QLD 4059, Australia; School of Optometry and Vision Science, Queensland University of Technology (QUT), Brisbane, QLD 4059, Australia.
Vision Res. 2021 Mar;180:1-10. doi: 10.1016/j.visres.2020.11.010. Epub 2020 Dec 21.
At the absolute threshold of vision, Hecht, Shlaer and Pirenne estimate that 5-14 photons are absorbed within a retinal area containing 500 rods. Other estimates of scotopic threshold vision based on stimuli with different durations and focal areas range up to ~100,000 photons. Given that rod density varies with retinal eccentricity and the magnitude of the intrinsic noise increases with increasing stimulus area and duration, here we determine whether the scotopic threshold estimates with focal stimuli can be extended to full-field stimulation and whether summation explains inter-study differences. We show that full-field threshold vision (1018 mm, 10 ms duration) is more sensitive than at absolute threshold, requiring the absorption of ~1000 photons across ~91.96 million rods. A summation model is presented integrating our and published data and using a nominal exposure duration, criterion frequency of seeing, rod density, and retinal area that largely explains the inter-study differences and allows estimation of rods per photon ratio for any stimulus size and duration. The highest signal to noise ratio is defined by a peak rod convergence estimated at 53:4:1:2 (rods:rod bipolar cells:AII amacrine cells:retinal ganglion cells), in line with macaque anatomical estimates that show AII amacrine cells form the bottleneck in the rod pathway to set the scotopic visual limit. Our model estimations that the rods per photon ratio under full-field stimulation is ~3000X higher than at absolute threshold are in accordance with visual summation effects and provide an alternative approach for understanding the limits of scotopic vision.
在视觉绝对阈值下，赫克特、施莱尔和皮雷纳估计，在一个包含约500个视杆细胞的视网膜区域内，有5 - 14个光子被吸收。基于不同持续时间和焦点区域刺激的暗视觉阈值的其他估计范围高达约100,000个光子。鉴于视杆细胞密度随视网膜离心率而变化，且内在噪声的幅度随刺激面积和持续时间的增加而增加，在此我们确定使用焦点刺激的暗视觉阈值估计是否可以扩展到全视野刺激，以及总和是否解释了研究间的差异。我们表明，全视野阈值视觉（~1018毫米，持续时间10毫秒）比绝对阈值时更敏感，在约9196万个视杆细胞上需要吸收约1000个光子。提出了一个总和模型，整合了我们的数据和已发表的数据，并使用了名义曝光持续时间、看见的标准频率、视杆细胞密度和视网膜面积，该模型在很大程度上解释了研究间的差异，并允许估计任何刺激大小和持续时间下的每个光子的视杆细胞比率。最高信噪比由估计为53:4:1:2（视杆细胞:视杆双极细胞:AII无长突细胞:视网膜神经节细胞）的视杆细胞峰值汇聚定义，这与猕猴的解剖学估计一致，表明AII无长突细胞在视杆细胞通路中形成瓶颈，从而设定暗视觉极限。我们的模型估计全视野刺激下每个光子的视杆细胞比率比绝对阈值时高约3000倍，这与视觉总和效应一致，并为理解暗视觉的极限提供了一种替代方法。
