[论文解读] Deep Learning Models of the Retinal Response to Natural Scenes

Lane McIntosh, Niru Maheswaranathan+3 more|arXiv (Cornell University)|Feb 6, 2017

Retinal Imaging and Analysis被引用 164

一句话总结

卷积神经网络能够准确预测自然场景下的视网膜节细胞反应，优于 LN/GLMs，对刺激类型具有泛化能力，并揭示视网膜内部机制。

ABSTRACT

A central challenge in neuroscience is to understand neural computations and circuit mechanisms that underlie the encoding of ethologically relevant, natural stimuli. In multilayered neural circuits, nonlinear processes such as synaptic transmission and spiking dynamics present a significant obstacle to the creation of accurate computational models of responses to natural stimuli. Here we demonstrate that deep convolutional neural networks (CNNs) capture retinal responses to natural scenes nearly to within the variability of a cell's response, and are markedly more accurate than linear-nonlinear (LN) models and Generalized Linear Models (GLMs). Moreover, we find two additional surprising properties of CNNs: they are less susceptible to overfitting than their LN counterparts when trained on small amounts of data, and generalize better when tested on stimuli drawn from a different distribution (e.g. between natural scenes and white noise). Examination of trained CNNs reveals several properties. First, a richer set of feature maps is necessary for predicting the responses to natural scenes compared to white noise. Second, temporally precise responses to slowly varying inputs originate from feedforward inhibition, similar to known retinal mechanisms. Third, the injection of latent noise sources in intermediate layers enables our model to capture the sub-Poisson spiking variability observed in retinal ganglion cells. Fourth, augmenting our CNNs with recurrent lateral connections enables them to capture contrast adaptation as an emergent property of accurately describing retinal responses to natural scenes. These methods can be readily generalized to other sensory modalities and stimulus ensembles. Overall, this work demonstrates that CNNs not only accurately capture sensory circuit responses to natural scenes, but also yield information about the circuit's internal structure and function.

研究动机与目标

• 理解视网膜节细胞如何编码自然场景刺激。
• 评估 CNN 作为预测视网膜对自然场景反应的模型，与 LN 和 GLM 基线相比。
• 研究在刺激分布（自然场景 vs 白噪声）之间的泛化。
• 识别 CNN 捕捉的类视网膜内部机制（抑制、适应性、变异性）。
• 探索架构增强（循环连接）以建模长期尺度的动态。

提出的方法

• 训练深度 CNN 以从自然场景序列和白噪声刺激预测节细胞尖峰。
• 将 CNN 与线性-非线性（LN）和广义线性模型（GLMs）进行比较。
• 使用 ADAM 优化带泊松负对数似然损失；应用 L2 和 L1 正则化。
• 改变网络深度、滤波器尺寸（>15×15）及层类型；在未见数据上评估。
• 可视化学习得到的第一层和第二层感受野以解释特征。
• 可选地用循环层增强 CNN 以捕捉更长时间尺度的适应。

实验结果

研究问题

• RQ1 CNN 是否能在预测自然场景下的视网膜反应方面超越 LN/GLM 模型？
• RQ2 CNN 是否在跨刺激分布（自然场景 vs 白噪声）上具有更好的泛化？
• RQ3 从 CNN 表征中可以看出视网膜的哪些内部机制（例如前馈抑制、适应性、次泊松变异性）？
• RQ4 循环连接是否提升对长期尺度自适应动力学的建模？
• RQ5 自然场景和白噪声训练数据得到的学习特征有何差异？

主要发现

• CNN 在预测自然场景和白噪声下的视网膜反应方面显著超过 LN 模型和 GLMs。
• CNN 的可靠性接近视网膜，并且在跨刺激分布方面的泛化优于更简单的模型。
• 用注入的潜在噪声进行训练可捕捉视网膜尖峰中观测到的次泊松变异性。
• CNN 通过前馈抑制呈现时间上更精确的放电，并在自然场景上显示出更广泛、多样化的第二层特征。
• 通过增加循环横向连接，CNN 能实现对比度自适应，成为对准确响应描述的涌现属性。
• 在一种刺激类别上训练的 CNN 相较于 GLMs 或 LN 模型，对其他刺激类别的泛化能力更强。