Department of Neuroscience, Columbia University, New York, NY, United States.
Front Neural Circuits. 2018 Apr 4;12:26. doi: 10.3389/fncir.2018.00026. eCollection 2018.
Confronted with an ever-changing visual landscape, animals must be able to detect relevant stimuli and translate this information into behavioral output. A visual scene contains an abundance of information: to interpret the entirety of it would be uneconomical. To optimally perform this task, neural mechanisms exist to enhance the detection of important features of the sensory environment while simultaneously filtering out irrelevant information. This can be accomplished by using a circuit design that implements specific "matched filters" that are tuned to relevant stimuli. Following this rule, the well-characterized visual systems of insects have evolved to streamline feature extraction on both a structural and functional level. Here, we review examples of specialized visual microcircuits for vital behaviors across insect species, including feature detection, escape, and estimation of self-motion. Additionally, we discuss how these microcircuits are modulated to weigh relevant input with respect to different internal and behavioral states.
面对不断变化的视觉环境，动物必须能够检测相关刺激并将此信息转化为行为输出。一个视觉场景包含大量信息：要完全解释它将是不经济的。为了最佳地执行此任务，存在神经机制来增强对感觉环境中重要特征的检测，同时同时过滤掉无关信息。这可以通过使用实现特定“匹配滤波器”的电路设计来完成，这些滤波器针对相关刺激进行了调整。根据这条规则，昆虫的特征明显的视觉系统已经进化到在结构和功能水平上简化特征提取。在这里，我们回顾了昆虫物种中各种重要行为的专门视觉微循环的例子，包括特征检测、逃避和自身运动估计。此外，我们还讨论了这些微循环如何根据不同的内部和行为状态来调整以权衡相关输入。
