Friederich Uwe, Billings Stephen A, Hardie Roger C, Juusola Mikko, Coca Daniel
Department of Automatic Control & Systems Engineering, the University of Sheffield, Mappin Street, Sheffield, S1 3JD, United Kingdom.
Department of Physiology, Development and Neuroscience, University of Cambridge, Downing Street, Cambridge, CB2 3DY, United Kingdom.
PLoS One. 2016 Jun 23;11(6):e0157993. doi: 10.1371/journal.pone.0157993. eCollection 2016.
More than five decades ago it was postulated that sensory neurons detect and selectively enhance behaviourally relevant features of natural signals. Although we now know that sensory neurons are tuned to efficiently encode natural stimuli, until now it was not clear what statistical features of the stimuli they encode and how. Here we reverse-engineer the neural code of Drosophila photoreceptors and show for the first time that photoreceptors exploit nonlinear dynamics to selectively enhance and encode phase-related features of temporal stimuli, such as local phase congruency, which are invariant to changes in illumination and contrast. We demonstrate that to mitigate for the inherent sensitivity to noise of the local phase congruency measure, the nonlinear coding mechanisms of the fly photoreceptors are tuned to suppress random phase signals, which explains why photoreceptor responses to naturalistic stimuli are significantly different from their responses to white noise stimuli.
五十多年前就有人提出，感觉神经元能够检测并选择性增强自然信号中与行为相关的特征。虽然我们现在知道感觉神经元经过调整以有效编码自然刺激，但直到现在还不清楚它们编码的刺激的统计特征是什么以及如何编码。在这里，我们对果蝇光感受器的神经编码进行逆向工程，并首次表明光感受器利用非线性动力学来选择性增强并编码时间刺激的相位相关特征，例如局部相位一致性，这些特征在光照和对比度变化时保持不变。我们证明，为了减轻局部相位一致性测量对噪声的固有敏感性，果蝇光感受器的非线性编码机制经过调整以抑制随机相位信号，这就解释了为什么光感受器对自然主义刺激的反应与它们对白噪声刺激的反应有显著差异。
