Department of Biomedical Engineering, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel.
Eur J Neurosci. 2012 Feb;35(3):436-44. doi: 10.1111/j.1460-9568.2011.07971.x.
Archer fish are known for their unique hunting method, where one fish in a group shoots down an insect with a jet of water while all the other fish are observing the prey's motion. To reap its reward, the archer fish must reach the prey before its competitors. This requires fast computation of the direction of motion of the prey, which enables the fish to initiate a turn towards the prey with an accuracy of 99%, at about 100 ms after the prey is shot. We explored the hypothesis that direction-selective retinal ganglion cells may underlie this rapid processing. We quantified the degree of directional selectivity of ganglion cells in the archer fish retina. The cells could be categorized into three groups: sharply (5%), broadly (37%) and non-tuned (58%) directionally selective cells. To relate the electrophysiological data to the behavioral results we studied a computational model and estimated the time required to accumulate sufficient directional information to match the decision accuracy of the fish. The computational model is based on two direction-selective populations that race against each other until one reaches the threshold and drives the decision. We found that this competition model can account for the observed response time at the required accuracy. Thus, our results are consistent with the hypothesis that the fast response behavior of the archer fish relies on retinal identification of movement direction.
射水鱼以其独特的狩猎方式而闻名，一群鱼中的一条鱼会用一股水流击落昆虫，而其他所有鱼都在观察猎物的运动。为了获得奖励，射水鱼必须在其竞争对手之前到达猎物。这需要快速计算猎物的运动方向，使鱼能够在猎物被击落约 100 毫秒后，以 99%的准确度向猎物转弯。我们探讨了一个假设，即方向选择性的视网膜神经节细胞可能是这种快速处理的基础。我们量化了射水鱼视网膜中神经节细胞的方向选择性程度。这些细胞可以分为三组：锐向（5%）、宽向（37%）和非调向（58%）方向选择性细胞。为了将电生理数据与行为结果联系起来，我们研究了一个计算模型，并估计了积累足够方向信息以匹配鱼的决策精度所需的时间。该计算模型基于两个相互竞争的方向选择性群体，直到其中一个达到阈值并驱动决策。我们发现，这种竞争模型可以解释观察到的响应时间和所需的准确性。因此，我们的结果与假设一致，即射水鱼的快速反应行为依赖于视网膜对运动方向的识别。
