Purchase this article with an account.
J.M. Hitt, F.A. Dodge, R.B. Barlow; Information Content in Responses of Limulus Optic Nerve Fibers . Invest. Ophthalmol. Vis. Sci. 2003;44(13):3253.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To measure information content of optic nerve spike trains generated by natural stimulation during the day and at night in the horseshoe crab, Limulus polyphemus. Methods: We recorded optic nerve activity from a single lateral eye in vivo responding to natural scenes day and night. All experiments were done in the laboratory with visual stimuli composed of natural scenes recorded underwater with a miniature camera (CrabCam) attached to a horseshoe crab. The visual stimuli were presented on a CRT monitor to mimic daytime conditions. For nighttime conditions the monitor output was attenuated by 4 log to match natural lighting at night. We calculated the information transmitted by the spike trains by cross-validating the response set and assigning responses to stimulus or noise categories based on their similarities to other responses in the data set. Spike train comparisons were calculated with the spike train similarity metric developed by Victor and Purpura. Responses were also analyzed with a theoretical object detector that responded to deviations of the response rate from its mean. Responses of the theoretical detector were analyzed with various thresholds in order to determine whether a response arose from signal or noise. Results: The spike train similarity metric distinguished signals from noise and separated responses to identical stimuli recorded from different crabs. Transmitted information is maximal (> 90%) when analyzed with a time window of 0.1 to 3.0 sec and declines to 50% to 60% correct when counting spikes and discarding all temporal information. The theoretical object detector correctly identified 90% of the stimuli when an optimal threshold was applied to the detector’s response rate. Conclusions: (1) Retinal signals analyzed with the spike train similarity metric and a theoretical object detector can be correctly identified as signal or noise. (2) The performance of these two analyzers match or exceed the visual performance of the horseshoe crabs measured psychophysically. Support: NSF, NIMH, NEI, Research to Prevent Blindness, and Lions of Central NY
This PDF is available to Subscribers Only