The messages in optic nerve fibers and their interpretation

Spike discharges are the principal carriers of information in the nervous system. Although both the ionic and the molecular mechanisms of spike generation have been studied extensively, the methods for analyzing a spike train that are currently employed have not changed much from those in use 20 years ago. There is an apparent need for a refinement of the methods used to analyze spike trains. We present here a summary of our recent results of an analysis of spike trains from retinal ganglion cells that is based on Wiener's theory of non-linear analysis or white-noise analysis. We found that spike trains carry, at least to a second-order approximation, as much information as is carried by the ganglion cell's postsynaptic potential (PSP). There is no loss of information when an analog signal, PSP, is converted into a point process, namely, spike discharges. It is indeed possible to predict the cell's PSP from a spike train. This finding has two important implications. First, the neuron network in the retina produces a PSP, the dynamics of which are optimal for triggering a spike discharge, or conversely, the spike-generation mechanism is optimized to match the dynamics of the network. The external stimulus that is optimal for production of a ganglion-cell discharge is represented as the cell's PSP. Second, there is structure encoded within the spike train; information on a second-order non-linearity is encoded by the relative timing of two consecutive spike discharges. Coding of non-linearity into a spike train is an efficient means of signal compression and is an important aspect of neurophysiology.