Spike Train Reproducibility in the LGN

In response to a dynamic visual stimulus, an LGN neuron produces a highly precise temporal pattern of firing. This is illustrated in the figure, where the times of action potentials are represented by dots, time going from left to right, experimental trial from top to bottom. The more reproducible the response to the same stimulus (red rasters), compared to the variability observed using different stimuli (blue rasters), the more information the spike train carries about the stimulus. The advantage of this analysis is that it does not rest on our ability to decode the cell's response, or even to know what it is about a stimulus that the cell encodes.

I have found that LGN firing patterns are highly reproducible and temporally precise, resulting in raw information rates in excess of 100 bits/sec and over 5 bits/spike. The timing of spikes is meaningful to a precision of at least 1 msec. In addition, some neurons encode information in temporal patterns of spikes. I am now interested in the question of how these precise patterns arise, and how cortical neurons might decode them. J.Neurosci.(2000): Abstract

Surprisingly, the precise times of spikes are reprodicible in different cells and even in different individuals. Thus spike timing is a largely universal representation for information about time-varying stimuli encountered in the real world. J Neurosci (2002): Abstract