Evaluation of the temporal correlation between two spike trains is an important issue when studying the dynamics of local neural networks. In the postcentral somatosensory cortex (SI), the temporal profile of neural activity during sustained stimulation is quite variable on the order of 100 msec, reflecting various thalamocortical inputs as well as various stages of the integrating process that might take place within local neural networks. Therefore, a numerical method is needed to evaluate the difference in the gross temporal profile over several hundreds of msec or more. To address this requirement, we tested a method using the dot product between a pair of spike trains. We applied the method to our own data, i. e., 88 pairs of SI neurons recorded simultaneously in one conscious monkey. Each discrete spike train was first smoothed by convolving with a Gaussian function, then the normalized dot product (similarity index) was taken between a pair of spike trains. We next conducted a simulation study using random spike trains (surrogate data set), in which the number of trials and the number of spikes of both units in each trial were set strictly to those of the original data set. The simulation was repeated 100 times for each pair, and the mean and standard deviation of the similarity index were taken. Finally, the difference in the similarity indexes was determined between the original and surrogate data set (z-score). The method successfully quantifies the gross temporal difference of activity in each pair of SI neurons.
ASJC Scopus subject areas
- Medicine (miscellaneous)
- Biochemistry, Genetics and Molecular Biology(all)