Abstract
The correlation among three nerve cells may be plotted as a three-dimensional graph. On the graph's base all possible sequences of three spikes (one from each spike train) may be represented. The relative abundance of each possible sequence is plotted vertically above this base. The correlation may be estimated by dividing the base into triangular bins and counting the number of occurrences of a particular sequence that fall within each bin. Three types of graphs may be constructed. The covariance function for the three spike trains represents the rate of occurrence of three spike sequences, the compound renewal-density function represents the effect of one spike on the combined firing of the other two cells, and the simple renewal-density function represents the way by which the firing times of two cells affect the firing rate of the third one. The covariance function is estimated by dividing the counts in each bin by the total time of measurement. The compound renewal density is estimated by dividing the covariance function by the firing rate of the triggering cell. The simple renewal density is estimated by dividing the covariance function by the cross correlation between the two triggering cells. The use of these analysis methods is illustrated by examination of data from single-unit recordings in the auditory cortex.
Original language | English |
---|---|
Pages (from-to) | 235-239 |
Number of pages | 5 |
Journal | IEEE Transactions on Biomedical Engineering |
Volume | BME-30 |
Issue number | 4 |
DOIs | |
State | Published - Apr 1983 |