A data analysis system for neural unit data, a stochastic point process, which borrows from data base management technology has been written. The data analysis is controlled by a supervisory system which consists of a command interpreter, data file handler, and program initiator. A description of the data file is stored and is accessible to the supervisory system along with the data requirements for each analysis program. The supervisory system passes data from the data files to the analysis programs. The data consist of sequences of neural unit data which can be collected as part of a response area paradigm, that is, a two-dimensional stimulus space which is explored systematically, and the response of the neural unit is reduced to a sequence of threshold crossing times. A variety of analysis can be performed on this data including Poststimulus time histograms, Interval histograms, Joint interval histograms, Latency histograms, Period histograms, Latency dot displays, Autocorrelograms, Calculation of phase and synchronization coefficients, Spike versus 'variable' curves, Isorate curves, Cumulative phase curves. The versatility of the system makes it relatively easy to include additional analysis in the system. Each result can be displayed, listed on a printer, plotted or written on a disc file for later analysis. Nearly all the programs have been coded in Fortran.
Journal description