Issues Concerning the Nature of Neuronal Response
A person whose intuition of neuronal reliability is shaped by post-stimulus time (PST) histograms (Gerstein and Kiang, 1960) and dot raster displays (see for example Schmidt et al, 1975) (myself included, in the early papers Legéndy, 1970, 1975) may well be tempted to assume that the single neuron is only reliable in broad statistical terms. To elicit reproducible behavior from a neuron of the visual cortex, for instance, one must sweep the receptive field 10–20 times, combine the sweeps into a PST histogram or a raster display, and examine the way the spikes are distributed. Two or three sweeps do not appear to be enough to get a clear idea of the neuron’s behavior, because, as seen in Fig. 2.1, the responses do not repeat spike-for-spike; some sweeps show more spikes, some fewer.