Figure 3.

Stereotyped neurons as population reference for stimulus onset in auditory cortex. (a) Sound wave and spectrogram for the auditory stimulus sequence on one example trial. Twelve different natural sounds (stimuli) were presented in pseudo-random sequence and with random intersound intervals (blue periods) on each trial. The sound wave is shown above the spectral representation (red colours indicate high power). (b) (i) Time course of trial-averaged responses to all 12 stimuli (each row represents a different stimulus) for one stereotyped (left) and one modulated (right) neuron. Time t = 0 corresponds to sound onset. (ii) Distribution of the latency variability (standard deviation across trials) and mean latency across neurons. Stereotyped and modulated neurons are colour-coded in red and blue, respectively, the dashed line in the left histogram indicates the threshold used to separate (i.e. define) the two groups. (c) Example response from one modulated neuron after being aligned to different reference frames. (i) Spike raster when aligned to stimulus onset time (t = 0). (ii) The same response but with each trial aligned to the response onset of a simultaneously recorded stereotyped reference neuron (here t = 0 corresponds to the onset latency of the reference neuron). (iii) The same response aligned to the onset of a modulated neuron. While the stereotyped reference preserves the temporal shape of the stimulus locked response, the use of a modulated neuron as reference results in a much more dispersed spike raster. (d) Information about stimulus identity obtained using the three considered reference frames in progressively longer time windows (starting at t = 0 and ending at each indicated time point). Lines denote the mean and shaded areas the standard error (s.e.m.) across the population (n = 48) of modulated neurons. Information was computed with a linear decoder and equation (2.2). Figure redrawn from [46].