Many neuroscience papers do analyses where they define a particular time period (say when a stimulus is shown), and then do an ANOVA to determine how many neurons are 'selective' during this time period as determined by how many neurons have p-values below a particular alpha level. Here we do a similar analysis where we plot the percentage of selective neurons as a function of time using a sliding 500 ms bin, that is sampled every 5 ms. Neurons here are considered selective if they have an ANOVA p-value of less than 0.01. The results look qualitatively similar to Figs. 2a, 2b, and S3a and S3b. However, in general we prefer to use decoding analyses to report our results since such analyses take into account whether different neurons actually have the same information (e.g., are all neurons only selective to 1 of the 8 stimuli), allow for more complex analyses concerning information coding, and they tend to give more robust results (e.g., less noise in the baseline period, etc.). Upper plots are analyses from the feature experiment and lower plots are from the spatial experiment.
