Sample neurons with their IDs in the datasets. The five panels from left to right are as follows: i, temporal evolutions of the mean firing rate under each stimulus condition, denoised by projecting to the first principal components capturing 80% of the variance of population activity of all well-tuned neurons in the same dataset (Murray et al., 2017); ii, spike raster plot under the preferred stimulus; iii, spike raster plot under the least preferred stimulus; iv, bar plot of average FF during foreperiod, during the stimulus-selective segments of the delay under the preferred stimulus, and during the stimulus-selective segments of the delay under the least preferred stimulus; and v, scatterplot of average FF versus mean firing rate across stimulus conditions during the stimulus-selective segments of the delay. A, A neuron with nonbursting stable activity in the ODR task. B, The neuron with maximal bursting behaviors in the ODR task. C, A neuron with nonbursting stable activity in the VDD task. D, The neuron with maximal bursting behaviors in the VDD task. E, A neuron with nonbursting stable activity in the MNM task. F, The neuron with maximal bursting behaviors in the MNM task.