Extracellular spike waveform analysis

We used the extracellular waveform width to differentiate between different putative neuronal types (Barthó et al., 2004; Mitchell et al., 2007; Rutishauser et al., 2015; Takahashi et al., 2015). For each neuron, we calculated the trough-to-peak width of the average extracellular action potential. The trough was identified as the time point when the waveform was largest, and the peak is the first local maximum after the trough. The distribution of spike widths was bimodal (see Fig. 4A), as often observed in extracellular recordings. We classified cells as being narrow or wide spiking by performing k-means clustering on the trough-to-peak width of the spikes, selecting for two k-means groups.

Cell type-specific modulation by IEDs. A, Histogram of the distribution of spike widths of all single units analyzed. The two peaks indicate the presence of two distinct populations of neurons with the cutoff at ∼0.5 ms. B, Distribution of spike widths of all the single units after splitting them into two groups: wide waveform cells (mean spike width of 0.81 ± 0.17) and narrow waveform cells (mean spike width of 0.31 ± 0.044 ms). C, Average waveforms of the two groups shown in B. D, Group average PSTH of all modulated wide (left) and narrow width (right) single units across all the brain areas shows that neurons modulated with narrow waveforms on average increase their firing rate during IEDs, whereas the modulation of wide waveform neurons is more heterogeneous, resulting in little on-average modulation. Red dashed line indicates SE across neurons.