Skip to main content
. 2023 Jan 4;43(1):93–112. doi: 10.1523/JNEUROSCI.1616-21.2022

Figure 9.

Figure 9.

Coherent population responses follow peak derivative events during slow modulations. We measured evoked responses in the global average firing rate of RS cells, reasoning that if a given amplitude feature evokes spiking in many cells, this coherent response would be reflected in the mean population activity. A, A schematic illustrating the identification of peakDrv events (orange circles) in sinusoidal AM stimuli. The left panel marks these events on a linear amplitude scale. For this analysis, peakDrv landmarks were identified within the logarithmically-transformed amplitude signal (right panel), corresponding to perceptual space. Note that, when the amplitude is expressed on a dB scale, peakDrv events coincide with the onsets (minima) of sinusoidal amplitude cycles (yellow circles). B, peakDrv events were identified within the log-transformed envelopes of vocoded speech stimuli. To focus on perceptually relevant amplitude modulations, we restricted analyses to peakDrv events that occurred between a local minimum followed by a local maximum at least 6 dB higher. The schematic in panel B illustrates the peakDrv events identified within an example speech segment. C, The stimulus envelope and corresponding RS population activity are depicted using the conventions of Figure 8. The trace at the bottom shows the mean activity across all RS cells for this stimulus segment. D, Mean population activity surrounding peakDrv landmarks was averaged across all events, plotted for each periodic modulation rate. Slower rates evoke stronger phasic activity in the population, and the short, positive latencies suggest that peakDrv could have a causal relationship to coherent activity in the population. E, Evoked population activity is plotted after averaging across peakDrv events in all vocoded speech stimuli (mean ± SD). This stimulus feature evoked a strong, phasic response in the aggregate population with ∼45-ms latency. The rising edge of the evoked response is sharp, as would be expected if the causal stimulus landmark were aligned in time.