Vibrotactile sensation is processed in the somatosensory cortex in the brain. Neurons can encode different properties of stimuli, including the frequency or magnitude of a sound or vibrotactile sensation, in their firing frequency. Pyramidal neurons in the primary sensory cortex (S1), however, have a limited firing rate and cannot match the frequency of fast vibrotactile inputs one-to-one. To assess how pyramidal neurons encode features of vibrotactile sensations delivered by primary afferents, UTCSP member Dr. Steven Prescott and colleagues created computational simulations and performed electrophysiological recordings on slices of these neurons in S1 to explore individual and population-level coding.
Written by:
Quinn Pauli
Edited by:
Georgia Hadjis
Leading up to Canada’s legalization of cannabis, cannabis companies partnered with academics to research the understanding, medical potential, and creation of cannabis compounds to make research claims for the cannabis market. UTCSP scientist Dr. Daniel Buchman, UofT Lawrence S. Bloomberg Faculty of Nursing co-PI Dr. Quinn Grundy, and colleagues conducted a meta-research study to 1) identify research with statements of disclosure addressing funding or financial relationships with Canadian cannabis companies, 2) describe the research being conducted with cannabis companies and sponsorship towards the studies, and 3) identify the demographics of the participants included.
An adaptive exponential integrate-and-fire computational model was first adjusted to match the properties of S1 pyramidal neurons. To simulate physiological electrical “noise” due to random opening and closing of ion channels or background synaptic input, the model was tested in the presence or absence of uncorrelated noise. Using this model, the authors found that without the inclusion of noise, higher frequency inputs were coded in a nearly identical pattern as lower frequency inputs due to limitations in neuronal firing rates. However, when noise approximating in vivo conditions was included in the model, cycles of stimulation were skipped irregularly, resulting in spiking intervals that correlated with the frequency of the input. By delivering and recording electrical signals from neurons directly, the authors also found that S1 neurons fire intermittently during high frequency inputs, which allowed population responses to be modulated in sync with input frequency.
The authors found that without the inclusion of noise, higher frequency inputs were coded in a nearly identical pattern as lower frequency inputs due to limitations in neuronal firing rates. However, when noise approximating in vivo conditions was included in the model, cycles of stimulation were skipped irregularly, resulting in spiking intervals that correlated with the frequency of the input.
Together, the simulations and electrophysiological recordings indicate that S1 neurons can encode information about the frequency and amplitude of periodic vibrotactile inputs using firing intervals and rate, respectively. Crucially, this coding is aided by physiological noise, which modulates the reliability of firing without disrupting its precision. Overall, this study provides novel insights on how neurons encode multiple stimulus features in their firing rates.
