Taxonomy of neural oscillation events in primate auditory cortex

Abstract

Electrophysiological oscillations in neocortex have been shown to occur as multi-cycle events, with onset and offset dependent on behavioral and cognitive state. To provide a baseline for state-related and task-related events, we quantified oscillation features in resting-state recordings. We used two invasively-recorded electrophysiology datasets: one from human, and one from non-human primate auditory system. After removing event related potentials, we used a wavelet transform based method to quantify oscillation features. We identified about 2 million oscillation events, classified within traditional frequency bands: delta, theta, alpha, beta, gamma, high gamma. Oscillation events of 1-44 cycles were present in at least one frequency band in 90% of the recordings, consistent across human and non-human primate. Individual oscillation events were characterized by non-constant frequency and amplitude. This result naturally contrasts with prior studies which assumed such constancy, but is consistent with evidence from event-associated oscillations. We measured oscillation event duration, frequency span, and waveform shape. Oscillations tended to exhibit multiple cycles per event, verifiable by comparing filtered to unfiltered waveforms. In addition to the clear intra -event rhythmicity, there was also evidence of inter -event rhythmicity within bands, demonstrated by finding that coefficient of variation of interval distributions and Fano Factor measures differed significantly from a Poisson distribution assumption. Overall, our study demonstrates that rhythmic oscillation events dominate auditory cortical dynamics.

Methodological Considerations for Studying Neural Oscillations

“…However, in empirical neurophysiological data, oscillations show large variability in their presence and extent across time, as well as across participants and cortical regions (Donoghue, Haller, et al, 2020;Frauscher et al, 2018;Groppe et al, 2013). Even when oscillations are present, they are highly variable (Jones, 2016;Neymotin, Barczak, et al, 2020), waxing and waning in short bursts and including longer, more tonic rhythms, with rapidly changing amplitude, frequency, and phase dynamics that are not easily captured by common analyses and predefined canonical frequency ranges. All of these properties, and more, need to be explicitly considered in order to accurately and reliably measure oscillatory neural activity.…”

Oscillatory Bursting as a Mechanism for Temporal Coupling and Information Coding

“…Lundqvist et al ( 2016 ) used a similar approach to detect an increase in gamma power of two standard deviations above the mean spectral power in that band, but also added a duration constraint of an increase lasting at least three cycles. Neymotin et al ( 2020a ) demonstrates that below three cycles, any “length measure” becomes unreliable, in that it overestimates the number of cycles. Hughes et al ( 2012 ) used an oscillation detection method to extract both sustained and transient rhythms from rat hippocampal recordings termed Better Oscillation Detection Method [BOSC; first described by Caplan et al ( 2001 )].…”

“…Even though these are merely selected examples and should not be considered as evidence for the specific durations of oscillatory activity, such recordings give us confidence that the basic phenomenon of a bursty oscillation exists. Recently, Neymotin et al ( 2020a ) quantified oscillation event features in resting-state invasive recordings from auditory cortex of humans and non-human primates. To our knowledge, this was the first attempt to characterize ongoing (single trial) oscillatory activity across different frequencies and species.…”

“…The BOSC method (Caplan et al, 2001 ; Hughes et al, 2012 ) attempts to avoid the detection of transient, non-oscillatory events by identifying oscillatory episodes at each frequency using both power and duration thresholds. Another approach to avoid bias due to externally-driven events removed the average evoked-responses waveforms from each cortical layer (Neymotin et al, 2020a ), though if this removal uses simple subtraction of the trial-averaged response from each single trial, it runs the risk of creating artifacts (Knuth et al, 2006 ). Tal and Abeles ( 2016 ) used an algorithm that may detect both oscillatory and non-oscillatory transient events using a template matching scheme.…”

“…For example, providing brief, strong thalamocortical activation to a hypothetical neocortical model would trigger production of a transient ERP-like event in the circuit, with a characteristic waveform (Neymotin et al, 2020b ). Although running a wavelet filter on such a waveform will produce high power at a frequency inverse to the duration of the ERP, since it was produced by a punctate event, this type of waveform should not be considered an oscillation (Neymotin et al, 2020a ). In contrast, specific synaptic connectivity and input patterns provided to a circuit model lead to production of sustained multi-cycle oscillations.…”