Subcortical neural generators of the envelope-following response in sleeping children: A transfer function analysis

Abstract

Multiple auditory structures, from cochlea to cortex, phase-lock to the envelope of complex stimuli. The relative contributions of these structures to the human surface-recorded envelope-following response (EFR) are still uncertain. Identification of the active contributor(s) is complicated by the fact that even the simplest two-tone ( f 1 & f 2 ) stimulus, targeting its ( f 2 − f 1 ) envelope, evokes additional linear ( f 1 & f 2 ) and non-linear ( 2 f 1 − f 2 ) phase-locked components as well as a transient auditory brainstem response (ABR). Here, we took advantage of the generalized primary tone phase variation method to isolate each predictable component in the time domain, allowing direct measurements of onset latency, duration and phase discontinuity values from which the involved generators were inferred. Targeting several envelope frequencies (0.22-1 kHz), we derived the EFR transfer functions along a vertical vertex-to-neck and a horizontal earlobe-to-earlobe recording channels, yielding respectively EFR-V and EFR-H waveforms. Subjects ( N = 30) were sleeping children with normal electrophysiological thresholds and normal oto-acoustic emissions. Both EFR-H and EFR-V phase-locking values (PLV) transfer functions had a low-pass profile, EFR-V showing a lower cut-off frequency than EFR-H. We also computed the frequency-latency relationships of both EFRs onset latencies. EFR-H data fitted a power-law function incorporating a frequencydependent traveling wave delay and a fixed one amounting to 1.2 ms. The fitted function nicely fell within five published estimations of the latency-frequency function of the ABR wave-I, thus pointing to a cochlear nerve origin. The absence of phase discontinuity and overall response durations that were equal to that of the stimulus indicated no contribution from a later generator. The recording of an entirely similar EFR-H response in a patient who had severe brainstem encephalitis with a normal, isolated, ABR wave-I but complete absence of later waves, further substantiated a cochlear nerve origin. Modeling of the EFR-V latency-frequency functions indicated a fixed transport time of 2 ms with respect to EFR-H onset, suggesting a cochlear nucleus (CN) origin, here also, without indication for multiple generators. Other features of the EFR-V response pointing to the CN were, at least for the EFR frequency below the cut-off values of the transfer functions, higher PLVs coupled with increased harmonic distortion. Such a behavior has been described in the so-called highly-synchronized neurons of the ventral cochlear nucleus (VCN). The present study compellingly demonstrated the advantage of isolating the EFR in the temporal domain so as to extract detailed spectro-temporal parameters that, combined with orthogonal recording channels, shed new light on the involved neural generators.

Preserved Auditory Steady State Response and Envelope-Following Response in Severe Brainstem Dysfunction Highlight the Need for Cross-Checking

“…This further strengthens the previous conclusion about the cochlear nerve origin of EFR-H waveforms in sleeping children proposed by Lucchetti et al (2021).…”

“…As it is highly improbable that a brainstem compression by the PFT mass effect would have increased conduction velocities along cochlear nerve axons, the shortened delay must have resulted from decreased processing time. The processing of cochlear nerve inputs by cochlear nucleus neurons leading to enhanced PLV outputs is thought to imply monaural coincidence detection (Lucchetti et al 2021) operated by the so-called high-sync cells found in the ventral part of the nucleus (Recio-Spinoso 2012) and identified as spherical and globular bushy cells as well as octopus cells (Keine & Rübsamen 2015; Lu et al 2018). The current view about signal processing by high-sync neurons is that it is controlled by a dynamic balance between excitation and inhibition (Keine & Rübsamen 2015).…”

“…Although the available data did not provide an explanation for this short value, this result was intriguing and called for raising an explanatory hypothesis. In a previous study on EFR transfer functions (Lucchetti et al 2021), the latency difference between the cochlear nerve and cochlear nucleus was modeled as the sum of a fixed transport time with an average value of 1.8 msec and a frequency-dependent processing time amounting to half the EFR period. The expected latency difference would therefore be 4.1 msec based on the model, whereas the measured difference in case PFT was 2.3 msec.…”

“…Recently, capitalizing on the generalization of a stimulus phase rotation method (Whitehead et al 1996) and on the capacity of orthogonal horizontal and vertical montages to segregate peripheral from central components (King et al 2016), two (and only two) distinct sequential EFR components were identified in sleeping children by Lucchetti et al (2018). Results showed that horizontal (earlobe to earlobe) and vertical (vertex to seventh cervical vertebrae) derivations respectively reflected activity from the cochlear nerve and cochlear nucleus (Lucchetti et al 2021). These conclusions were based on the dependency of onset latencies on the tonotopic location of stimulation, the ratios of phase-locking values (PLVs) and harmonic distortion levels between channels in control subjects, and the findings in one patient with brainstem encephalitis.…”