ABSTRACT: Frequency-following responses (FFRs) are scalp-recorded electrophysiological ‘neurophonic’ potentials that reflect phase-locked activity from neural ensembles across the auditory pathway. FFRs provide a neural snapshot of the integrity of supra-threshold speech processing that can be measured non-invasively using a minimal electrophysiological set-up that already exists in audiology clinics, has high test-retest reliability, and requires minimal subject preparation. The original project, titled “Online modulation of auditory brainstem responses to speech”, examined the extent to which FFRs, which were thought to primarily reflect subcortical auditory processing, were influenced by experience-dependent plasticity. The previous proposal systematically tested a predictive tuning model that proposed that subcortical auditory processing is not hard-wired in adults, and that there is continuous fine-tuning of the representation of stimulus features guided by top-down expectations. An evolving perspective is that the FFR should be considered an integrated response from both subcortical and cortical neural ensembles. There is a critical need to understand cortical contributions to the FFR to realize the fundamental translational potential as a biomarker for many clinical conditions. In this renewal application, the primary focus is to understand the properties of the cortical source of the FFR at a mechanistic level, as well as the larger role of cortico-collicular modulatory influences on the FFR. Using a highly complementary and cross-disciplinary team of PIs, this proposal builds on key scientific insights gained in the first funding period with the explicit goal of accelerating pre- clinical to clinical translation. Using a cross-species (human, macaque, guinea pigs), cross-level (cells to meso-scale), neurocomputational approach, this proposal systematically deconstructs the role of the cortex in the generation and modulation of the FFR. Aim 1 will measure scalp- recorded FFRs and intracranial cortical activity in human patients, macaques, and guinea pigs to characterize cortical phase-locking limits, laminar and frequency dependence, and hemispheric asymmetry. Aim 2 will measure scalp-recorded and intracranial FFRs to human and non-human vocalizations using a harmonized protocol in all three species. Using representational similarity analyses to quantify cross-species and cross-level similarities, Aim 2 will examine the influence of predictability, category relevance, and subject arousal on the FFRs. Aim 3 leverages this information to build a novel computational model that consists of a core feedforward module that is modulated by a feedback cortico-collicular module. Predictions from this model will be systematically validated in human patients with Heschl’s gyrus lesions, and using chemogenetic experiments to reversibly suppress cortico-collicular feedback in animal models.