Age-Dependent Auditory Plasticity: Critical Periods in Subcortical and Cortical Encoding

Search strategy

A systematic search was performed in PubMed/MEDLINE, Scopus, and Web of Science databases for the period January 1963 to December 2025. The lower date boundary was chosen to capture foundational work on critical periods originating with Hubel and Wiesel [3], while the upper boundary ensured inclusion of the most recent molecular and clinical evidence. Searches were conducted independently in each database using the following search string, applied consistently across all three platforms: (“auditory plasticity” OR “critical period” OR “sensitive period”) AND (“subcortical” OR “brainstem” OR “inferior colliculus” OR “medial geniculate” OR “cortical” OR “auditory cortex”) AND (“age” OR “development” OR “aging” OR “maturation”). In PubMed/MEDLINE, search terms were additionally mapped to Medical Subject Headings (MeSH) where available; free-text searching was used in parallel to maximize retrieval sensitivity. No language filter was applied at the database search stage. Reference lists of all included articles were additionally hand-searched to identify relevant studies not captured by the electronic search alone.

Inclusion and exclusion criteria

Inclusion criteria were: 1) experimental or review studies examining auditory plasticity at subcortical and/or cortical levels; 2) studies directly addressing age-dependent changes, critical periods, or sensitive periods in the auditory system; 3) peer-reviewed articles published in English. Exclusion criteria were: 1) studies addressing only peripheral hearing loss at the cochlear level without examining central auditory consequences; 2) studies focused exclusively on other sensory modalities (e.g., visual, somatosensory) with no generalization to the auditory system; 3) case reports, editorials, conference abstracts, and letters to the editor; 4) non-English language publications. Studies using visual or somatosensory models were retained when they provided direct mechanistic evidence applicable to auditory critical period regulation (e.g., PNN biology, excitatory/inhibitory [E/I] balance).

Screening process and data synthesis

A total of 3,128 records were retrieved across the three databases. Duplicate removal was performed using reference management software (Zotero v6.0; Corporation for Digital Scholarship), with automated detection supplemented by manual verification, yielding 2,437 unique records. All 2,437 records underwent title and abstract screening against the inclusion and exclusion criteria, after which 312 full texts were retrieved for detailed assessment. Of these 312 full texts, 234 were excluded for the following reasons: focused exclusively on peripheral cochlear pathology without addressing central plasticity (n=89); non-auditory sensory modality with no generalization to auditory system (n=54); case reports, editorials, or conference proceedings (n= 47); insufficient data on age-dependent or critical/sensitive period plasticity (n=31); non-English language (n=13). This yielded 78 studies meeting all inclusion criteria (Fig. 1). Of these, 31 examined subcortical plasticity primarily, 29 examined cortical critical periods primarily, and 18 investigated both levels or clinical applications. All 78 included studies informed the thematic synthesis; a representative subset is directly cited in the narrative text, while the complete set of included studies, organized by developmental stage and auditory neuraxis level, is presented in Supplementary Table 1 (in the online-only Data Supplement). Data charting was performed by a single reviewer (S.E.U.) using a standardized extraction form. Each included study was coded for: 1) level of investigation (subcortical/cortical/both); 2) species model (human/ rodent/non-human primate/mixed); 3) primary methodology (electrophysiology/molecular/imaging/behavioral/clinical/review); 4) age group or developmental stage examined; and 5) main finding relevant to auditory plasticity. Thematic synthesis was used to integrate findings across studies, with themes generated inductively from the data and then organized according to the GWM principles.

Fig. 1.

PRISMA-ScR flow diagram of the study selection process. Records were retrieved from three databases (PubMed/MEDLINE, Scopus, Web of Science; January 1963–December 2025). Following duplicate removal (n=691), title/abstract screening (n=2,437), and full-text assessment (n=312), 78 studies met all inclusion criteria. Full-text exclusion reasons are shown within the diagram.

Developmental trajectory of the auditory brainstem

The traditional view holds that ABR properties reach adultlike maturity within the first two postnatal years [13]. This long-standing view was substantially revised by the landmark study of Skoe, et al. [9], who recorded ABR to both click and speech stimuli in 586 normal-hearing individuals. Their findings suggest that developmental changes extend well beyond 2 years of age, that a pronounced developmental “overshoot” period may exist between ages 5–11, and that different response components follow independent developmental trajectories. This finding suggests that the brainstem may not be a monolithic structure but rather an assembly of functionally specialized circuits developing along independent timelines.

However, the cross-sectional design of Skoe, et al. [9] limits causal inferences; longitudinal data are needed. The scarcity of data from the 1–2 year age range prevents complete mapping of developmental transitions. Future studies should prioritize longitudinal designs and source localization methods.

Tonotopic critical period

Barkat, et al. [4] showed in mice that tone exposure altered primary auditory cortex (A1) tonotopic maps within a three-day window but did not change maps in the ventral medial geniculate body (MGBv), a finding consistent with cortical primacy during thalamocortical development. Deletion of the Icam5 molecule accelerated spine maturation, shifting the critical period earlier. de Villers-Sidani, et al. [5] defined the spectral tuning critical period as P11–P13 and frequency sweep sensitivity as approximately P14 in rats. Current single-nucleus RNA sequencing (snRNA-seq) studies revealed that tone exposure alters gene expression patterns in interneurons (Kcnc1, Nrgn increase), reduces myelin genes (Mbp) in oligodendrocytes, and reverses PNN core protein expression [22].

Inhibitory circuits and molecular brakes as gatekeepers

The E/I balance hypothesis proposes that maturation of PV+ interneurons initiates the critical period, while consolidation of PNNs closes it [1,23]. PNNs are chondroitin sulfate proteoglycan networks surrounding the soma and proximal dendrites of PV+ cells. They form in an activity-dependent manner during development and function as molecular brakes restricting plasticity upon maturation [24-26]. Cisneros-Franco and de Villers-Sidani [10] showed that chemogenetic silencing of PV+ interneurons in the adult auditory cortex reactivated critical period-like plasticity, paralleled by anatomical changes in PNN expression. Takesian, et al. [27] showed that L1 interneurons serve as a nodal point where thalamocortical inputs and neuromodulatory signals converge, and that the Lynx1 protein may contribute to critical period closure by modulating nicotinic acetylcholine receptor (nAChR) sensitivity.

Rupert and Shea [28] summarized shared features of plasticity during developmental critical periods and adult learning: 1) cortical disinhibition through decreased stimulus-evoked firing rates of PV+ cells, 2) connectivity remodeling of PV+ inputs to pyramidal cells, and 3) elevation of PV and PNN expression marking window closure. These shared mechanisms suggest that developmental and adult plasticity may share overlapping molecular substrates, although direct causal evidence in the human auditory system remains limited.

cMost cortical critical period literature derives from rodent models. The human auditory cortex has a much longer developmental trajectory, complex laminar architecture, and species-specific experiences. Rodent timelines cannot be directly transferred; however, core mechanisms such as E/I balance, PNNs, and Lynx1 may be conserved. Human-scale biomarkers are urgently needed, such as GABA-edited MR spectroscopy and electroencephalography (EEG) gamma oscillations.

Sequential critical periods and hierarchical organization

The cortex does not have a single monolithic critical period but rather temporally sequenced multiple windows. Tonotopic map reorganization closes earliest; tuning bandwidth, binaural processing, and frequency modulation selectivity follow sequentially [29,30]. Kral [2] proposed that this sequencing reflects a transition from feature representation to object representation. Environmental acoustic enrichment can support restoration of degraded cortical processing [31,32].

Core principles

Principle 1—Graded Opening/Closing: plasticity windows open sequentially along the auditory neuraxis from bottom to top. Each level is governed by its own molecular onset and closure triggers. Supporting evidence includes: early maturation of auditory brainstem responses [13] and rodent evidence showing defined cortical critical periods in A1 [4,5]; ABR components mature earlier than cortical CAEP components in humans [7,9]. Unresolved questions: whether windows overlap partially or are strictly sequential in humans; which molecular signals initiate opening at subcortical levels. Testable prediction: see Hypothesis 2. Principle 2—Bidirectional Coupling: subcortical and cortical plasticity interact in both directions. During development, cortical maturation appears to influence subcortical encoding via corticofugal pathways; in adulthood, local homeostatic mechanisms become more prominent. Supporting evidence includes (primarily animal): corticofugal projections modulate subcortical responses [33,34]; musical training effects on human brainstem encoding are consistent with top-down modulation, though causality has not been established [14,16]. Unresolved question: whether subcortical plasticity in humans is primarily driven by cortical input or local mechanisms. Testable prediction: see Hypothesis 1. Principle 3—Feature-Specific Timing: different auditory features have distinct critical/sensitive period timelines. Simpler features (e.g., basic frequency tuning) close earlier; complex features (e.g., binaural processing, speech-in-noise) close later. Supporting evidence includes: feature-dependent sensitive periods in rodent A1 [29,30]; de Villers- Sidani, et al. [5] identified distinct windows for spectral tuning (P11–P13) and frequency sweep sensitivity (approximately P14) in rats. Unresolved question: precise human timelines for feature-specific windows remain unmapped. Testable prediction: see Hypothesis 2. Principle 4—Reverse Cascade: in aging, peripheral deafferentation triggers a maladaptive cascade of central reorganization that propagates from lower to higher neuraxis levels. Supporting evidence includes (primarily rodent): age-related GABAergic and glycinergic decline in CN and IC [18]; altered central gain in aged rats [19]. Unresolved questions: whether the cascade is sequential or parallel across levels; whether early amplification can interrupt its propagation. Testable prediction: see Hypothesis 3.

Testable hypotheses

The following hypotheses are derived from the GWM to enable experimental testing:

Hypothesis 1: optogenetic silencing of corticofugal pathways during development will reduce subcortical plasticity during the cortical critical period but will have minimal effect during the subcortical critical period (test of Principle 2).

Hypothesis 2: different components of the brainstem response (onset latency vs. frequency encoding vs. response consistency) will have different developmental peaks, and experience-dependent plasticity will be observed only in components where developmental change continues (test of Principle 3).

Hypothesis 3: in age-related hearing loss, early amplification (hearing aid) will preserve inhibitory neurotransmitter levels in the CN and IC compared to unamplified controls (test of Principle 4).

Hypothesis 4: chemogenetic silencing of PV+ interneurons or enzymatic degradation of PNNs in the adult auditory cortex, combined with auditory training, will produce greater frequency map reorganization than training alone (pharmacological reopening hypothesis).

Hypothesis 5: in late-implanted children (>7 years), subcortical cABR components (temporal processing) will show greater improvement than cortical P1 components, because the subcortical window closes later than the cortical window (clinical prediction of the GWM).

The GWM is a conceptual framework, not an empirically established model. Its principles are drawn from converging but indirect evidence across species and levels of analysis. None has been validated through simultaneous multi-level recordings in the same organism across the full lifespan. Principle 1 (graded sequencing) is inferred from separate studies at different neuraxis levels, often in different species. Direct validation requires simultaneous multi-level recordings in the same individual. Principle 2 (bidirectional coupling) rests primarily on animal optogenetic and chemogenetic data; its relevance to humans remains inferential. Principle 3 (feature-specific timing) is well-supported in rodents but has not been systematically mapped in humans. Principle 4 (reverse cascade) is consistent with epidemiological and rodent data; whether aging changes develop sequentially or in parallel requires longitudinal verification. The five testable hypotheses below are designed to address these gaps and provide empirical grounding for each principle.

Cochlear implantation timing

Sharma, et al. [7,35] identified a sensitive period of approximately 3.5 years using the cortical auditory evoked potential (CAEP) P1 component as a biomarker. In a study of 245 congenitally deaf children, the majority implanted before age 3.5 achieved normal P1 latencies, while nearly all implanted after age 7 showed abnormal responses. Results between 3.5–7 years were variable. Earlier cochlear implantation has been associated with more favorable central auditory development and auditory outcomes [7,35]. Cross-modal reorganization (visual and somatosensory areas invading the auditory cortex) observed in late-implanted children is thought to constitute a key biological basis of the sensitive period [36,37], although the causal relationship between cross-modal reorganization and outcomes remains an area of active investigation. Clinical Recommendation 1 (Evidence Level: Moderate-High | Evidence Type: human prospective cohort data and CAEP biomarker studies; no RCT is ethically feasible for this outcome): the available human evidence suggests that cochlear implantation in congenitally deaf children is associated with more favorable outcomes when performed before 3.5 years of age, with the strongest outcomes associated with implantation within the first year of life. Late implantation should not be abandoned; the GWM predicts that subcortical plasticity windows close later than cortical ones (Hypothesis 5), suggesting that subcortical components of auditory rehabilitation may remain partially accessible even after the cortical sensitive period has closed. Rehabilitation strategies for late-implanted children should therefore explicitly target subcortical as well as cortical levels.

Age-graded auditory training

Experience-dependent plasticity provides a neurobiological basis for age-targeted auditory training programs. Training paradigms have shown improvements in brainstem encoding in children with dyslexia [38], neural timing in older adults [20], and frequency tracking in tonal language learning [17]. Musical training is one of the rare interventions that appears to target multiple plasticity windows simultaneously, with effects documented at both subcortical and cortical levels [14,39,40]. Clinical Recommendation 2 (Evidence Level: Moderate | Evidence Type: Human quasi-experimental and correlational studies; limited RCT support; window-specific age assignments partially derived from rodent models): auditory training programs are most likely to be effective when matched to the plasticity windows appropriate for the targeted skill. The following age-based guidelines are suggested on the basis of converging human electrophysiological data and rodent sensitive period mapping: 0–5 years for basic frequency discrimination; 5–12 years for temporal processing and speech-in-noise skills; adulthood and elderly for speech envelope tracking and cognitive-auditory integration. The specific age boundaries are derived partly from rodent data and partly from cross-sectional human studies; direct prospective human evidence for window-specific training efficacy remains limited. These guidelines should be treated as evidence-informed starting points rather than established protocols, and prospective controlled studies in each age group are needed.

Neuroprotective amplification

Central changes in presbyacusis extend beyond peripheral loss. Human epidemiological evidence consistently links untreated hearing loss to accelerated cognitive decline and increased dementia risk [41,42], suggesting that peripheral deafferentation has consequences extending well into the central nervous system. Mechanistic evidence from rodent studies indicates that peripheral deafferentation drives progressive central inhibitory decline and maladaptive gain changes at multiple neuraxis levels [18,19], providing a plausible neurobiological basis for the neuroprotective amplification hypothesis. However, it should be noted that while the association between hearing loss and dementia is robust in human data, the causal role of central maladaptive plasticity in this relationship—and the capacity of hearing aids to interrupt it—has not been established in prospective RCTs. Ongoing trials (e.g., the ACHIEVE study) are expected to provide more definitive evidence. Clinical Recommendation 3 (Evidence Level: Moderate | Evidence Type: Human epidemiological data for the hearing loss-dementia association; rodent mechanistic data for central pathway preservation; no prospective RCT yet available for hearing aid neuroprotection): early and sustained amplification should be considered not only for the purpose of peripheral gain but also to minimize central deafferentation and its downstream consequences. The concept of neuroprotective amplification—fitting hearing aids earlier and more consistently than currently practiced—is supported by the available evidence but awaits confirmation from prospective human studies.

Pharmacological plasticity enhancement

Clinical Recommendation 4 (Evidence Level: Low | Evidence Type: Rodent in vivo and in vitro models only; no human data available): pharmacological strategies targeting the molecular brakes of cortical critical periods—including chemogenetic silencing of PV+ interneurons [10], enzymatic degradation of PNNs [25], and Lynx1 inhibition—have shown proof-of-concept in rodent models. These approaches are entirely preclinical and must not be applied in clinical practice outside of properly approved research protocols. Their inclusion here is intended solely to highlight a promising future research direction. Translation to humans would require extensive safety profiling, dose-response characterization, and human trial phases. The combination of such approaches with intensive auditory rehabilitation, once safety is established, represents a high-priority experimental paradigm.

Table 2 summarizes the key mechanisms, critical period timelines, species models, and clinical relevance for each level of the auditory neuraxis, and provides a concise reference for matching intervention strategies to the appropriate developmental stage and evidence level.

Table 2.

Summary of mechanisms, critical period timelines, and clinical relevance across the auditory neuraxis

Open questions

First priority: molecular timers of subcortical critical periods (PNNs, PV+ maturation, myelin) should be characterized in the IC and CN. Second priority: subcortical-cortical causality should be tested through optogenetic/chemogenetic manipulation of corticofugal projections during developmental windows. Third priority: human-scale plasticity biomarkers should be developed using combinations of cABR, N1/P2, GABA-MRS, and gamma EEG. Fourth priority: The auditory plasticity-cognitive decline relationship should be investigated to understand the mechanistic basis of hearing loss as a dementia risk factor [41,42].

Conclusion

In this scoping review, we mapped and synthesized the available evidence on the mechanisms, developmental timelines, and clinical implications of age-dependent auditory plasticity across subcortical and cortical levels. Four contributions emerged: 1) critical methodological appraisals, including identification of a persistent translational gap between rodent and human data; 2) the GWM—a conceptual framework proposing that plasticity is organized as a series of hierarchically sequenced windows with distinct molecular timers at each neuraxis level; 3) five testable hypotheses to provide empirical grounding for the model’s principles; and 4) clinical recommendations graded by evidence level and type, with human and animal evidence clearly distinguished. Auditory plasticity is not a single window but a series of windows that open and close in a graded manner for different levels and features. If validated, the GWM may have important implications for cochlear implant rehabilitation, auditory training design, and neuroprotective amplification. Rigorously testing the hypotheses it generates is the necessary next step.