medRxiv · Jan 30, 2026
Extended High-Frequency Hearing Loss, Not Cochlear Synaptopathy, Predicts Speech Recognition in a Population Cohort
By Cederroth CR
Abstract Age-related hearing loss is the leading sensory deficit in older adults, yet audiometric thresholds at conventional frequencies often poorly predict speech understanding. Two competing hypotheses have emerged: extended high-frequency (eHF) hearing loss beyond 8 kHz may unmask variance in speech performance, while hidden hearing loss from cochlear synaptopathy - detectable via the reduction of auditory brainstem response (ABR) wave I amplitude - may degrade higher order temporal processing independent of audiometry. Here, in 526 ears from 263 tinnitus-free adults in the Swedish Tinnitus Outreach Project (STOP) cohort, we show that eHF pure-tone average (10–16 kHz) is the single most age-sensitive auditory measure, explaining 64% of age-related variance compared to only 16% for conventional audiometry (PTA4, 0.5-4 kHz). EHF strongly predicted speech-in-noise performance, substantially more than any other audiological marker (i.e., PTA4, otoacoustic emissions, or ABR). Whereas Wave I amplitude showed no effect, Wave I latency showed robust predictive value in the decline of word recognition. While in young adults, delayed Wave I latency had little impact on word recognition, in older adults, a 1.8 ms delay lead to ∼25% drop in word recognition. These findings challenge the translational relevance of cochlear synaptopathy to age-related speech deficits and identify auditory nerve conduction delays, not synaptic loss, as the peripheral neural mechanism underlying speech comprehension decline in aging. EHF audiometry emerges as the dominant clinical predictor of speech outcomes, with implications for clinical routine and interventional clinical trials in heterogeneous human populations.