Age-related hearing loss (presbycusis) at moderate levels (>40 dB HL) has been recognized as an important risk factor for cognitive decline. However, whether individuals with mild hearing loss (audiogram thresholds between 25 and 40 dB HL) or even those with normal audiograms (<25 dB HL) have a higher risk of dementia is still debated. Importantly, these early stages of presbycusis are the most common among the elderly, indicating the need to screen and identify individuals with early presbycusis that have an increased risk of cognitive decline. Unfortunately, in this group of patients, audiogram thresholds are not sufficiently sensitive to detect all the hearing impairments related to cognitive decline. Consequently, at the individual level, audiogram thresholds are not good estimators of the dementia risk in the group with mild hearing loss or normal hearing thresholds. Here, we propose to use distortion product otoacoustic emissions (DPOAE) as an objective and sensitive tool to estimate the risk of clinically relevant cognitive decline in elders with normal hearing o mild hearing loss. We assessed neuropsychological, brain magnetic resonance imaging, and auditory analyses on 94 subjects aged >64 years old. In addition, cognitive and functional performance was evaluated with the Clinical Dementia Rating Sum of Boxes (CDR SoB), assessed through structured interviews conducted by neurologists blind to the DPOAE results. We found that cochlear dysfunction, measured by DPOAE -and not by conventional audiometry-was associated with CDR SoB classification and brain atrophy in the group with mild hearing loss (25 to 40 dB) and normal hearing (<25 dB). Our findings suggest that DPOAE may be a non-invasive tool for detecting neurodegeneration and cognitive decline in the elderly, potentially allowing for early intervention.