Age dependent deficits in speech recognition in quiet and noise are reflected in MGB activity and cochlear onset coding

• Published in: NeuroImage (Orlando, Fla.) Vol. 305; p. 120958
• Main Authors: Dapper, Konrad, Wolpert, Stephan M., Schirmer, Jakob, Fink, Stefan, Gaudrain, Etienne, Başkent, Deniz, Singer, Wibke, Verhulst, Sarah, Braun, Christoph, Dalhoff, Ernst, Rüttiger, Lukas, Munk, Matthias H.J., Knipper, Marlies
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.01.2025; Elsevier Limited; Elsevier
• Subjects: Adolescent; Adult; Age; Aged; Aging; Aging - physiology; Atrophy; Auditory nerve; Brain stem; Cochlea; Cochlea - physiology; Cochlear synaptopathy; Cognitive science; EEG; EEG recording; Electroencephalography; Ethics; Female; Frequency dependence; Hearing aid; Hearing loss; Humans; Information processing; Latency; Localization; Male; Medial geniculate body; Middle Aged; Neural coding; Neuroscience; Noise; Noise reduction; Otolaryngology; Speech; Speech Perception - physiology; Speech recognition; Ultrastructure; Voice recognition; Young Adult; PTA-EHF; EHF; DPOAE; PT; PNOT; PTA-LF; OLSA; ASSR; PTA-HF; ABR; PLL; PTT; ANOVA; Aging; TENV; SR; TFS; CWT; PEA; Hearing aid; PED; PCA; PTA4; Cochlear synaptopathy; PC; pDPOAE; Speech recognition; ERF; PTA; EEG recording; one-way analysis of variance; one-way analysis of variance ASSR; German Oldenburger sentence test; EEG recording Speech recognition Cochlear synaptopathy Aging Hearing aid ABR; distortion-product otoacoustic emission; event-related fields OLSA; auditory brainstem responses ANOVA; auditory steady-state responses CWT; event-related fields; continuous wavelet transform DPOAE; extended high frequencies ERF; continuous wavelet transform; auditory steady-state responses; distortion-product otoacoustic emission EHF; extended high frequencies; auditory brainstem responses
• ISSN: 1053-8119; 1095-9572; 1095-9572
• DOI: 10.1016/j.neuroimage.2024.120958

•Speech cues above and below phase locking evoke different MGB activity•Speech envelope cues in quiet are reflected in likely MGB amplitudes•Extended high-frequency thresholds are reflected in MGB delay related to speech cues.•MGB delay may predict speech recognition in quiet and noise independent of age.•MGB activity can reflect cochlear synaptopathy regardless of age. The slowing and reduction of auditory responses in the brain are recognized side effects of increased pure tone thresholds, impaired speech recognition, and aging. However, it remains controversial whether central slowing is primarily linked to brain processes as atrophy, or is also associated with the slowing of temporal neural processing from the periphery. Here we analyzed electroencephalogram (EEG) responses that most likely reflect medial geniculate body (MGB) responses to passive listening of phonemes in 80 subjects ranging in age from 18 to 76 years, in whom the peripheral auditory responses had been analyzed in detail (Schirmer et al., 2024). We observed that passive listening to vowels and phonemes, specifically designed to rely on either temporal fine structure (TFS) for frequencies below the phase locking limit (<1500 Hz), or on the temporal envelope (TENV) for frequencies above phase locking limit, entrained lower or higher neural EEG responses. While previous views predict speech content, particular in noise to be encoded through TENV, here a decreasing phoneme-induced EEG amplitude over age in response to phonemes relying on TENV coding could also be linked to poorer speech-recognition thresholds in quiet. In addition, increased phoneme-evoked EEG delay could be correlated with elevated extended high-frequency threshold (EHF) for phoneme changes that relied on TFS and TENV coding. This may suggest a role of pure-tone threshold averages (PTA) of EHF for TENV and TFS beyond sound localization that is reflected in likely MGB delays. When speech recognition thresholds were normalized for pure-tone thresholds, however, the EEG amplitudes remained insignificant, and thereby became independent of age. Under these conditions, poor speech recognition in quiet was found together with a delay in EEG response for phonemes that relied on TFS coding, while poor speech recognition in ipsilateral noise was observed as a trend of shortened EEG delays for phonemes that relied on TENV coding. Based on previous analyses performed in these same subjects, elevated thresholds in extended high-frequency regions were linked to cochlear synaptopathy and auditory brainstem delays. Also, independent of hearing loss, poor speech-performing groups in quiet or with ipsilateral noise during TFS or TENV coding could be linked to lower or better outer hair cell performance and delayed or steeper auditory nerve responses at stimulus onset. The amplitude and latency of MGB responses to phonemes requiring TFS or TENV coding, dependent or independent of hearing loss, may thus be a new predictor of poor speech recognition in quiet and ipsilateral noise that links deficits in synchronicity at stimulus onset to neocortical activity. Amplitudes and delays of speech EEG responses to syllables should be reconsidered for future hearing-aid studies. A: Over age hearing loss is over a wide frequency range (PTA-4, PTA-EHF) associated with a decrease in EEG (MGB) response amplitude. This predominantly effects speech recognition on the level of TENV coding for OLSA in quiet (likely through high-spontaneous firing rate (high-SR) low threshold auditory fibers) and ipsilateral noise (likely through low- spontaneous firing rate (low-SR) high threshold auditory fibers). B: Over age hearing loss in extended high frequency range (PTA-EHF) is associated with a prolonged latency (delay) of EEG (MGB) response. This predominantly effects speech recognition on the level of TFS and TENV coding both in quiet and in ipsilateral noise, again possibly linked to synaptopathy of high-SR fibers (TFS, quiet) and low-SR fibres (TENV, noise) coding in the cochlea. C: Independent of age and hearing threshold a delay in EEG (MGB) response is correlated with speech recognition deficits for TFS coding in quiet (likely synaptopathy of high-SR fibers) and a shortening of EEG (MGB) response is correlated with TENV coding in ipsilateral noise (likely low-SR fibers ). [Display omitted]