Aging Affects Neural Synchronization to Speech-Related Acoustic Modulations

Abstract

As people age, speech perception problems become highly prevalent, especially in noisy situations. In addition to peripheral hearing and cognition, temporal processing plays a key role in speech perception. Temporal processing of speech features is mediated by synchronized activity of neural oscillations in the central auditory system. Previous studies indicate that both the degree and hemispheric lateralization of synchronized neural activity relate to speech perception performance. Based on these results, we hypothesize that impaired speech perception in older persons may, in part, originate from deviances in neural synchronization. In this study, auditory steady-state responses that reflect synchronized activity of theta, beta, low and high gamma oscillations (i.e., 4, 20, 40, and 80 Hz ASSR, respectively) were recorded in young, middle-aged, and older persons. As all participants had normal audiometric thresholds and were screened for (mild) cognitive impairment, differences in synchronized neural activity across the three age groups were likely to be attributed to age. Our data yield novel findings regarding theta and high gamma oscillations in the aging auditory system. At an older age, synchronized activity of theta oscillations is increased, whereas high gamma synchronization is decreased. In contrast to young persons who exhibit a right hemispheric dominance for processing of high gamma range modulations, older adults show a symmetrical processing pattern. These age-related changes in neural synchronization may very well underlie the speech perception problems in aging persons.

Keywords: aging; amplitude modulations; auditory steady-state response (ASSR); hemispheric lateralization; neural oscillations; synchronization; temporal processing.

Figure 1

Audiometric thresholds. Median audiometric thresholds (dB HL) of the young, middle-aged, and older participants are represented by the black, gray, and dashed lines, respectively. Error bars indicate the interquartile range.

Figure 2

Electrode configuration and selection. 64 active electrodes placed according to the 10-10 electrode system and two complementary electrodes, CMS and DRL. The electrodes that were selected for further analyses, are encircled in black.

Figure 3

Whole-head voltage topographic maps. The EEG voltage maps are plotted with reference-free interpolation and show the scalp distribution of the instantaneous voltage at the maximum positive peak of the grand mean averaged response period. The 4, 20, 40, and 80 Hz grand mean response period is obtained by averaging 256 artifact-free epochs from all participants (N = 53) for the three sides of stimulation (L, R, BI). The panels display the EEG voltage topography for the 4, 20, 40, and 80 Hz modulation frequencies (from left to right panels, respectively) from six different viewpoints. Red buttons represent the electrode locations. The voltage distributions are plotted on a color shade minimum-maximum scale going from blue (negative polarity) to red (positive polarity), in which a darker tint reflects a higher voltage value.

Figure 4

FFT spectra. EEG frequency spectra of a young participant, obtained by the procedure outlined in Section EEG Processing. The spectra represent the neural responses to the 4, 20, 40, and 80 Hz acoustic modulations that were recorded over the electrode selection in the right hemisphere (see Figure 2), when presenting the stimuli to the right ear. Spectral amplitudes are converted to dB by applying a logarithmic transformation with reference to 1 nV. The arrows indicate the 4, 20, 40, and 80 Hz ASSR, i.e., the spectral peak in the respective modulation frequency bin that stands out above the adjacent frequency bins which contain EEG noise, induced by the auditory stimulus.

Figure 5

Auditory evoked potentials topographic maps. The evoked potentials are referenced to the vertex electrode (Cz) and represent the time-domain grand mean averaged response to four cycles of the 4, 20, 40, or 80 Hz modulated noise. The 4, 20, 40, and 80 Hz grand mean response period is obtained by averaging 256 artifact-free epochs from all participants (N = 53) for the three sides of stimulation (L, R, BI). The waveforms are plotted at the approximate electrode locations viewed from the top of the head with the nose at the top of the figure.

Figure 6

Neural SNR. Overview of the neural SNRs (dB) in response to the 4, 20, 40, and 80 Hz acoustic modulations. Age groups (young, middle-aged, older) are displayed per column. The bars are clustered per hemisphere (LH, RH) and represent the average neural SNR per side of stimulation, i.e., L (dark fill), R (dotted fill), and BI (gray fill). Error bars indicate the 95% CI. ^*p ≤ 0.05; ^***p ≤ 0.001.

Figure 7

Laterality index. Overview of the LIs for the 4, 20, 40, and 80 Hz acoustic modulations. The markers are clustered per age group (young, middle-aged, older) and represent the average LI per side of stimulation, i.e., L (triangles), R (circles), and BI (squares). Error bars indicate the 95% CI. ^*p ≤ 0.05; ^**p ≤ 0.01.