BOLD fMRI responses to amplitude-modulated sounds across age in adult listeners

Abstract

Age-related alterations in the auditory system have been suggested to affect the processing of temporal envelope amplitude modulations (AM) at different levels of the auditory hierarchy, yet few studies have used functional magnetic resonance imaging (fMRI) to study this noninvasively in humans with high spatial resolution. In this study, we utilized sparse-sampling fMRI at 3 Tesla (3T) to investigate regional blood oxygenation level-dependent (BOLD) responses to AM noise stimuli in 65 individuals ranging in age from 19 to 77 years. We contrasted BOLD responses to AM noise stimuli modulated at 4 Hz or 80 Hz with responses to unmodulated stimuli. This allowed us to derive functional measures of regional neural sensitivity to the imposed AM. Compared with unmodulated noise, slowly varying 4 Hz AM noise stimuli elicited significantly greater BOLD responses in the left and right auditory cortex along the Heschl's gyrus (HG). BOLD responses to the 80 Hz AM stimuli were significantly greater than responses to unmodulated stimuli in putatively primary auditory cortical regions in the lateral HG. BOLD responses to 4 Hz AM stimuli were significantly greater in magnitude than responses to 80 Hz AM stimuli in auditory cortical regions. We find no discernible effects of age on the functional recruitment of the auditory cortex by AM stimuli. While the results affirm the involvement of the auditory cortex in processing temporal envelope rate information, they provide no support for age-related effects on these measures. We discuss potential caveats in assessing age-related changes in responses to AM stimuli in the auditory pathway.

Keywords: age-related effects; amplitude modulation; auditory; envelope; fMRI.

Fig. 1.

(a) Schematic representation of an experimental run for a given participant. Each functional run consisted of stimulus trials, silent trials, and catch trials. The timing of acquisitions is visualized with light gray vertical bars. The colors are used to indicate different stimulus types. Bars of lower height indicate catch trials. (b) Group-mean responses in auditory regions of interests (ROIs) for the three stimulus types. Each panel reflects average percent signal change (PSC) in a given ROI. The top row shows data from ROIs in the right hemisphere and the bottom row shows data from ROIs in the left hemisphere. Asterisks below each panel are used to indicate significant differences between responses to different stimulus types (purple: 4 Hz vs. 80 Hz, orange: 80 Hz vs. unmodulated, red: 4 Hz vs. unmodulated). Regions of interest include the primary auditory cortex (A1), lateral belt complex (LBelt), medial belt complex (MBelt), retroinsular area (RI), auditory 4 complex (A4), parabelt complex (PBelt), medial geniculate body (MGB), inferior colliculus (IC), superior olivary complex (SOC), and cochlear nucleus (CN). (c) Schematic representation of stimulus waveforms for three realizations of different noise types. Stimuli in stimulus trials were either unmodulated or modulated at 4 Hz or 80 Hz. The different stimulus types are depicted in different colors. (d–g) Whole-brain second-level results. Each panel shows thresholded maps oft-statistics (based on maps of family-wise error rate (FWER) correctedp-values).

Fig. 2.

(a) Visualization of responses to the different noise stimuli in the left and right A1, as well as in the left and right superior olivary complex (SOC), plotted against age of the participants. Each panel depicts regression lines derived from linear regression models that each includes an intercept term and age of the participants. Shaded areas represent 95% confidence intervals of the regression lines obtained via bootstrapping (10000 iterations). Vertical lines indicate standard deviations after dividing participants into two age groups. The “younger” group comprises data from 33 participants (16 females) aged 40 years or younger, while the “older” group comprises data from 32 participants aged over 40 years (17 females). Notice that this figure divides listeners into two groups for visualization purposes, but that the second-level models included age as a continuous variable. (b) Audiograms for each ear. Lines represent data from the “younger” (light blue) and the “older” (light gray) groups, with error bars representing standard deviation. Error bars have been adjusted around the center frequency for improved visualization clarity.