Functional Organization of Auditory and Reward Systems in Aging

Abstract

The intrinsic organization of functional brain networks is known to change with age, and is affected by perceptual input and task conditions. Here, we compare functional activity and connectivity during music listening and rest between younger (n = 24) and older (n = 24) adults, using whole-brain regression, seed-based connectivity, and ROI-ROI connectivity analyses. As expected, activity and connectivity of auditory and reward networks scaled with liking during music listening in both groups. Younger adults show higher within-network connectivity of auditory and reward regions as compared with older adults, both at rest and during music listening, but this age-related difference at rest was reduced during music listening, especially in individuals who self-report high musical reward. Furthermore, younger adults showed higher functional connectivity between auditory network and medial prefrontal cortex that was specific to music listening, whereas older adults showed a more globally diffuse pattern of connectivity, including higher connectivity between auditory regions and bilateral lingual and inferior frontal gyri. Finally, connectivity between auditory and reward regions was higher when listening to music selected by the participant. These results highlight the roles of aging and reward sensitivity on auditory and reward networks. Results may inform the design of music-based interventions for older adults and improve our understanding of functional network dynamics of the brain at rest and during a cognitively engaging task.

Figure 1.

Functional activity across liking ratings. Univariate second-level results showing activity at each level of liking (rows 1–4) and linear contrast of liking (row 5) for YAs (A) and OAs (B) across liking ratings (p < .05 p-FDR corrected for voxel height and cluster size). (C) Subcortical structures involved with listening to loved music for YAs (red), OAs (blue), and both (purple; voxel height p < .05 p-FDR corrected; cluster size k > 10).

Figure 2.

Age-related differences in ROI–ROI connectivity for various levels of liking. (A) Correlation matrices showing the relationship between auditory and reward regions for different liking ratings for YAs and OAs. (B) Bar graphs depicting overall connectivity between auditory–auditory, reward–reward, and auditory–reward regions for YAs (orange) and OAs (blue) for hated, neutral, liked, and loved musical stimuli.

Figure 3.

Age differences in auditory network seed-based connectivity: significant clusters for YA > OA contrast in auditory network seed, significant at the p < .05, FDR-corrected level for both voxel height and cluster size. (A) Visualization of significant clusters in the right hemisphere during resting-state scan, including clusters favoring YAs (warm tones) and clusters favoring OAs (cool tones). (B) Bar graphs depicting effect sizes for all significant clusters in resting-state contrast. Clusters with higher connectivity in YAs (red) are depicted in the top graph, and clusters with higher connectivity in OAs (blue) are depicted in the bottom graph. Cluster labels are MNI coordinates for center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster. (C) Visualization of significant clusters in the right hemisphere during the music listening task, including clusters favoring YAs (warm tones) and clusters favoring OAs (cool tones). (D) Bar graphs depicting effect sizes for all significant clusters in music listening contrast. Clusters with higher connectivity in YAs (red) are depicted in the top graph, and clusters with higher connectivity in OAs (blue) are depicted in the bottom graph. (E) Line graph depicting ROI–ROI connectivity between auditory network regions and mPFC comparing across groups and liking ratings. Error bars depict ± 2 SEs, and marginal means are adjusted for covariates of gender, BMRQ, and within-group age.

Figure 4.

Age differences in reward network seed-based connectivity: significant clusters for YA > OA contrast in reward network seed, significant at the p < .05, FDR-corrected level for both voxel height and cluster size. (A) Visualization of significant clusters in right hemisphere during resting-state scan, including clusters favoring YAs (warm tones) and clusters favoring OAs (cool tones). (B) Bar graphs depicting effect sizes for all significant clusters in resting-state contrast. Clusters with higher connectivity in YAs (red) are depicted in the top graph, and clusters with higher connectivity in OAs (blue) are depicted in the bottom graph. Cluster labels are MNI coordinates for center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster. (C) Visualization of significant clusters in the right hemisphere during the music listening task, including clusters favoring YAs (warm tones) and clusters favoring OAs (cool tones). (D) Bar graphs depicting effect sizes for all significant clusters in music listening contrast. Clusters with higher connectivity in YAs (red) are depicted in the top graph, and clusters with higher connectivity in OAs (blue) are depicted in the bottom graph.

Figure 5.

Task differences in seed-based connectivity: significant clusters for music > rest contrast, significant at the p < .05, FDR-corrected level for both voxel height and cluster size. (A) Visualization of significant clusters from auditory network seed in right hemisphere of YAs, including clusters favoring music listening (warm tones) and clusters favoring rest (cool tones). (B) Bar graphs depicting effect sizes for all significant clusters in young adult contrast. Clusters with higher connectivity during music listening (red) are depicted in the top graph, and clusters with higher connectivity during rest (blue) are depicted in the bottom graph. Cluster labels are MNI coordinates for the center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster. (C) Visualization of significant clusters from the auditory network seed in the right hemisphere of OAs, including clusters favoring music listening (warm tones) and clusters favoring rest (cool tones). (D) Bar graphs depicting effect sizes for all significant clusters in OA contrast. Clusters with higher connectivity during music listening (red) are depicted in the top graph, and clusters with higher connectivity during rest (blue) are depicted in the bottom graph. (E) Visualization of significant clusters from the reward network seed in YAs, and bar graphs depicting effect sizes for all significant clusters. Cluster labels are MNI coordinates for the center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster.

Figure 6.

Differences of musical preference in seed-based connectivity: significant clusters for effects of musical selection (self-selected > familiar western) and linear effects of musical liking in auditory and reward network seed-based connectivity ([−3, −1, 1, 3], two-tailed t test; p < .05, FDR-corrected for voxel height and cluster size). (A) Visualization of significant clusters from the reward network seed showing effects of selection, and bar graphs depicting effect sizes. Cluster labels are MNI coordinates for the center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster. (B) Visualization of significant clusters from the auditory network seed showing linear effects of liking, and bar graphs depicting effect sizes. Cluster labels are MNI coordinates for the center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster. (C) Visualization of significant clusters from the reward network seed showing linear effects of liking, and bar graphs depicting effect sizes. Cluster labels are MNI coordinates for the center of mass of each significant cluster, as well as the most prevalent anatomical region included within each cluster.