Cognitive flexibility is a fundamental component of adaptive human behavior, enabling individuals to shift perspectives, modify strategies, and coordinate a complex array of executive functions. This capacity emerges early in life and continues to develop throughout adulthood, supporting essential cognitive processes such as decision-making, inhibitory control, and working memory.
Cognitive flexibility is often impaired in both neurologic and neuropsychiatric disorders, such as anxiety and dementia, as well as during normal aging. Maintaining cognitive flexibility in late adulthood can reduce age-related cognitive impairment. Therefore, determining the neural processes that support flexibility across adulthood is critically important.
Cognitive flexibility depends on the coordinated activity of distributed brain systems and is supported by both regional specialization and large-scale network integration. The structural and functional changes associated with aging include synaptic pruning, cortical myelination, and reorganization of white matter pathways. Although many neuroimaging studies characterize age-related white matter changes as degenerative, histological evidence suggests that not all such changes are pathological; some may reflect adaptive differentiation. Accordingly, this study examines the relationship between macromolecular properties of white matter, cognitive flexibility, and age during natural aging rather than disease.
A Neurosynth meta-analysis was conducted to identify brain regions associated with set-shifting using the search term switching. The analysis was performed based on 14,371 studies and extracted 5,124 significant neural activations with a specificity value of 0.885 using an effect size threshold of >2. These regions were mapped onto four anatomical atlases to outline cortical and subcortical areas involved in cognitive flexibility.
White matter tracts connecting these regions were reconstructed using the Human Connectome Project (HCP) 1065 tractography template (1,065 subjects, ages 22-37 years) and compared to the HCP-Aging template (422 subjects, ages 39-100+ years). The parameters used in tractography were 10,000,000 streamlines, a quantitative anisotropy (QA) threshold of 0.1, and the tract lengths ranging from 5 to 400 mm.
A subset of UK Biobank participants (n = 301) was included for age-related analyses, comprising younger adults (mean age: 22 ± 3 years; 43% male, 57% female) and older adults (mean age: 68 ± 6 years; 45% male, 54.2% female). Cognitive flexibility was measured with the help of NIH Toolbox tasks that measured dimensional change card sorting, flanker inhibitory control, and list sorting working memory. White matter characteristics obtained from magnetic resonance imaging (MRI) included the mean standardized T1-weighted intensity (m), reflecting macromolecular density, and kurtosis (k) of the T1/ T2-FLAIR ratio, which reflects myelin-related homogeneity.
Statistical analyses included independent t-tests, linear regression, and Bayesian-Pearson correlation. The cognitive flexibility white matter structure identified consisted of 32 tracts that included the major long-range pathways, such as the cingulum bundle, arcuate fasciculus, uncinate fasciculus, and superior longitudinal fasciculus. Comparisons between the HCP1065 and HCP-Aging templates revealed only slight reorganizations, accounting for approximately 0.3% of the tracts. Performance differed significantly between age groups across all cognitive tasks (p < 0.001). Set-shifting and working memory declined with age in older adults (β₁ = −0.462, p = 0.001; β₁ = −0.339, p = 0.014), whereas inhibitory control showed age-related decline primarily in older adults, with additional effects observed in younger adults (β₁ = 0.892, p = 0.034).
White matter analyses revealed significant age-group differences in macromolecular density and homogeneity within the aggregated tract system (m: t[30] = 9.26, p < 0.001; k: t[27] = 3.24, p = 0.003), with both measures reduced in older adults. Homogeneity decreased with increasing age among the older adults (r = -0.643, p < 0.001) and showed associations with biological sex.
Associations of age between functions and structure were age-dependent. In younger adults, cognitive performance showed weaker associations with white matter properties, with inhibitory control demonstrating a modest negative association with the right uncinate fasciculus (r = −0.202, p = 0.043). In contrast, older adults exhibited multiple tract-specific associations linking white matter density and homogeneity to set-shifting, inhibitory control, and working memory.
In summary, this study demonstrates that cognitive flexibility is supported by a specialized white matter system that undergoes substantial differentiation across adulthood. In early adulthood, these tracts appear relatively undifferentiated, whereas later stages of life are characterized by selective declines in macromolecular density and homogeneity, reflecting myelin-related changes that are associated with functional performance. Notably, these alterations are indicative of normal aging rather than pathology. By integrating meta-analytic functional mapping with macromolecular MRI measures, this work can be characterized by the value of a function-centered approach to understanding brain aging. The findings suggest that age-related white matter differentiation plays a critical role in cognitive flexibility and may provide a foundation for future normative models of healthy brain aging.
References: Wolfe T, Gassel A, Calvert ML, et al. Population-level age effects on the white matter structure subserving cognitive flexibility in the human brain. eNeuro. 2026;ENEURO.0179-25.2025. doi:10.1523/ENEURO.0179-25.2025
