Superficial white matter (SWM) lies immediately beneath the cortical mantle and consists primarily of short association fibers, called the U-fibers of Meynert (1), which are among the slowest myelinating fibers and the myelination process may extend into the fourth decade of life. Normal development and aging of SWM have seldom been systematically examined and understanding non-disease-related changes in SWM during normal aging can inform pathological changes associated with age-related neurodegenerative disorders. Using magnetization transfer (MT) imaging and an innovative multimodal image analysis approach (VBSS), this study aimed to vertexwise characterize the effects of age on SWM across the entire cortex in healthy adults from 30 to 85 years of age.

Subject. The subject sample consisted of 66 healthy adults (age: 60.22 ± 14.51 years, IQ: 106.80 ± 12.20, education: 16.03 ± 3.00 years, Cumulative Illness Rating Scale: 3.80 ± 2.52, Mini-Mental Status Exam score: 29.20 ± 0.95, Hamilton Depression Rating Scale score: 1.00 ± 1.38, hemoglobin A1c level: 5.62 ± 0.34% (38 ± 3.7 mmol/mol).

MRI Data Acquisition. MRI scans were performed on a Philips Achieva 3T scanner with an 8-element phased-array head coil. MT images were acquired using a 3D spoiled gradient-echo sequence with multi-shot EPI readout: TR/TE=64/15ms, flip angle=9°, FOV=24 cm, 67 axial slices, slice thickness/gap =2.2 mm/no gap, EPI factor=7, reconstructed voxel size = 0.83×0.83×2.2 mm³, with a nonselective five-lobed Sinc-Gauss off-resonance MT prepulse (B1/Δf/dur=10.5μT/1.5kHz/24.5ms). T1-weighted (T1w) magnetization prepared rapid acquisition gradient echo (MPRAGE) image was acquired: TR/TE = 8.4/3.9 ms, flip angle = 8º, FOV = 24 cm, 134 axial slices/no gap, reconstructed voxel size = 0.83 × 0.83 ×1.1 mm³.

Image Processing. For each subject, inter-subject cortical spatial normalization and cortical thickness were estimated on the T1w image using FreeSurfer (2-4). Magnetization transfer ratio (MTR) image was calculated from MT images: MTR= (M₀-M_s)/M₀ (M_s with and M₀ without the MT prepulse) and was co-registered to the T1w image. With the MT->T1w intra-subject transformation, SWM MTR was averaged and projected onto individual WM surface, which was further projected onto the common template surface using the T1w -> template surface transformation. SWM MTR at each vertex was determined by averaging the MTR values of SWM sampled along the WM surface normal from 1 mm to up to 5 mm of distance to the WM surface. Surface-based smoothing with a 10-mm FWHM Gaussian kernel was applied. General linear model analyses were performed to test the effects of gender and age on SWM MTR and cortical thickness at each vertex, with linear and quadratic terms of age (age and age2) as continuous covariates. A cluster-size threshold was estimated with Monte Carlo simulation (10,000 iterations of simulation) and was used to correct for multiple comparisons (5).

The effects of age on SWM MTR were illustrated in Fig. 1 (corrected p < 0.0001). Significant age-related changes of SWM MTR was observed in widespread SWM regions across all brain lobes. In contrast, there is no significant change of SWM MTR with age in the SWM regions associated with basic motor and sensory functions and some tertiary heteromodal regions. To illustrate the heterogeneous patterns of relationship between SWM MTR and age, SWM MTR was plotted by age at 25 SWM vertices in Fig. 2. There is no substantial spatial overlap in the age effects between SWM MTR and cortical thickness.Our findings suggest regionally selective and temporally heterochronologic changes of SWM MTR with age. Distinct age effects on SWM and cortical GM suggest that the age-related changes of SWM MTR found were not due to cortical atrophy but reflected alterations in myelin status in SWM with age. As a sensitive marker of myelin integrity, divergent trajectories of SWM MTR with age, (i.e., SWM MTR peaks or decreases at different ages or rates) suggest that SWM myelin change occurs at varied paces across the cortex: (1) Inverted U-shaped trajectories of SWM MTR with age in the rostral middle frontal, inferior temporal, and temporoparietal regions, which suggests myelination and protracted maturation of SWM till age 40-50 years and accelerating demyelination at age 60 years and beyond, (2) Linear decline of SWM MTR with age in the middle and superior temporal, and pericalcarine areas, which may be interpreted as early maturation and less acceleration in age-related degeneration of SWM, (3) No significant changes in SWM MTR in the primary motor (precentral), somatosensory (postcentral), auditory (Hechl’s gyri) regions, which indicates early maturation and resistance to age-related deterioration of SWM. Regionally varied susceptibility to myelin degradation in SWM during normal aging may relate to different myelination mechanisms and fiber profiles in the regions.