https://orcid.org/0000-0001-9244-9267
Abstract
Poor motor function is associated with brain atrophy and cognitive impairment. Less is known about the relationship between motor domains and brain atrophy and whether associations are affected by cerebrovascular burden and/or physical activity.
We analyzed data from 726 Baltimore Longitudinal Study of Aging participants (mean age 70.6 ± 10.1 years, 56% women, 27% Black), 525 of whom had repeat MRI scans over an average of 5.0 ± 2.1 years. Two motor domains, manual dexterity and gross motor, were operationalized as latent variables. Associations between the latent variables and cortical and subcortical brain volumes of interest were examined using latent growth curve modeling, adjusted for demographics, white matter hyperintensities, and physical activity.
Both higher manual dexterity and gross motor function were cross-sectionally associated with smaller ventricular volume and greater white matter volumes in the frontal, parietal, and temporal lobes (all p < .05). Manual dexterity was also cross-sectionally associated with parietal gray matter (B = 0.14; 95% CI: 0.05, 0.23), hippocampus (B = 0.10; 95% CI: 0.01, 0.20), postcentral gyrus (B = 0.11; 95% CI: 0.01, 0.20), and occipital white matter (B = 0.10; 95% CI: 0.01, 0.21) volumes, and gross motor function with temporal gray matter volume (B = 0.16; 95% CI: 0.05, 0.26). Longitudinally, both higher manual dexterity and gross motor function were associated with less temporal white matter and occipital gray matter atrophy (all p < .05). Manual dexterity was also associated with a slower rate of ventricular enlargement (B = −0.17; 95% CI: −0.29, −0.05) and less atrophy of occipital white matter (B = 0.39; 95% CI: 0.04, 0.71).
Among cognitively normal middle- and older-aged adults, manual dexterity and gross motor function exhibited shared as well as distinct associations with brain atrophy over time.
Age-related decline in motor function affects the performance of daily activities and has been linked to emerging cognitive impairment and reduced life expectancy (1,2). Motor function deficits include (but are not limited to) compromised manual dexterity, low musculoskeletal strength, slow gait speed, high movement variability, and poor balance (3–6). With aging, brain gray and white matter volumes decline (7,8) and these structural changes are thought to underlie motor function deficits (9,10). A growing body of literature has identified a connection between brain structure and motor function in older adults (11–13). For instance, research from the Baltimore Longitudinal Study of Aging (BLSA) has shown parameters of gait and muscular strength to be associated with overall and region-specific gray and white matter brain volumes among cognitively normal older adults (14–16), suggesting that motor function may be an important indicator of brain health (17,18).
Most previous studies of the interplay between motor function and brain structure have primarily focused on gait speed, a single yet highly integrated motor function measure. The current work aims to evaluate a broader range of motor performance activities to possibly improve and/or increase the sensitivity to detect associations between motor parameters and brain structure. Additionally, past research focused primarily on brain gray matter volume (9,12,13,19) and white matter hyperintensities (20,21), which largely represent damage attributable to cerebral small vessel disease (22). This form of vascular brain injury occurs with increasing frequency in older age and is linked to reduced brain volume, and may affect associations between motor function and brain volume (23,24). Indeed, pathological findings in regions of white matter hyperintensities include loss of myelin and axons (25), which may disrupt corticospinal pathways necessary for neuromuscular signaling and recruitment, resulting in disrupted motor function. A better understanding of whether associations between motor function and white matter volume reflect or are independent of vascular damage is needed. As physical activity has been found to have beneficial effects on motor function as well as brain structure (26–28), we also evaluated these relationships as a function of physical activity.
In this study, we aim to identify neural correlates of different motor function domains in aging. Specifically, we examined the cross-sectional and longitudinal associations between different motor function domains and gray and white matter brain volume as well as subsequent brain atrophy over up to 8 years of follow-up while accounting for cerebrovascular burden and physical activity.
Method
Participants
The BLSA is a study of human aging initiated in 1958. All BLSA participants are community-dwelling adults free of major chronic conditions and cognitive and functional impairments at the time of enrollment. Once enrolled, participants are followed for life and undergo comprehensive health, cognitive, neuroimaging, and functional assessments every 1–4 years depending on age (<60: every 4 years, 60–79: every 2 years, ≥80: every year). Trained staff administer all assessments following standardized protocols. Additional study enrollment and design details have been previously described (29). The sample for the current study consists of cognitively normal participants free of Parkinson’s or history of stroke who underwent physical examinations, health history assessments, functional testing, and magnetic resonance imaging (MRI) scans. Participants classified as cognitively impaired at baseline or at any subsequent visit were excluded from the present study. The Internal Review Board of the Intramural Research Program of the National Institutes of Health approved the study protocol. At each examination, participants provided written informed consent.
Motor Function Assessments
Motor function tests were selected based on recent work identifying these measures as predictive of mobility disability in older adults (30). The Purdue Pegboard test (31) assesses visuomotor integration and manual dexterity, which reflects coordinated movement in the upper extremities and psychomotor speed. Participants pick up pegs and place them sequentially into small holes on the board as quickly as possible in 30 seconds over 2 trials per hand. The number of pegs correctly placed was averaged separately for dominant and nondominant hands. Grip strength was measured in kilograms (kg) of force using a Jamar hydraulic hand dynamometer, using the highest value over 3 trials in each hand. The Expanded Short Physical Performance Battery (ExSPPB) (32), an expanded version of the original SPPB test, was used to measure physical functioning over a broader range. Participants were instructed to stand up from a straight-backed chair with folded arms across the chest and sit down 10 times as quickly as possible. The rate of repeated chair stands (chair stands/second) was calculated for the fastest set of 5 chair stands. The standing balance test consists of 3 progressively more difficult standing tests: semitandem, full-tandem, and single-leg stance, each held for 30 seconds. Time to hold each pose was recorded. Binary variables for full-tandem and single-leg stance were also computed (able to hold for 30 seconds vs not). Usual gait speed was measured over a 6-m course in an uncarpeted corridor where participants were instructed to walk at a “normal comfortable pace.” Two timed trials were conducted to derive usual gait speed in m/s; the faster of the 2 trials was used for analyses. The Long-Distance Corridor Walk test (32) measured fast-paced walking endurance. Participants walked back and forth along a course marked by orange traffic cones for ten 40-m laps (400-m total). Total time to complete as well as time to complete each lap was recorded. Lap time variation was defined as the detrended standard deviation of residuals of lap time over the 10 laps. As previously described (33), linear mixed effects models regressed lap time on number of laps, which was then used to calculate the standard deviation of the residuals from the model, which served as the lap time variation measure. All the motor function assessments were completed at one study visit.
Neuroimaging Protocol
The MRI scans were acquired on a 3T Philips Achieva scanner (Philips Healthcare, Andover, MA). Three-dimensional (3-D) T1-weighted magnetization-prepared rapid gradient echo scans, T2-weighted dual-echo scans, and fluid-attenuated inversion recovery (FLAIR) scans were collected. T1-weighted scans used the following parameters: echo time (TE) = 3.2 milliseconds, repetition time (TR) = 6.8 milliseconds, flip angle = 8°, image matrix = 256 × 256, 170 slices, pixel size 1 × 1 mm, slice thickness = 1.2 mm, sagittal acquisition. The structural images were processed using an automated multiatlas approach. After correction of intensity inhomogeneities (34) a multiatlas skull stripping algorithm was applied for the removal of extra-cranial tissues (35). Each T1-weighted scan was automatically segmented into a set of anatomical gray and white matter regions of interest (ROIs) using a multiatlas label fusion method, MUSE (36). This segmentation method has been extensively validated in the BLSA MRI data set (36,37). Intracranial volume was the sum of gray matter, white matter, and cerebral spinal fluid volumes. For our analyses, a priori brain ROIs included gray and white matter volume within the frontal, parietal, temporal, and occipital lobes, and the middle frontal gyrus, precentral gyrus, postcentral gyrus, hippocampus, and striatum. We also investigated ventricular volume as a measure of central brain atrophy. All brain ROIs were expressed as a percentage of intracranial volume to account for individual variability in head size. White matter hyperintensities were quantified using FLAIR and T1-weighted images based on a deep learning-based method (38). The baseline MRI scans were conducted during the same study visit as the motor function assessments.
Covariate Measures
All participants complete a comprehensive set of health-related questionnaires and measurements at each study visit. Baseline variables investigated as potential confounders included age, sex, race, education (years), body mass index (BMI), physical activity, and white matter hyperintensities. Age, sex, race, and education were determined by self-report. Body mass index was calculated from measured height and weight (kg/m2). White matter hyperintensities were derived from MRI. Physical activity was evaluated using a modified leisure-time physical activity questionnaire (39). Participants reported the frequency and duration of specific activities (eg, household chores) over the past 7 days. Based on the metabolic equivalent of each activity, the physical activity measure was expressed as a kcal/week score and participants were categorized into inactive, low, moderate, or high active groups.
Statistical Analysis
We tested the associations between motor function domains and brain volume changes over time using structural equation modeling (SEM). SEM methods facilitate the representation of multiple motor functions as latent variables, based upon the shared covariation among indicators that form a measurement model for that given construct. Exploratory factor analysis was conducted and 2 latent factors corresponding to manual dexterity and gross motor function were identified. Measurement models were then estimated for manual dexterity and gross motor function using confirmatory factor analysis. Manual dexterity was measured by 2 indicators from the Purdue Pegboard test corresponding to the number of pegs correctly placed within 30 seconds using the dominant and nondominant hand. Gross motor function was measured by the following 6 indicators: (i) hand grip strength, (ii) chair stand pace (ExSPPB), (iii) ability to hold full-tandem, and (iv) single-leg stance for 30 seconds (standing balance test), (v) usual gait speed, and (vi) log-transformed lap time variation (long-distance corridor walk test), with the sign changed to be consistent with the polarity of other motor function variables. Confirmatory factor analyses were conducted using maximum likelihood estimation for manual dexterity and weighted least squares estimation with theta parameterization for gross motor function. Model fit was evaluated by global model fit indices including the comparative fit index (CFI), Tucker-Lewis fit index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). Criteria for excellent model fit include CFI of 0.95 or greater, TLI of 0.95 or greater, RMSEA of 0.05 or lower, and SRMR of 0.08 or lower (40).
Baseline and follow-up participant characteristics were summarized using means (standard deviations) and frequencies (percentages). Baseline and longitudinal changes in brain volume were assessed for each brain ROI using latent growth curve models. Latent intercept and slope factors represent baseline and annual linear trajectories of brain volume for each specific region. These latent factors of each brain ROI were regressed on latent manual dexterity and gross motor function in separate models. Coefficients for intercept outcomes represent differences in standardized baseline brain volume per standard deviation (SD) of difference in baseline manual dexterity and gross motor factor scores. Coefficients for trajectories represent annual change of standardized brain volume per SD difference in the manual dexterity and gross motor factor scores. Factor loadings on constructs were constrained to be equal across time points, and residual variances of the latent constructs were constrained to be equal. All models were adjusted for age, sex, race, education, BMI, white matter hyperintensities, and physical activity. All continuous covariates were mean-centered prior to model fitting. The factor loadings were fixed according to the fitted measurement models. The models were estimated with a Bayesian estimator with potential scale reduction values approximately equal to 1.0. Sensitivity analyses tested age (median split) interactions. Statistical significance was set at an α level of .05. Descriptive statistical analyses were conducted using R, version 3.6.2 (R Foundation). The SEM analysis was performed using Mplus 8.4 (Muthén & Muthén, Los Angeles, CA).
Results
We examined 726 participants (mean age 70.6 ± 10.1 years, 56% women, 27% Black) with baseline motor function assessments and structural brain MRI data and 525 participants with follow-up MRI (3.4 ± 1.3 MRI scans over 5.0 ± 2.1 years). A flowchart of participant selection and baseline participant characteristics are detailed in Figure 1 and Table 1. All manual dexterity and gross motor measurement models exhibited excellent model fit (RMSEA ≤ 0.06, CFI ≥ 0.96, TLI ≥ 0.94, and SRMR ≤ 0.05) and factor loadings for indicators in the 2 measurement models ranged from 0.36 to 0.88 (see Figure 2 for SEM diagrams).
Characteristics of Study Participants
| Baseline Variables | Entire Sample | Sample With ≥ 2 Visits |
|---|---|---|
| (N = 726) | (N = 525) | |
| Age, y, mean (SD) [range] | 70.6 (10.1) [50.2, 95.1] | 71.67 (9.7) [50.2, 95.1] |
| Female, N (%) | 403 (55.5) | 296 (56.4) |
| Black, N (%) | 193 (26.6) | 141 (26.9) |
| Education, y, mean (SD) [range] | 17.6 (2.6) [9.0, 30.0] | 17.6 (2.5) |
| BMI, kg/m2, mean (SD) [range] | 27.2 (4.5) [17.8, 45.7] | 27.1 (4.4) [17.8, 45.7] |
| Physical activity, N (%) | ||
| Inactive | 57 (7.9) | 41 (7.8) |
| Low | 270 (37.2) | 194 (37.0) |
| Moderate | 230 (31.7) | 176 (33.5) |
| Very active | 169 (23.3) | 114 (21.7) |
| Cardiovascular disease, N (%) | 56 (7.7) | 43 (8.2) |
| Diabetes, N (%) | 119 (16.4) | 84 (16.0) |
| Hypertension, N (%) | 328 (45.2) | 244 (46.5) |
| Hyperlipidemia, N (%) | 431 (59.4) | 318 (60.6) |
| Osteoarthritis, N (%) | 383 (52.8) | 278 (53.0) |
| Lung disease, N (%) | 93 (12.8) | 69 (13.1) |
| Cancer, N (%) | 209 (28.8) | 153 (29.1) |
| Manual dexterity, mean (SD) [range] | ||
| Pegboard dominant | 12.4 (2.1) [6.0, 18.0] | 12.3 (2.0) [6.0, 17.5] |
| Pegboard nondominant | 11.9 (2.0) [5.0, 18.0] | 11.9 (2.0) [5.0, 18.0] |
| Gross motor function, mean (SD) [range] | ||
| Hand grip strength, kg | 32.2 (10.9) [12.0, 72.0] | 31.6 (10.4) [12.0, 64.0] |
| Chair stands, stands/s | 0.50 (0.17) [0.0, 11.4] | 0.50 (0.16) [0.0, 11.4] |
| Standing balance: full-tandem, N (%) | 646 (89.8) | 462 (89.0) |
| Standing balance: single-leg, N (%) | 413 (57.4) | 289 (55.7) |
| Usual gait speed, m/s | 1.18 (0.23) [0.42, 1.94] | 1.18 (0.23) [0.42, 1.94] |
| Lap time variation | 0.81 (0.38) [0.19, 3.64] | 0.81 (0.39) [0.19, 3.64] |
| Neuroimaging data, cm3, mean (SD) [range] | ||
| White matter hyperintensities | 6.7 (9.5) [0.1, 80.3] | 6.8 (9.6) [0.2, 80.3] |
| Ventricular volume | 33.3 (19.7) [7.6, 131.9] | 33.8 (20.0) [7.6, 131.8] |
| Frontal white matter | 172.3 (21.9) [99.6, 250.4] | 171.8 (21.3) [100.1, 250.4] |
| Parietal white matter | 86.0 (10.6) [55.8, 121.7] | 85.8 (10.4) [55.8, 121.2] |
| Temporal white matter | 102.4 (12.2) [65.7, 142.4] | 102.1 (12.2) [65.7, 142.4] |
| Occipital white matter | 42.4 (5.9) [24.6, 62.3] | 42.3 (5.9) [24.6, 61.5] |
| Frontal gray matter | 179.3 (19.9) [112.7, 253.5] | 178.8 (20.3) [112.7, 253.5] |
| Parietal gray matter | 87.2 (9.6) [62.5, 116.3] | 86.9 (9.5) [62.5, 115.8] |
| Temporal gray matter | 99.6 (11.6) [62.9, 141.0] | 99.3 (11.7) [62.9, 141.0] |
| Occipital gray matter | 72.0 (8.8) [45.8, 98.4] | 71.8 (8.8) [47.8, 98.4] |
| Hippocampus | 7.5 (0.8) [4.8, 10.0] | 7.5 (0.8) [4.8, 10.0] |
| Middle frontal gyrus | 30.6 (4.1) [18.3, 46.8] | 30.6 (4.2) [18.3, 46.8] |
| Precentral gyrus | 22.8 (2.8) [13.9, 30.6] | 22.8 (2.8) [13.9, 30.6] |
| Postcentral gyrus | 17.7 (2.2) [12.1, 26.6] | 17.8 (2.2) [12.6, 24.7] |
| Striatum | 14.5 (1.7) [9.9, 22.5] | 14.5 (1.8) [9.9, 22.5] |
| Baseline Variables | Entire Sample | Sample With ≥ 2 Visits |
|---|---|---|
| (N = 726) | (N = 525) | |
| Age, y, mean (SD) [range] | 70.6 (10.1) [50.2, 95.1] | 71.67 (9.7) [50.2, 95.1] |
| Female, N (%) | 403 (55.5) | 296 (56.4) |
| Black, N (%) | 193 (26.6) | 141 (26.9) |
| Education, y, mean (SD) [range] | 17.6 (2.6) [9.0, 30.0] | 17.6 (2.5) |
| BMI, kg/m2, mean (SD) [range] | 27.2 (4.5) [17.8, 45.7] | 27.1 (4.4) [17.8, 45.7] |
| Physical activity, N (%) | ||
| Inactive | 57 (7.9) | 41 (7.8) |
| Low | 270 (37.2) | 194 (37.0) |
| Moderate | 230 (31.7) | 176 (33.5) |
| Very active | 169 (23.3) | 114 (21.7) |
| Cardiovascular disease, N (%) | 56 (7.7) | 43 (8.2) |
| Diabetes, N (%) | 119 (16.4) | 84 (16.0) |
| Hypertension, N (%) | 328 (45.2) | 244 (46.5) |
| Hyperlipidemia, N (%) | 431 (59.4) | 318 (60.6) |
| Osteoarthritis, N (%) | 383 (52.8) | 278 (53.0) |
| Lung disease, N (%) | 93 (12.8) | 69 (13.1) |
| Cancer, N (%) | 209 (28.8) | 153 (29.1) |
| Manual dexterity, mean (SD) [range] | ||
| Pegboard dominant | 12.4 (2.1) [6.0, 18.0] | 12.3 (2.0) [6.0, 17.5] |
| Pegboard nondominant | 11.9 (2.0) [5.0, 18.0] | 11.9 (2.0) [5.0, 18.0] |
| Gross motor function, mean (SD) [range] | ||
| Hand grip strength, kg | 32.2 (10.9) [12.0, 72.0] | 31.6 (10.4) [12.0, 64.0] |
| Chair stands, stands/s | 0.50 (0.17) [0.0, 11.4] | 0.50 (0.16) [0.0, 11.4] |
| Standing balance: full-tandem, N (%) | 646 (89.8) | 462 (89.0) |
| Standing balance: single-leg, N (%) | 413 (57.4) | 289 (55.7) |
| Usual gait speed, m/s | 1.18 (0.23) [0.42, 1.94] | 1.18 (0.23) [0.42, 1.94] |
| Lap time variation | 0.81 (0.38) [0.19, 3.64] | 0.81 (0.39) [0.19, 3.64] |
| Neuroimaging data, cm3, mean (SD) [range] | ||
| White matter hyperintensities | 6.7 (9.5) [0.1, 80.3] | 6.8 (9.6) [0.2, 80.3] |
| Ventricular volume | 33.3 (19.7) [7.6, 131.9] | 33.8 (20.0) [7.6, 131.8] |
| Frontal white matter | 172.3 (21.9) [99.6, 250.4] | 171.8 (21.3) [100.1, 250.4] |
| Parietal white matter | 86.0 (10.6) [55.8, 121.7] | 85.8 (10.4) [55.8, 121.2] |
| Temporal white matter | 102.4 (12.2) [65.7, 142.4] | 102.1 (12.2) [65.7, 142.4] |
| Occipital white matter | 42.4 (5.9) [24.6, 62.3] | 42.3 (5.9) [24.6, 61.5] |
| Frontal gray matter | 179.3 (19.9) [112.7, 253.5] | 178.8 (20.3) [112.7, 253.5] |
| Parietal gray matter | 87.2 (9.6) [62.5, 116.3] | 86.9 (9.5) [62.5, 115.8] |
| Temporal gray matter | 99.6 (11.6) [62.9, 141.0] | 99.3 (11.7) [62.9, 141.0] |
| Occipital gray matter | 72.0 (8.8) [45.8, 98.4] | 71.8 (8.8) [47.8, 98.4] |
| Hippocampus | 7.5 (0.8) [4.8, 10.0] | 7.5 (0.8) [4.8, 10.0] |
| Middle frontal gyrus | 30.6 (4.1) [18.3, 46.8] | 30.6 (4.2) [18.3, 46.8] |
| Precentral gyrus | 22.8 (2.8) [13.9, 30.6] | 22.8 (2.8) [13.9, 30.6] |
| Postcentral gyrus | 17.7 (2.2) [12.1, 26.6] | 17.8 (2.2) [12.6, 24.7] |
| Striatum | 14.5 (1.7) [9.9, 22.5] | 14.5 (1.8) [9.9, 22.5] |
Note: BMI = body mass index; SD = standard deviation.
Characteristics of Study Participants
| Baseline Variables | Entire Sample | Sample With ≥ 2 Visits |
|---|---|---|
| (N = 726) | (N = 525) | |
| Age, y, mean (SD) [range] | 70.6 (10.1) [50.2, 95.1] | 71.67 (9.7) [50.2, 95.1] |
| Female, N (%) | 403 (55.5) | 296 (56.4) |
| Black, N (%) | 193 (26.6) | 141 (26.9) |
| Education, y, mean (SD) [range] | 17.6 (2.6) [9.0, 30.0] | 17.6 (2.5) |
| BMI, kg/m2, mean (SD) [range] | 27.2 (4.5) [17.8, 45.7] | 27.1 (4.4) [17.8, 45.7] |
| Physical activity, N (%) | ||
| Inactive | 57 (7.9) | 41 (7.8) |
| Low | 270 (37.2) | 194 (37.0) |
| Moderate | 230 (31.7) | 176 (33.5) |
| Very active | 169 (23.3) | 114 (21.7) |
| Cardiovascular disease, N (%) | 56 (7.7) | 43 (8.2) |
| Diabetes, N (%) | 119 (16.4) | 84 (16.0) |
| Hypertension, N (%) | 328 (45.2) | 244 (46.5) |
| Hyperlipidemia, N (%) | 431 (59.4) | 318 (60.6) |
| Osteoarthritis, N (%) | 383 (52.8) | 278 (53.0) |
| Lung disease, N (%) | 93 (12.8) | 69 (13.1) |
| Cancer, N (%) | 209 (28.8) | 153 (29.1) |
| Manual dexterity, mean (SD) [range] | ||
| Pegboard dominant | 12.4 (2.1) [6.0, 18.0] | 12.3 (2.0) [6.0, 17.5] |
| Pegboard nondominant | 11.9 (2.0) [5.0, 18.0] | 11.9 (2.0) [5.0, 18.0] |
| Gross motor function, mean (SD) [range] | ||
| Hand grip strength, kg | 32.2 (10.9) [12.0, 72.0] | 31.6 (10.4) [12.0, 64.0] |
| Chair stands, stands/s | 0.50 (0.17) [0.0, 11.4] | 0.50 (0.16) [0.0, 11.4] |
| Standing balance: full-tandem, N (%) | 646 (89.8) | 462 (89.0) |
| Standing balance: single-leg, N (%) | 413 (57.4) | 289 (55.7) |
| Usual gait speed, m/s | 1.18 (0.23) [0.42, 1.94] | 1.18 (0.23) [0.42, 1.94] |
| Lap time variation | 0.81 (0.38) [0.19, 3.64] | 0.81 (0.39) [0.19, 3.64] |
| Neuroimaging data, cm3, mean (SD) [range] | ||
| White matter hyperintensities | 6.7 (9.5) [0.1, 80.3] | 6.8 (9.6) [0.2, 80.3] |
| Ventricular volume | 33.3 (19.7) [7.6, 131.9] | 33.8 (20.0) [7.6, 131.8] |
| Frontal white matter | 172.3 (21.9) [99.6, 250.4] | 171.8 (21.3) [100.1, 250.4] |
| Parietal white matter | 86.0 (10.6) [55.8, 121.7] | 85.8 (10.4) [55.8, 121.2] |
| Temporal white matter | 102.4 (12.2) [65.7, 142.4] | 102.1 (12.2) [65.7, 142.4] |
| Occipital white matter | 42.4 (5.9) [24.6, 62.3] | 42.3 (5.9) [24.6, 61.5] |
| Frontal gray matter | 179.3 (19.9) [112.7, 253.5] | 178.8 (20.3) [112.7, 253.5] |
| Parietal gray matter | 87.2 (9.6) [62.5, 116.3] | 86.9 (9.5) [62.5, 115.8] |
| Temporal gray matter | 99.6 (11.6) [62.9, 141.0] | 99.3 (11.7) [62.9, 141.0] |
| Occipital gray matter | 72.0 (8.8) [45.8, 98.4] | 71.8 (8.8) [47.8, 98.4] |
| Hippocampus | 7.5 (0.8) [4.8, 10.0] | 7.5 (0.8) [4.8, 10.0] |
| Middle frontal gyrus | 30.6 (4.1) [18.3, 46.8] | 30.6 (4.2) [18.3, 46.8] |
| Precentral gyrus | 22.8 (2.8) [13.9, 30.6] | 22.8 (2.8) [13.9, 30.6] |
| Postcentral gyrus | 17.7 (2.2) [12.1, 26.6] | 17.8 (2.2) [12.6, 24.7] |
| Striatum | 14.5 (1.7) [9.9, 22.5] | 14.5 (1.8) [9.9, 22.5] |
| Baseline Variables | Entire Sample | Sample With ≥ 2 Visits |
|---|---|---|
| (N = 726) | (N = 525) | |
| Age, y, mean (SD) [range] | 70.6 (10.1) [50.2, 95.1] | 71.67 (9.7) [50.2, 95.1] |
| Female, N (%) | 403 (55.5) | 296 (56.4) |
| Black, N (%) | 193 (26.6) | 141 (26.9) |
| Education, y, mean (SD) [range] | 17.6 (2.6) [9.0, 30.0] | 17.6 (2.5) |
| BMI, kg/m2, mean (SD) [range] | 27.2 (4.5) [17.8, 45.7] | 27.1 (4.4) [17.8, 45.7] |
| Physical activity, N (%) | ||
| Inactive | 57 (7.9) | 41 (7.8) |
| Low | 270 (37.2) | 194 (37.0) |
| Moderate | 230 (31.7) | 176 (33.5) |
| Very active | 169 (23.3) | 114 (21.7) |
| Cardiovascular disease, N (%) | 56 (7.7) | 43 (8.2) |
| Diabetes, N (%) | 119 (16.4) | 84 (16.0) |
| Hypertension, N (%) | 328 (45.2) | 244 (46.5) |
| Hyperlipidemia, N (%) | 431 (59.4) | 318 (60.6) |
| Osteoarthritis, N (%) | 383 (52.8) | 278 (53.0) |
| Lung disease, N (%) | 93 (12.8) | 69 (13.1) |
| Cancer, N (%) | 209 (28.8) | 153 (29.1) |
| Manual dexterity, mean (SD) [range] | ||
| Pegboard dominant | 12.4 (2.1) [6.0, 18.0] | 12.3 (2.0) [6.0, 17.5] |
| Pegboard nondominant | 11.9 (2.0) [5.0, 18.0] | 11.9 (2.0) [5.0, 18.0] |
| Gross motor function, mean (SD) [range] | ||
| Hand grip strength, kg | 32.2 (10.9) [12.0, 72.0] | 31.6 (10.4) [12.0, 64.0] |
| Chair stands, stands/s | 0.50 (0.17) [0.0, 11.4] | 0.50 (0.16) [0.0, 11.4] |
| Standing balance: full-tandem, N (%) | 646 (89.8) | 462 (89.0) |
| Standing balance: single-leg, N (%) | 413 (57.4) | 289 (55.7) |
| Usual gait speed, m/s | 1.18 (0.23) [0.42, 1.94] | 1.18 (0.23) [0.42, 1.94] |
| Lap time variation | 0.81 (0.38) [0.19, 3.64] | 0.81 (0.39) [0.19, 3.64] |
| Neuroimaging data, cm3, mean (SD) [range] | ||
| White matter hyperintensities | 6.7 (9.5) [0.1, 80.3] | 6.8 (9.6) [0.2, 80.3] |
| Ventricular volume | 33.3 (19.7) [7.6, 131.9] | 33.8 (20.0) [7.6, 131.8] |
| Frontal white matter | 172.3 (21.9) [99.6, 250.4] | 171.8 (21.3) [100.1, 250.4] |
| Parietal white matter | 86.0 (10.6) [55.8, 121.7] | 85.8 (10.4) [55.8, 121.2] |
| Temporal white matter | 102.4 (12.2) [65.7, 142.4] | 102.1 (12.2) [65.7, 142.4] |
| Occipital white matter | 42.4 (5.9) [24.6, 62.3] | 42.3 (5.9) [24.6, 61.5] |
| Frontal gray matter | 179.3 (19.9) [112.7, 253.5] | 178.8 (20.3) [112.7, 253.5] |
| Parietal gray matter | 87.2 (9.6) [62.5, 116.3] | 86.9 (9.5) [62.5, 115.8] |
| Temporal gray matter | 99.6 (11.6) [62.9, 141.0] | 99.3 (11.7) [62.9, 141.0] |
| Occipital gray matter | 72.0 (8.8) [45.8, 98.4] | 71.8 (8.8) [47.8, 98.4] |
| Hippocampus | 7.5 (0.8) [4.8, 10.0] | 7.5 (0.8) [4.8, 10.0] |
| Middle frontal gyrus | 30.6 (4.1) [18.3, 46.8] | 30.6 (4.2) [18.3, 46.8] |
| Precentral gyrus | 22.8 (2.8) [13.9, 30.6] | 22.8 (2.8) [13.9, 30.6] |
| Postcentral gyrus | 17.7 (2.2) [12.1, 26.6] | 17.8 (2.2) [12.6, 24.7] |
| Striatum | 14.5 (1.7) [9.9, 22.5] | 14.5 (1.8) [9.9, 22.5] |
Note: BMI = body mass index; SD = standard deviation.
Flowchart for final analytic sample used in the study.
(A) Parameterization of the latent variable model used to test associations of baseline manual dexterity with trajectories of brain volumes. Observed brain volume measures, motor function, and covariates are displayed in squares. In circles are latent variables representing level of (“intercept”) and annual rate of change in (“slope”) brain volumes; these are handled as random effects in the between-persons level of the model but constructed using within-individual data. Manual dexterity is a latent variable itself representing the common covariation between pegboard tests. (B). Parameterization of the latent variable model used to test associations of baseline gross motor function with trajectories of brain volumes. Observed brain volume measures, motor function, and covariates are displayed in squares. In circles are latent variables representing level of (“intercept”) and annual rate of change in (“slope”) brain volumes; these are handled as random effects in the between-persons level of the model but constructed using within-individual data. Gross motor is a latent variable itself representing the shared covariation between handgrip strength, chair stands, full tandem and single leg stance, usual gait speed, and lap time variation.
The primary results are displayed in Tables 2 and 3. Both better manual dexterity and gross motor function were cross-sectionally associated with smaller ventricular volume and greater white matter volumes in the frontal, parietal, and temporal lobes. Better manual dexterity was also cross-sectionally associated with larger volumes of parietal gray matter (B = 0.14; 95% CI: 0.05, 0.23), hippocampus (B = 0.10; 95% CI: 0.01, 0.20), postcentral gyrus (B = 0.11; 95% CI: 0.01, 0.20), and occipital white matter (B = 0.10; 95% CI: 0.01, 0.21). The only distinct gross motor association was observed with greater temporal gray matter volume (B = 0.16; 95% CI: 0.05, 0.26). Longitudinally, better manual dexterity and gross motor function were associated with an attenuated rate of change in temporal white matter and occipital gray matter atrophy (Tables 2 and 3). Better manual dexterity was also associated with slower annual increase in ventricular enlargement (B = −0.17; 95% CI: −0.29, −0.05) and slower rate of occipital white matter atrophy (B = 0.39; 95% CI: 0.04, 0.71).
SEM Estimates of Cross-Sectional and Longitudinal Associations Between Manual Dexterity and Brain Volumes
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.788 (0.094) | −0.212 | −0.306, −0.122 | 1.314 (0.116) | −0.174 | −0.291, −0.052 |
| Frontal white matter | 12.931 (0.329) | 0.144 | 0.060, 0.229 | −4.113 (1.193) | 0.255 | −0.217, 0.589 |
| Parietal white matter | 13.570 (0.355) | 0.174 | 0.08, 0.268 | −1.972 (0.621) | 0.330 | −0.043, 0.641 |
| Temporal white matter | 15.820 (0.428) | 0.105 | 0.008, 0.196 | −2.761 (0.854) | 0.420 | 0.011, 0.683 |
| Occipital white matter | 10.936 (0.310) | 0.103 | 0.005, 0.210 | −0.958 (0.352) | 0.393 | 0.037, 0.712 |
| Frontal gray matter | 13.919 (0.388) | 0.065 | −0.016, 0.146 | −3.176 (0.712) | 0.193 | −0.175, 0.565 |
| Parietal gray matter | 12.541 (0.339) | 0.139 | 0.050, 0.226 | −3.374 (0.698) | 0.184 | −0.196, 0.551 |
| Temporal gray matter | 13.736 (0.394) | 0.079 | −0.012, 0.167 | −2.469 (0.529) | 0.139 | −0.137, 0.473 |
| Occipital gray matter | 12.794 (0.380) | 0.048 | −0.046, 0.144 | −1.302 (0.581) | 0.449 | 0.062, 0.787 |
| Hippocampus | 10.546 (0.270) | 0.101 | 0.008, 0.198 | −1.523 (0.156) | −0.083 | −0.229, 0.085 |
| Middle frontal gyrus | 10.266 (0.301) | −0.002 | −0.098, 0.092 | −2.159 (0.413) | 0.021 | −0.232, 0.327 |
| Precentral gyrus | 11.024 (0.304) | 0.100 | −0.002, 0.198 | −2.295 (0.461) | 0.033 | −0.212, 0.302 |
| Postcentral gyrus | 9.438 (0.249) | 0.110 | 0.014, 0.203 | −3.405 (1.31) | 0.158 | −0.177, 0.528 |
| Striatum | 9.999 (0.281) | 0.070 | −0.030, 0.170 | −0.899 (0.421) | −0.033 | −0.489, 0.408 |
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.788 (0.094) | −0.212 | −0.306, −0.122 | 1.314 (0.116) | −0.174 | −0.291, −0.052 |
| Frontal white matter | 12.931 (0.329) | 0.144 | 0.060, 0.229 | −4.113 (1.193) | 0.255 | −0.217, 0.589 |
| Parietal white matter | 13.570 (0.355) | 0.174 | 0.08, 0.268 | −1.972 (0.621) | 0.330 | −0.043, 0.641 |
| Temporal white matter | 15.820 (0.428) | 0.105 | 0.008, 0.196 | −2.761 (0.854) | 0.420 | 0.011, 0.683 |
| Occipital white matter | 10.936 (0.310) | 0.103 | 0.005, 0.210 | −0.958 (0.352) | 0.393 | 0.037, 0.712 |
| Frontal gray matter | 13.919 (0.388) | 0.065 | −0.016, 0.146 | −3.176 (0.712) | 0.193 | −0.175, 0.565 |
| Parietal gray matter | 12.541 (0.339) | 0.139 | 0.050, 0.226 | −3.374 (0.698) | 0.184 | −0.196, 0.551 |
| Temporal gray matter | 13.736 (0.394) | 0.079 | −0.012, 0.167 | −2.469 (0.529) | 0.139 | −0.137, 0.473 |
| Occipital gray matter | 12.794 (0.380) | 0.048 | −0.046, 0.144 | −1.302 (0.581) | 0.449 | 0.062, 0.787 |
| Hippocampus | 10.546 (0.270) | 0.101 | 0.008, 0.198 | −1.523 (0.156) | −0.083 | −0.229, 0.085 |
| Middle frontal gyrus | 10.266 (0.301) | −0.002 | −0.098, 0.092 | −2.159 (0.413) | 0.021 | −0.232, 0.327 |
| Precentral gyrus | 11.024 (0.304) | 0.100 | −0.002, 0.198 | −2.295 (0.461) | 0.033 | −0.212, 0.302 |
| Postcentral gyrus | 9.438 (0.249) | 0.110 | 0.014, 0.203 | −3.405 (1.31) | 0.158 | −0.177, 0.528 |
| Striatum | 9.999 (0.281) | 0.070 | −0.030, 0.170 | −0.899 (0.421) | −0.033 | −0.489, 0.408 |
Note: Intercept: mean (SE) of brain volume; slope: mean (SE) brain volume annual change; all brain regions are standardized to intracranial volume. SEM = structural equation modeling. Models adjusted for age, sex, race, education, body mass index, physical activity, and white matter hyperintensities. BOLD indicates a significant association.
SEM Estimates of Cross-Sectional and Longitudinal Associations Between Manual Dexterity and Brain Volumes
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.788 (0.094) | −0.212 | −0.306, −0.122 | 1.314 (0.116) | −0.174 | −0.291, −0.052 |
| Frontal white matter | 12.931 (0.329) | 0.144 | 0.060, 0.229 | −4.113 (1.193) | 0.255 | −0.217, 0.589 |
| Parietal white matter | 13.570 (0.355) | 0.174 | 0.08, 0.268 | −1.972 (0.621) | 0.330 | −0.043, 0.641 |
| Temporal white matter | 15.820 (0.428) | 0.105 | 0.008, 0.196 | −2.761 (0.854) | 0.420 | 0.011, 0.683 |
| Occipital white matter | 10.936 (0.310) | 0.103 | 0.005, 0.210 | −0.958 (0.352) | 0.393 | 0.037, 0.712 |
| Frontal gray matter | 13.919 (0.388) | 0.065 | −0.016, 0.146 | −3.176 (0.712) | 0.193 | −0.175, 0.565 |
| Parietal gray matter | 12.541 (0.339) | 0.139 | 0.050, 0.226 | −3.374 (0.698) | 0.184 | −0.196, 0.551 |
| Temporal gray matter | 13.736 (0.394) | 0.079 | −0.012, 0.167 | −2.469 (0.529) | 0.139 | −0.137, 0.473 |
| Occipital gray matter | 12.794 (0.380) | 0.048 | −0.046, 0.144 | −1.302 (0.581) | 0.449 | 0.062, 0.787 |
| Hippocampus | 10.546 (0.270) | 0.101 | 0.008, 0.198 | −1.523 (0.156) | −0.083 | −0.229, 0.085 |
| Middle frontal gyrus | 10.266 (0.301) | −0.002 | −0.098, 0.092 | −2.159 (0.413) | 0.021 | −0.232, 0.327 |
| Precentral gyrus | 11.024 (0.304) | 0.100 | −0.002, 0.198 | −2.295 (0.461) | 0.033 | −0.212, 0.302 |
| Postcentral gyrus | 9.438 (0.249) | 0.110 | 0.014, 0.203 | −3.405 (1.31) | 0.158 | −0.177, 0.528 |
| Striatum | 9.999 (0.281) | 0.070 | −0.030, 0.170 | −0.899 (0.421) | −0.033 | −0.489, 0.408 |
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.788 (0.094) | −0.212 | −0.306, −0.122 | 1.314 (0.116) | −0.174 | −0.291, −0.052 |
| Frontal white matter | 12.931 (0.329) | 0.144 | 0.060, 0.229 | −4.113 (1.193) | 0.255 | −0.217, 0.589 |
| Parietal white matter | 13.570 (0.355) | 0.174 | 0.08, 0.268 | −1.972 (0.621) | 0.330 | −0.043, 0.641 |
| Temporal white matter | 15.820 (0.428) | 0.105 | 0.008, 0.196 | −2.761 (0.854) | 0.420 | 0.011, 0.683 |
| Occipital white matter | 10.936 (0.310) | 0.103 | 0.005, 0.210 | −0.958 (0.352) | 0.393 | 0.037, 0.712 |
| Frontal gray matter | 13.919 (0.388) | 0.065 | −0.016, 0.146 | −3.176 (0.712) | 0.193 | −0.175, 0.565 |
| Parietal gray matter | 12.541 (0.339) | 0.139 | 0.050, 0.226 | −3.374 (0.698) | 0.184 | −0.196, 0.551 |
| Temporal gray matter | 13.736 (0.394) | 0.079 | −0.012, 0.167 | −2.469 (0.529) | 0.139 | −0.137, 0.473 |
| Occipital gray matter | 12.794 (0.380) | 0.048 | −0.046, 0.144 | −1.302 (0.581) | 0.449 | 0.062, 0.787 |
| Hippocampus | 10.546 (0.270) | 0.101 | 0.008, 0.198 | −1.523 (0.156) | −0.083 | −0.229, 0.085 |
| Middle frontal gyrus | 10.266 (0.301) | −0.002 | −0.098, 0.092 | −2.159 (0.413) | 0.021 | −0.232, 0.327 |
| Precentral gyrus | 11.024 (0.304) | 0.100 | −0.002, 0.198 | −2.295 (0.461) | 0.033 | −0.212, 0.302 |
| Postcentral gyrus | 9.438 (0.249) | 0.110 | 0.014, 0.203 | −3.405 (1.31) | 0.158 | −0.177, 0.528 |
| Striatum | 9.999 (0.281) | 0.070 | −0.030, 0.170 | −0.899 (0.421) | −0.033 | −0.489, 0.408 |
Note: Intercept: mean (SE) of brain volume; slope: mean (SE) brain volume annual change; all brain regions are standardized to intracranial volume. SEM = structural equation modeling. Models adjusted for age, sex, race, education, body mass index, physical activity, and white matter hyperintensities. BOLD indicates a significant association.
SEM Estimates of Cross-Sectional and Longitudinal Associations Between Gross Motor Function and Brain Volumes
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.627 (0.093) | −0.200 | −0.296, −0.101 | 1.207 (0.112) | −0.089 | −0.225, 0.048 |
| Frontal white matter | 12.944 (0.327) | 0.154 | 0.061, 0.249 | −3.645 (1.131) | 0.239 | −0.262, 0.652 |
| Parietal white matter | 13.490 (0.365) | 0.130 | 0.030, 0.237 | −1.844 (0.701) | 0.249 | −0.220, 0.644 |
| Temporal white matter | 16.129 (0.426) | 0.177 | 0.071, 0.279 | −2.236 (0.654) | 0.508 | 0.042, 0.774 |
| Occipital white matter | 11.034 (0.328) | 0.109 | −0.004, 0.224 | −0.824 (0.391) | 0.203 | −0.206, 0.645 |
| Frontal gray matter | 13.956 (0.385) | 0.070 | −0.020, 0.160 | −2.749 (0.565) | 0.124 | −0.263, 0.616 |
| Parietal gray matter | 12.206 (0.332) | 0.058 | −0.033, 0.156 | −2.809 (0.553) | 0.250 | −0.192, 0.716 |
| Temporal gray matter | 14.024 (0.383) | 0.155 | 0.051, 0.258 | −2.151 (0.473) | 0.278 | −0.073, 0.635 |
| Occipital gray matter | 13.012 (0.383) | 0.101 | −0.010, 0.211 | −0.816 (0.355) | 0.532 | 0.085, 0.796 |
| Hippocampus | 10.546 (0.27) | 0.097 | −0.007, 0.202 | −1.523 (0.156) | −0.004 | −0.200, 0.205 |
| Middle frontal gyrus | 10.505 (0.282) | 0.091 | −0.019, 0.201 | −2.233 (0.474) | 0.042 | −0.294, 0.377 |
| Precentral gyrus | 10.833 (0.309) | 0.026 | −0.077, 0.133 | −2.171 (0.399) | 0.081 | −0.229, 0.398 |
| Postcentral gyrus | 9.11 (0.25) | −0.019 | −0.116, 0.084 | −3.025 (0.898) | 0.007 | −0.377, 0.464 |
| Striatum | 9.967 (0.278) | 0.031 | −0.080, 0.139 | −0.846 (0.374) | −0.049 | −0.542, 0.399 |
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.627 (0.093) | −0.200 | −0.296, −0.101 | 1.207 (0.112) | −0.089 | −0.225, 0.048 |
| Frontal white matter | 12.944 (0.327) | 0.154 | 0.061, 0.249 | −3.645 (1.131) | 0.239 | −0.262, 0.652 |
| Parietal white matter | 13.490 (0.365) | 0.130 | 0.030, 0.237 | −1.844 (0.701) | 0.249 | −0.220, 0.644 |
| Temporal white matter | 16.129 (0.426) | 0.177 | 0.071, 0.279 | −2.236 (0.654) | 0.508 | 0.042, 0.774 |
| Occipital white matter | 11.034 (0.328) | 0.109 | −0.004, 0.224 | −0.824 (0.391) | 0.203 | −0.206, 0.645 |
| Frontal gray matter | 13.956 (0.385) | 0.070 | −0.020, 0.160 | −2.749 (0.565) | 0.124 | −0.263, 0.616 |
| Parietal gray matter | 12.206 (0.332) | 0.058 | −0.033, 0.156 | −2.809 (0.553) | 0.250 | −0.192, 0.716 |
| Temporal gray matter | 14.024 (0.383) | 0.155 | 0.051, 0.258 | −2.151 (0.473) | 0.278 | −0.073, 0.635 |
| Occipital gray matter | 13.012 (0.383) | 0.101 | −0.010, 0.211 | −0.816 (0.355) | 0.532 | 0.085, 0.796 |
| Hippocampus | 10.546 (0.27) | 0.097 | −0.007, 0.202 | −1.523 (0.156) | −0.004 | −0.200, 0.205 |
| Middle frontal gyrus | 10.505 (0.282) | 0.091 | −0.019, 0.201 | −2.233 (0.474) | 0.042 | −0.294, 0.377 |
| Precentral gyrus | 10.833 (0.309) | 0.026 | −0.077, 0.133 | −2.171 (0.399) | 0.081 | −0.229, 0.398 |
| Postcentral gyrus | 9.11 (0.25) | −0.019 | −0.116, 0.084 | −3.025 (0.898) | 0.007 | −0.377, 0.464 |
| Striatum | 9.967 (0.278) | 0.031 | −0.080, 0.139 | −0.846 (0.374) | −0.049 | −0.542, 0.399 |
Notes: Intercept: mean (SE) of brain volume; slope: mean (SE) brain volume annual change; all brain regions are standardized to intracranial volume. SEM = structural equation modeling. Models adjusted for age, sex, race, education, body mass index, physical activity, and white matter hyperintensities. BOLD indicates a significant association.
SEM Estimates of Cross-Sectional and Longitudinal Associations Between Gross Motor Function and Brain Volumes
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.627 (0.093) | −0.200 | −0.296, −0.101 | 1.207 (0.112) | −0.089 | −0.225, 0.048 |
| Frontal white matter | 12.944 (0.327) | 0.154 | 0.061, 0.249 | −3.645 (1.131) | 0.239 | −0.262, 0.652 |
| Parietal white matter | 13.490 (0.365) | 0.130 | 0.030, 0.237 | −1.844 (0.701) | 0.249 | −0.220, 0.644 |
| Temporal white matter | 16.129 (0.426) | 0.177 | 0.071, 0.279 | −2.236 (0.654) | 0.508 | 0.042, 0.774 |
| Occipital white matter | 11.034 (0.328) | 0.109 | −0.004, 0.224 | −0.824 (0.391) | 0.203 | −0.206, 0.645 |
| Frontal gray matter | 13.956 (0.385) | 0.070 | −0.020, 0.160 | −2.749 (0.565) | 0.124 | −0.263, 0.616 |
| Parietal gray matter | 12.206 (0.332) | 0.058 | −0.033, 0.156 | −2.809 (0.553) | 0.250 | −0.192, 0.716 |
| Temporal gray matter | 14.024 (0.383) | 0.155 | 0.051, 0.258 | −2.151 (0.473) | 0.278 | −0.073, 0.635 |
| Occipital gray matter | 13.012 (0.383) | 0.101 | −0.010, 0.211 | −0.816 (0.355) | 0.532 | 0.085, 0.796 |
| Hippocampus | 10.546 (0.27) | 0.097 | −0.007, 0.202 | −1.523 (0.156) | −0.004 | −0.200, 0.205 |
| Middle frontal gyrus | 10.505 (0.282) | 0.091 | −0.019, 0.201 | −2.233 (0.474) | 0.042 | −0.294, 0.377 |
| Precentral gyrus | 10.833 (0.309) | 0.026 | −0.077, 0.133 | −2.171 (0.399) | 0.081 | −0.229, 0.398 |
| Postcentral gyrus | 9.11 (0.25) | −0.019 | −0.116, 0.084 | −3.025 (0.898) | 0.007 | −0.377, 0.464 |
| Striatum | 9.967 (0.278) | 0.031 | −0.080, 0.139 | −0.846 (0.374) | −0.049 | −0.542, 0.399 |
| Cross-Sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| Intercept Mean | Beta | 95% CI | Slope Mean | Beta | 95% CI | |
| Ventricular volume | 1.627 (0.093) | −0.200 | −0.296, −0.101 | 1.207 (0.112) | −0.089 | −0.225, 0.048 |
| Frontal white matter | 12.944 (0.327) | 0.154 | 0.061, 0.249 | −3.645 (1.131) | 0.239 | −0.262, 0.652 |
| Parietal white matter | 13.490 (0.365) | 0.130 | 0.030, 0.237 | −1.844 (0.701) | 0.249 | −0.220, 0.644 |
| Temporal white matter | 16.129 (0.426) | 0.177 | 0.071, 0.279 | −2.236 (0.654) | 0.508 | 0.042, 0.774 |
| Occipital white matter | 11.034 (0.328) | 0.109 | −0.004, 0.224 | −0.824 (0.391) | 0.203 | −0.206, 0.645 |
| Frontal gray matter | 13.956 (0.385) | 0.070 | −0.020, 0.160 | −2.749 (0.565) | 0.124 | −0.263, 0.616 |
| Parietal gray matter | 12.206 (0.332) | 0.058 | −0.033, 0.156 | −2.809 (0.553) | 0.250 | −0.192, 0.716 |
| Temporal gray matter | 14.024 (0.383) | 0.155 | 0.051, 0.258 | −2.151 (0.473) | 0.278 | −0.073, 0.635 |
| Occipital gray matter | 13.012 (0.383) | 0.101 | −0.010, 0.211 | −0.816 (0.355) | 0.532 | 0.085, 0.796 |
| Hippocampus | 10.546 (0.27) | 0.097 | −0.007, 0.202 | −1.523 (0.156) | −0.004 | −0.200, 0.205 |
| Middle frontal gyrus | 10.505 (0.282) | 0.091 | −0.019, 0.201 | −2.233 (0.474) | 0.042 | −0.294, 0.377 |
| Precentral gyrus | 10.833 (0.309) | 0.026 | −0.077, 0.133 | −2.171 (0.399) | 0.081 | −0.229, 0.398 |
| Postcentral gyrus | 9.11 (0.25) | −0.019 | −0.116, 0.084 | −3.025 (0.898) | 0.007 | −0.377, 0.464 |
| Striatum | 9.967 (0.278) | 0.031 | −0.080, 0.139 | −0.846 (0.374) | −0.049 | −0.542, 0.399 |
Notes: Intercept: mean (SE) of brain volume; slope: mean (SE) brain volume annual change; all brain regions are standardized to intracranial volume. SEM = structural equation modeling. Models adjusted for age, sex, race, education, body mass index, physical activity, and white matter hyperintensities. BOLD indicates a significant association.
In sensitivity analyses, we observed significant age × manual dexterity interactions with respect to ventricular volume (interaction coefficient = −0.161; 95% CI = −0.237, −0.082) and parietal white matter (interaction coefficient = 0.085; 95% CI = 0.003, 0.166). Furthermore, there was a significant age × gross motor interaction with respect to ventricular volume (interaction coefficient = −0.148; 95% CI = −0.23, −0.064). No significant interactions were observed in the longitudinal models.
Discussion
In this sample of cognitively normal middle- and older-aged adults, different motor function domains—manual dexterity and gross motor function—show shared and distinct cross-sectional and longitudinal associations with brain atrophy over time. These findings extend previous research by: (i) investigating different motor domains, (ii) examining longitudinal brain volume changes in both gray and white matter cortical and subcortical regions, and (iii) accounting for cerebrovascular burden and physical activity. Our findings provide additional evidence of shared and distinct neural correlates underlying manual dexterity and gross motor function that are independent of cerebrovascular and lifestyle factors.
We observed that both better manual dexterity and gross motor function were associated with preserved temporal white matter volume. These relationships were present in both cross-sectional and longitudinal analyses, including greater temporal white matter volume at baseline and slower progression of white matter atrophy over time. The finding that both constructs of motor function are associated with white matter volume in the temporal lobe fit with research demonstrating white matter structure in the temporal lobe is highly involved in sensory-motor integration (41). Previous studies investigating manual dexterity have reported cross-sectional associations between cerebellar volume (42) and subcortical regions (43). However, others have not observed associations between manual dexterity and brain gray matter volume or white matter integrity in cognitively normal adults (44). The mixed findings reported in the literature may be due to neuroimaging methodology (eg, voxel-based analyses, automated segmentation), varying sample sizes, and participant demographic differences (eg, age). One important consideration is participant cerebrovascular burden. Because white matter hyperintensities are associated with poorer performance on the Purdue Pegboard test (20), the present study provides evidence of the connection between manual dexterity and temporal lobe white matter volume independent of cerebrovascular burden.
Our finding that gross motor function is associated with temporal white matter volume at baseline and follow-up expands on previous research. In a large cross-sectional study (n > 1 500), Sullivan and colleagues reported associations between gross motor function, measured using gait speed, gray matter volume within the temporal and parietal lobes, and total brain volume (45). However, a similar population-based study of community-dwelling older adults reported that gait speed was associated with basal ganglia volume only, among the 16 gray matter regions investigated (46). In addition to usual gait speed, other motor measures have been linked to brain volume including muscular strength (47) and physical fitness (48,49). Although a link between aspects of motor function and brain volume appears to be supported in the literature, there have been contradictions. For instance, a recent study that leveraged data from the UK Biobank did not observe associations between handgrip strength and total brain or hippocampal volume (50). The operationalization of motor function is one apparent explanation for the differing findings reported in the literature. In the present study, latent variable modeling was used to combine 6 motor function measures into a single gross motor domain, which may result in a stronger predictor of brain volume than a single measure of motor function. Indeed, a recent small cross-sectional study (n = 70) found that motor measures that incorporate multiple aspects of functional mobility (eg, dynamic gait and balance) were stronger predictors of brain gray matter regions within the temporal and parietal lobes than gait speed alone (19). The present study extends upon these findings and demonstrates that gross motor function is associated with preserved brain structure in the temporal lobe independent of cerebrovascular burden and physical activity.
Both better manual dexterity and gross motor function were also associated with slower progression of occipital gray matter atrophy over time. Mechanisms underlying these associations are not clear. The Purdue Pegboard test involves visual input (51), which may be related to occipital gray matter structure. The observed association with gross motor function and occipital gray matter atrophy shares some similarities with previous research that reported an association between gait speed and occipital white matter (14). Note that we detected longitudinal associations between these 2 motor domains (ie, manual dexterity, gross motor) and occipital gray matter atrophy over time. There were no observed cross-sectional associations at baseline. This suggests that the rate of occipital gray matter atrophy is more closely tied to motor function. Perhaps there was minimal variance in occipital gray matter volume at baseline, which may have limited our ability to detect a cross-sectional association.
We found unique longitudinal neural correlates of manual dexterity, which were not observed with gross motor function. Better manual dexterity, but not gross motor function, was associated with slower progression of ventricular enlargement, and an attenuated rate of white matter atrophy within the occipital lobe. Previous neuroimaging research that used the Purdue Pegboard test has relied primarily on cross-sectional designs and focused on subcortical regions (42–44). The finding that performance on this manual dexterity test is associated with future ventricular enlargement, a measure of central brain atrophy, extends previous research. With respect to occipital white matter, these longitudinal findings compliment a cross-sectional investigation that found performance on the Purdue Pegboard test associated with diffusion tensor imaging measures of white matter integrity in the occipital region, but not in the frontal, parietal, or temporal regions (52).
The cross-sectional neural correlates with both motor function domains include global brain health indicated by ventricular volume and white matter volumes in the frontal, parietal, and temporal lobes. Manual dexterity was associated with hippocampal volume, whereas gross motor function was associated with total temporal gray matter volume. Because the domain of gross motor function involves several motor components, this may have contributed to the association with overall temporal gray matter. Manual dexterity was also cross-sectionally associated with parietal regions, including the postcentral gyrus, which plays a key role in the planning and execution of motor reach tasks (53). Testing the interaction effect of age suggests manual dexterity and gross motor function may be more strongly associated with ventricular volume and parietal white matter in older adults; however, no other interactions were observed in the other brain regions, indicating that the strength of the observed cross-sectional associations is similar within middle- and older-age participants. The dispersed cross-sectional relationships observed in the current study expand upon previous research and demonstrate that manual dexterity and gross motor performance have shared and distinct associations with cortical and subcortical gray and white matter regions.
The BLSA consists of a well-characterized sample of community-dwelling middle- and older-aged adults with longitudinal brain imaging data. However, there are study limitations to consider. We cannot determine whether brain atrophy contributed to manual dexterity or gross motor function decline. It is possible that structural brain changes contribute to future change in manual dexterity and/or gross motor function as previous studies have demonstrated that brain atrophy contributes to future gait speed decline (14,17). Further, we did not test hemispherical differences within our sample (ie, right vs left cortical regions). Future studies that assess changes in motor function and lateralization are needed to better elucidate directionality and specificity of the observed associations. Although this study incorporated numerous motor function variables, the latent construct of manual dexterity relied on Purdue Pegboard test performance. Although this test is a valid and reliable measure of manual dexterity, research has shown that low vision can also influence performance (51). Thus, we cannot rule out the possibility that visual deficits may have affected our manual dexterity measure. Lastly, although our models were adjusted for several important covariates, there are additional clinical factors that may have affected the strength of the observed associations.
In summary, the present study provides evidence that within cognitively normal middle- and older-aged community-dwelling adults, better manual dexterity and gross motor function are associated with larger brain gray and white volume across cortical and subcortical regions, and an attenuated rate of temporal, occipital, and central brain atrophy. Notably, these findings accounted for cerebrovascular and lifestyle factors shown to affect brain gray and white matter volume. This work contributes to research investigating the connection between motor function and the brain by detailing shared and distinct associations of manual dexterity and gross motor function with brain gray and white matter atrophy.
Funding
This research was supported by the Intramural Research Program of the National Institute on Aging of the National Institutes of Health and by grant R01AG061786. Data used in the analyses were obtained from the Baltimore Longitudinal Study of Aging, a study performed by the National Institute on Aging Intramural Research Program. R.J.D. was supported by an NIA grant K01AG080122.
Conflict of Interest
E.M.S., L.F., and J.A.S. currently serve on the editorial board of JGMS. The other authors declare no conflict of interest.
Author Contributions
Drafting/revising the manuscript for content: R.J.D., H.W., A.L.G., J.A.S., Y.A., C.D., Y.C., E.M.S., L.F., S.M.R., and Q.T. Study concept or design: R.D., H.W., A.L.G., J.A.S., Y.A., E.M.S., L.F., S.M.R., and Q.T. Analysis or interpretation of data: R.D., H.W., A.L.G., J.A.S., Y.A., C.D., Y.C., E.M.S., L.F., S.M.R., and Q.T.
