High-resolution structural connectivity mediates age-related differences in functional connectivity and fluid cognition

Jenna L Merenstein; Allen W Song; David J Madden

doi:10.1093/braincomms/fcaf376

. 2025 Sep 28;7(5):fcaf376. doi: 10.1093/braincomms/fcaf376

High-resolution structural connectivity mediates age-related differences in functional connectivity and fluid cognition

Jenna L Merenstein ^1,^2,^✉, Allen W Song ^3,⁴, David J Madden ^5,⁶

PMCID: PMC12528985 PMID: 41113674

Abstract

Magnetic resonance imaging studies using diffusion-weighted imaging suggest that age-related cognitive decline and alterations in brain function, in healthy adults, are at least partly explained by the degradation of white matter pathways connecting distributed brain regions. Studies of younger adults and animal models suggest that more precise estimates of white matter connectivity may be achieved by higher resolution, relative to standard spatial resolution, diffusion-weighted imaging. Here, in a cross-sectional study of healthy adults across the lifespan (n = 140; ages 18–88 years; 72 females), we compared age-related differences in measures of white matter structural connectivity from standard (1.5 mm³ voxels; 3.375 µl volume) and high-resolution (1 mm³; 1 µl volume) diffusion-weighted imaging, and their ability to explain age-related differences in functional connectivity and cognition. We assessed cognition using tests of memory, executive function, and perceptual-motor speed, and assessed structural and functional (resting-state functional magnetic resonance imaging) connectivity using graph theory. Results revealed more pronounced age-related decreases in structural connectivity for sensorimotor, ventral attention, and sub-cortical networks for high-resolution than standard diffusion-weighted imaging. Age-related decreases in functional connectivity were evident across the brain and mediated by high-resolution structural connectivity in the default mode network. Age-related decline in fluid cognition was mediated by within-network connectivity from only high-resolution diffusion-weighted imaging, but by a combination of high-resolution and standard diffusion-weighted imaging for between-network connectivity. Thus, relative to standard diffusion-weighted imaging, high-resolution diffusion-weighted imaging may better capture age-related differences in white matter connectivity and their constraint on age-related alterations in brain function and cognitive performance.

Keywords: graph theoretical analyses, multiplexed sensitivity encoding, diffusion-weighted imaging, healthy aging, white matter

Merenstein et al. reported that structural brain connectivity, as measured by high-resolution diffusion-weighted MRI, improved the ability to predict age-related alterations in brain function and cognitive performance, relative to standard spatial resolution. These results may assist with the early identification of adults who are at heightened risk for cognitive decline.

Graphical Abstract

Introduction

Healthy aging is accompanied by some degree of decline in fluid, speed-dependent cognition.^1-5 Theories based on studies assessing brain structure using MRI techniques such as diffusion-weighted imaging (DWI) propose that the disconnection of white matter (WM) pathways between distributed brain regions contributes to age-related cognitive decline.^6-9 However, higher resolution DWI acquisitions may help us better understand age-related structural disconnection, and its relation to brain function, which is necessary to inform future interventions aimed at slowing neurocognitive aging.

White matter plays a crucial role in cognitive function, such that heavily myelinated fibres propagate action potentials among different grey matter (GM) regions in an efficient manner.^7,10,11 In theory, declines in DWI measures of WM structural connectivity should constrain functional MRI (fMRI) measures of GM functional connectivity in healthy aging,^12,13 such that faster axonal propagation of action potentials between pre- and post-synaptic neurons should benefit cognitive performance.⁷ However, studies report mixed results regarding the presence and direction of these structure-function associations in adults across the lifespan.^14-17 For example, some studies have identified significant age-related decreases in the association between brain structure and brain function,^18,19 and other studies have reported relatively independent age-related effects on structural and functional connectivity.^16,17,20,21 One possibility for this discrepancy in the literature is that functional connectivity measurements vary with task demands (e.g. cognitive load versus resting-state), whereas WM connectivity is instead relatively static at the time of measurement.^22,23 A second possibility, to be explored here, is that these mixed findings may be due to the relatively low spatial resolution of standard DWI acquisitions attenuating the ability to detect these associations. Compared to standard DWI, higher resolution DWI should better estimate WM fibres in small, tightly curved regions, and voxels with multiple tissue types,^24-26 and in turn, better capture age-related differences in functional connectivity.

To facilitate comparisons between different MRI modalities, an innovative approach is to assess both structural and functional connectivity using graph theory,²⁷ which divides the brain into networks comprised of anatomical nodes (i.e. GM regions) and the edges among them (i.e. WM connections). Graph theory has been a successful tool for identifying age-related differences in structural and functional connectivity, and for assessing relations between these measures.^17,18,20,28 To date, however, no study has tested whether high-resolution structural connectivity better explains age-related variance in functional connectivity, when directly compared to standard resolution DWI.

Here, we used a sample of 140 healthy adults (18–88 years of age) who underwent resting-state fMRI (rsfMRI) and standard resolution DWI (1.5 mm³ voxels; 3.375 µl volume) and high-resolution DWI (1 mm³; 1 µl volume) data acquisition. An extensive test battery assessed both general and specific forms of fluid cognition, and graph theoretical analyses assessed structural and functional connectivity among cortical and sub-cortical networks.^29,30 The goals of this study were 3-fold. First, we aimed to replicate our previous work using a subset of these data (n = 61) demonstrating that the 3-fold increase in resolution from the high-resolution relative to standard acquisition uniquely identified assessed patterns of age-related decreases in WM connectivity both within- and between-networks.³¹ We conducted these analyses across the entire brain to assess whether this pattern was more evident in certain networks, such as those with greater superficial WM regions (e.g. sensorimotor network) versus long-range connections (e.g. default mode network). Second, we aimed to replicate previous reports of age-related decreases in within-network functional connectivity, without any corresponding difference in between-network connectivity.^17,32 We extended this work by testing whether these patterns are evident even when including adults aged 80+ years.³³ Our third goal, representing a novel extension that has not yet been assessed in prior work,³¹ tested our primary hypothesis: that high-resolution structural connectivity should significantly mediate age-related differences in functional connectivity and fluid cognition, beyond standard resolution DWI.

Studies of healthy younger adults report stronger relations between structural and functional connectivity in primary sensory networks.³⁴ However, studies of healthy aging report larger age-related differences in functional connectivity for association (e.g. default mode) than primary sensory (e.g. visual) networks.^32,35 Thus, we expected that healthy aging may instead be associated with stronger structure-function relations for association than sensory networks,³⁶ and that this may reflect their vulnerability to underlying age-related structural degradation.

Materials and methods

Participants

This study was conducted in compliance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and the protocol was reviewed and approved by the Institutional Review Board for Duke University Medical Center. All participants were compensated for their time and provided informed consent.

Participants were recruited from the Raleigh-Durham, North Carolina area and surrounding communities using a combination of printed or virtual flyers distributed to local hospitals, community centres and other local organizations. Study participation initially involved a screening and behavioural testing session, followed by an MRI scan that was acquired approximately one month later. Inclusion criteria were: at least a high school education, free of major medical conditions as assessed by a modified version of the Christensen et al.³⁷ questionnaire (e.g. diabetes, uncontrolled hypertension, epilepsy, prior loss of consciousness and recent diagnosis of cancer), no current prescription medications known to affect cognition (e.g. stimulants and antidepressants), intact general cognition on the Mini-Mental State Exam³⁸ (score ≥ 27; n = 61) or the Montreal Cognitive Assessment³⁹ (score ≥ 26; n = 79), age-expected performance on the Wechsler Adult Intelligence Scale-III vocabulary subtest⁴⁰ (score > 50th percentile), no evidence of major depression on the Beck Depression Inventory⁴¹ (score ≤ 16), and adequate visual acuity on the Freiburg Visual Acuity Test⁴² (Snellen score < 20/40).

For the initial screening session, 252 healthy adults from the community participated (18–88 years of age). We excluded 73 of these participants due to attrition (n = 8), failing to meet the inclusion criteria or MRI safety requirements (e.g. body size, claustrophobia, implanted ferrous metal; n = 63), or for being outliers (>3 standard deviations) on one or more of the cognitive tests (n = 2; described below, Psychometric and Cognitive Tasks). The remaining 179 participants completed MRI scanning, but we were prohibited from acquiring the high-resolution DWI data for 12 individuals due to time constraints. We excluded 27 additional participants due to extreme head motion, MRI artefacts or issues with image pre-processing, or poor performance (accuracy < 75%) on a visual attention task completed during fMRI scanning. Table 1 includes details on the final sample of 140 participants 18–88 years of age (51.4% female, 8.6% Hispanic, 83.6% White). The age distribution is as follows: 18–29 (n = 26), 30 s (n = 22), 40 s (n = 22), 50 s (n = 17), 60 s (n = 22), 70 s (n = 20) and 80 s (n = 11).

Table 1.

Participant characteristics

Measure	Mean	(SD)	r with age	P-value
Education (years)	17.050	(2.249)	0.057	0.506
MMSE (n = 61)	29.541	(0.621)	0.004	0.978
MoCA (n = 79)	28.354	(1.197)	−0.279	0.013
BDI	3.514	(3.851)	0.030	0.723
Vocabulary	56.000	(4.686)	0.002	0.981
Colour vision	13.886	(0.368)	−0.257	0.002
Visual acuity	−0.015	(0.090)	−0.367	<0.001
Scan Interval	36.971	(27.444)	−0.036	0.675

Open in a new tab

The n = 140 for all tests except the Mini-Mental State Exam (MMSE)³⁸ and Montreal Cognitive Assessment (MoCA).³⁹ Values are presented as mean (SD). BDI = score on the Beck Depression Inventory⁴¹; Vocabulary = raw score on the vocabulary subtest of the Wechsler Adult Intelligence Scale III⁴⁰; Colour Vision = score on Dvorine colour plates⁴³; Visual Acuity = inversed logarithm of the minimum angle of resolution for the Freiburg visual acuity test,⁴² Scan Interval = number of days between the initial screening session and MRI scan. Significant effects are presented in bold.

Psychometric and cognitive tasks

As described previously,^20,44-46 participants completed a fluid cognition battery comprised of 12 individual tests. Perceptual-motor speed was measured by three computer-based reaction time (RT) tests (simple RT, two versions of choice RT), and number correct in 85 s. from the National Institutes of Health (NIH) Toolbox Pattern Comparison Test.⁴⁷ Executive function was measured by the Digit Symbol Substitution Task, Flanker Task (incompatible versus compatible RT), Trail Making Test Part B versus A,⁴⁸ and NIH Toolbox Card Sort Test.⁴⁷ Memory was measured by a computer-based 6-item change detection task,⁴⁹ 20-min. delayed verbal recall of 16 words, the Logical Memory Test,⁴⁰ and the NIH Toolbox Picture Sequence Test.⁴⁷

Imaging data acquisition

Imaging data acquisition occurred at the brain imaging and analysis centre at Duke University Medical Center across two separate experiments of visual attention that occurred between 2021 and 2024. Most of the data (n = 92) were acquired on a 3T GE MR750 MRI scanner equipped with an eight-channel head coil. In 2022, the scanner underwent a series of upgrades, and the remainder of the data (n = 48) were acquired on a 3T GE Ultra High-Performance MRI scanner equipped with a 48-channel head coil. Participants wore earplugs to protect them from scanner noise, and head motion was minimized by using foam padding.

For 61 participants, a high-resolution T1-weighted image volume was acquired using a 3D fast inverse-recovery-prepared spoiled gradient recalled echo (SPGR) sequence with these parameters: voxel size = 0.5 × 0.5 × 0.5 mm, acquisition matrix = 512 × 512 × 292 mm, repetition time (TR) = 2200 ms, echo time (TE) = 3.1 ms, flip angle = 8°, sensitivity encoding (SENSE) factor = 2. For the remaining 79 participants, a high-resolution T1-weighted image volume was acquired using a 3D inverse-recovery-prepared SPGR sequence with the following parameters: TR = 2132.8 ms, TE = minimum full echo, flip angle = 8°, voxel size = 1 × 1 × 1 mm, acquisition matrix = 256 × 256 × 116 mm (n = 9) or 256 × 256 × 124 (n = 70) and SENSE factor = 2.

For all 140 participants, whole-brain standard resolution DWI data were acquired using a single-shot spin-echo echo planar imaging (EPI) sequence with these parameters: acquisition matrix = 144 × 144 × 83 mm, TR = 4620 ms, TE = 64.1 ms, flip angle = 90°, voxel size = 1.5 × 1.5 × 1.5 mm, SENSE factor = 1 and multi-band factor = 3. Diffusion-weighted gradients were applied in 90 directions with b values of 1500 s/mm² and 3000 s/mm² and with two non-diffusion-weighted b = 0 images. For most participants (n = 129), a second diffusion sequence was acquired with six phase-encoding directions of opposite polarity using the same parameters as the main sequence, except that TR = 4971 ms. Subsequent analyses of these data were limited only to the data acquired using the b = 1500 s/mm² shell to facilitate more comparable analyses with the high-resolution data.

For all 140 participants, whole-brain high-resolution DWI data were acquired using a four-shot multiplexed sensitivity encoding (MUSE) sequence. For 74 participants, single-band data were acquired using a spin-echo EPI sequence with these parameters: voxel size = 1 × 1 × 1 mm, acquisition matrix = 256 × 256 × 90 mm (n = 61) or 256 × 256 × 116 mm (n = 13), TR = 13 000 ms, TE = 58 ms, flip angle = 90°, SENSE factor = 1 and multi-band factor = 1. Diffusion-weighted gradients were applied in 15 directions with a b value of 1000 s/mm² (n = 61) or 800 s/mm² (n = 13) and with two non-diffusion-weighted b = 0 s/mm² images. For the remaining 66 participants, multi-band data were acquired using a spin-echo EPI sequence with these parameters: voxel size = 1 × 1 × 1 mm, acquisition matrix = 164 × 328 × 122 mm, TR = 8767 ms, TE = 60.8 ms, flip angle = 90°, SENSE factor = 1 and multi-band factor = 2. Diffusion-weighted gradients were applied in 25 directions with a b value of 800 s/mm² and with two non-diffusion-weighted b = 0 images. For nearly all participants with multi-band data (n = 64), a second diffusion sequence of one b = 0 s/mm² volume and one 800 s/mm² volume was acquired with phase-encoding directions of opposite polarity using identical parameters.

For all 140 participants, rsfMRI data were acquired (eyes open, central fixation) using a T2*-weighted gradient-echo EPI sequence with these parameters: voxel size = 2 × 2 × 2 mm, acquisition matrix = 128 × 128 × 50 mm (n = 61) or 128 × 128 × 63 mm (n = 79), TR = 1500 ms, TE = 30 ms, flip angle = 60°, SENSE factor = 2 and number of volumes = 300. For the 79 participants with 63 slices acquired, a set of field maps were acquired using an EPI sequence with phase-encoding directions of opposite polarity and identical parameters to the rsfMRI data, except that TR = 1000 ms and flip angle = 90°.

During scanning, we also acquired data from task-related fMRI, susceptibility-weighted angiography, and fluid attenuated inversion recovery imaging, most of which has already been reported in separate articles.^17,45,46,50

Processing of DWI data

We first extracted the standard resolution DWI data limited only to the b = 1500 s/mm² shell (dwiextract) to avoid the potential confounding effect of gradient strength in comparisons between the two acquisitions of interest.^51,52 All DWI data were then pre-processed using MRtrix3⁵³ based on the following sequence of steps: denoising (dwidenoise), correction for motion, eddy current-induced distortions and susceptibility-induced distortions (dwifslpreproc), bias-correction (dwibiascorrect) and generation of a skull-stripped whole-brain mask (dwi2mask). The only exception to this pipeline was that the correction of susceptibility-induced distortions was not performed for the single-band MUSE data because the conventional SENSE reconstruction procedure reconstructed full field-of-view images from four DWI segments and allowed for the calculation of shot-to-shot phase variations.⁵⁴ A trained researcher visually inspected all processed DWI scans and deemed them to be acceptable for the coverage of the brain mask, sufficient correction of motion and susceptibility-induced distortions, the absence of MR artefacts (e.g. ghosting, radio frequency inhomogeneities), and no major anatomical abnormalities.

To estimate the structural connectome, we first used the Dhollander⁵⁵ algorithm to estimate tissue-specific response functions from GM, WM and CSF (dwi2response). We then derived fibre orientation distribution (FOD) maps (dwi2fod) using multi-shell, multi-tissue constrained spherical deconvolution (mstmt-csd) based on two compartments (WM and CSF) because the data comprised only a single shell, and performed normalization (mtnormalize) of the resulting FOD maps.^53,56,57 Next, we generated a GM/WM boundary mask based on the high-resolution T1-weighted image volume (5ttgen and 5tt2gmwmi). For each DWI acquisition, the average skull-stripped, pre-processed b = 0 s/mm² image was registered to the GM/WM boundary mask using a six degrees of freedom linear registration in FSL’s flirt.^58-60 The inverse of this transformation matrix was then used to register the GM/WM boundary mask to native diffusion space (transformconvert, mrtransform). Using the FOD maps, we generated 10 million tracts (tckgen) using anatomically constrained probabilistic tractography (minimum streamline length = 1 mm, maximimum length = 250 mm, FOD amplitude cutoff = 0.06). Using anatomically constrained probabilistic tractography helps ensures a biologically plausible termination of streamlines in seed regions across the brain (GM nodes within the Brainnetome atlas, described under Graph Theoretical Analyses) by forcing their termination within the GM/WM boundary mask.⁶¹ We used spherical-deconvolution informed filtering of tractograms 2 (tcksift2) to assign a weight to each streamline, relative to the underlying apparent fibre density.⁶² Instead of removing streamlines, this approach incorporates a weighted contribution of all streamlines to the spherical deconvolution diffusion model,⁶³ represents the fibre density in each voxel more accurately, and helps prevent issues related to overfitting.⁶²

Processing of rsfMRI data

All rsfMRI data were pre-processed using fMRIPrep⁶⁴ based on the following sequence of steps: generating a skull-stripped functional reference image from the first four image volumes, estimating susceptibility-induced distortions, estimating head-motion parameters with respect to the functional reference image (transformation matrices, and six corresponding rotation and translation parameters), and correcting for slice timing (AFNI’s 3dShift). No spatial smoothing was performed.⁶⁵ For the 61 participants that field maps were not acquired for, we used a boundary-based registration to correct for susceptibility-induced distortions by aligning the functional reference image to the high-resolution T1-weighted image volume with inverted intensity.⁶⁶ For the remaining 79 participants, susceptibility-induced distortions were correcte4d by using 3dQwarp to estimate the B0-non-uniformity map from the reverse polarity EPI field maps.⁶⁷ Lastly, we resampled the corrected functional time-series onto the participants’ native space by applying a single, composite transform that was used for motion and susceptibility-induced distortion correction.

After processing in fMRIPrep, the rsfMRI data were then fed into xcpEngine (https://xcpengine.readthedocs.io/) to remove the first four brain volumes, perform confound regression, scrub high-motion volumes, and apply a temporal band-pass filter (0.01–0.08 Hz). Confound regression included frame-to-frame motion estimates, mean signals from WM and CSF, the mean global signal, and the temporal and quadratic derivatives of these signals, yielding 36 nuisance regressors.⁶⁸ Before motion scrubbing, mean frame-wise displacement (FD)⁶⁹ was 0.130 mm (SD = 0.053) and exhibited a positive correlation with age, r = 0.259, P = 0.002. We used spectral interpolation to scrub all resting-state volumes with FD > 0.50 mm. If fewer than five volumes remained between two outlier volumes, we scrubbed both outliers and all the intervening volumes to account for residual motion.^70-72 Mean duration of the scrubbed rsfMRI data across all participants was 7.40 min; (SD = 0.24), with a minimum of 5.85 min.

Graph theoretical analyses

To assess connectivity, we used a standard network partition based on rsfMRI data derived from a separate study of healthy younger adults.²⁹ The networks were based on a parcellation of 210 cortical and 36 sub-cortical GM nodes from the Brainnetome atlas³⁰ in Montreal Neurological Institute (MNI) 152 template space (1 mm³ resolution). For the estimation of both structural and functional connectivity, each cortical node was assigned one of the following seven networks: dorsal attention, default mode, frontoparietal, limbic, sensorimotor, ventral attention and visual networks.²⁹ We combined the 36 sub-cortical nodes into a single sub-cortical network mask, similar to prior work.^46,73,74 Unfortunately, the slice prescription for both the high-resolution DWI and the rsfMRI data from the first experiment (n = 61) did not have sufficient spatial coverage of inferior temporal regions, mostly in the limbic network, for ≥20 participants. In addition to its insufficient spatial coverage here, the limbic network is highly prone to EPI-related artefacts and high signal-to-noise ratios,⁷⁵ and we therefore excluded the 26 limbic network nodes from our calculation of structural and functional connectivity, leaving a final set of 220 usable Brainnetome nodes. Applying the same functionally defined network partition for both the DWI and rsfMRI data allows for more direct comparisons between these modalities and aligns with our prior study that used this same partition to assess only structural connectivity.³¹ Our goal here was to assess how the measurement of structural connections between GM regions in various functional networks differed by DWI spatial resolution, rather than to characterize the unique modular structure of each imaging domain.

To assess structural connectivity for each participant and DWI acquisition type, the estimated streamlines in native diffusion space were transformed to standard MNI 152 1 mm³ space (tcktransform). We created symmetrical structural connectivity matrices by assigning the 10 million weighted streamlines to the final set of 220 Brainnetome nodes (tck2connectome), and then used a 2 mm radial search to asign the endpoint of each streamline to the nearest GM node.⁶² In the resulting unthresholded 220 × 220 matrices,⁷⁶ the value in each cell corresponding to the sum of the weighted contribution of the streamlines connecting that pair of nodes (scaled by node volume).⁷⁷

To assess resting-state functional connectivity for each participant, we used aligned the residualized image from xcpEngine (i.e. after motion scrubbing) to standard MNI 152 1 mm³ space using the antsRegistration tool. Symmetrical functional connectivity matrices were created using the correlations (Pearson r) of the time-series data for each of the 220 Brainnetome nodes, separately for each participant. In the resulting unweighted 220 × 220 matrices, negative correlations were set to zero,²⁷ the diagonal was excluded, and the r values were transformed to Fisher z values.

Lastly, for the estimation of both structural and functional connectivity, we used the Brain Connectivity Toolbox²⁷ to estimate graph theoretical measures of within-network connectivity (i.e. average connectivity strength among nodes belonging to a single network) and between-network connectivity (i.e. average connectivity strength between nodes belonging to a single network and all nodes beyond that network), separately for each of the seven networks of interest (i.e. excluding the limbic network).

Experimental design and statistical analysis

Statistical analyses were primarily conducted within R Studio (version 2024.04.2) with a handful of analyses (i.e. partial correlations with more than 1 covariate) conducted within JASP (version 0.17.1). A χ² test indicated that distribution of the different scanner features did not significantly differ among younger (18–39 years of age), middle-age (40–59 years of age) and older (60–88 years of age) adults, χ² = 0.883, P_uncorrected = 0.643. However, scanner setup was nonetheless used as a covariate for all subsequent analyses with the MRI data. All results were corrected for multiple comparisons by applying false discovery rate (FDR) procedures.⁷⁸

For the psychometric and cognitive data, the unrotated first factor derived from a principal axis factor analysis of all 12 tests was used to represent general fluid cognition, and memory, executive function, and perceptual-motor speed were represented by the first factor from each set of four tests within those individual domains.^9,20,79-81 Our primary interest was the association between age and general fluid cognition and was tested using Pearson correlational analyses, but we also conducted similar exploratory analyses between age and each of the individual cognitive measures.

For the structural connectivity data, we first assessed the degree of similarity between the standard resolution and high-resolution measures for within- and between-network connectivity of each of the seven networks by conducting partial correlational analyses, while covarying for scanner setup. We then conducted partial correlational analyses between age and each of the seven network connectivity measures, separately for the standard resolution and high-resolution data, while covarying for scanner setup. We compared the size of the age-related correlations between the standard resolution and high-resolution acquisitions using the two-tailed Steiger’s z test approach for comparing correlated correlation coefficients.^82,83

For the functional connectivity data, we conducted partial correlational analyses between age and each of the seven within- and between-network network connectivity measures, while covarying for scanner setup and head motion (FD). To test the hypothesis that high-resolution structural connectivity explains age-related differences in functional connectivity, beyond the standard resolution structural connectivity measures, we conducted parallel mediation analyses using the PROCESS macro in R Studio.⁸⁴ Significant indirect effects, corresponding to the interaction between model paths, were determined by bias-corrected 95% confidence intervals (CIs) that did not contain zero after 10 000 bootstrap replacements. In these models, we included age as the predictor, scanner setup and head motion as covariates, functional connectivity as the outcome, and high-resolution and standard resolution structural connectivity measures as parallel mediators, with separate analyses for each network (e.g. frontoparietal structural connectivity as a mediator of age-related differences in frontoparietal functional connectivity). These parallel mediation analyses were conducted separately for measures of within- or between-network connectivity and were limited only to functional connectivity outcome measures that exhibited a significant association with age.

Lastly, we sought to understand whether any of the three forms of connectivity (resting-state functional, high-resolution structural and standard resolution structural) explained age-related differences in general fluid cognition. We again conducted parallel mediation analyses in the exact same way as above, except that now all three forms of connectivity were included as parallel mediators and general fluid cognition performance was modelled as the outcome variable. These models were again conducted separately for measures of within- or between-network connectivity, and the choice of mediator variables was limited only to measures of connectivity from networks that exhibited a significant association with age.

Results

Age-related differences in cognitive performance

To assess age-related differences in cognition, we focused on the general fluid cognition score but also explored the individual domain factor scores for memory, executive function and perceptual-motor speed. The individual domain factors were each significantly correlated with one another, Pearson r ≥ 0.479, P_FDR < 0.001. To account for unique domain-specific variance in performance, we residualized the individual domain scores to account for performance on the two factor scores outside of that domain.^20,79,81 We then conducted Pearson correlations between age and each cognitive factor, with FDR correction applied for comparisons across four measures. Results indicated that increased age was significantly correlated with lower performance for general fluid cognition, perceptual-motor speed, and memory, but was not significantly related to executive function (Fig. 1).

Structural connectivity

Correlations Between DWI Acquisitions

To examine associations between high-resolution and standard resolution structural connectivity, we used separate partial correlations (covaried for scanner setup), with FDR correction applied for comparisons across seven networks. Results indicated that the two acquisitions were generally highly correlated for within-network connectivity, with correlation coefficients ranging from 0.269 to 0.668 (Table 2; r High-Res—Standard). The correlations for between-network connectivity were also highly correlated, with correlation coefficients ranging from 0.375 to 0.743 (Table 2, r High-Res—Standard).

Table 2.

Age correlations with white matter structural connectivity

	r Standard—age		r High-res—age		r High-res—standard		Steiger’s z
Within-network connectivity
VIS	−0.405	***	−0.277	**	0.580	***	1.771
SMN	−0.178		−0.468	*******	0.463	***	3.589	*
DAN	−0.256	**	−0.152		0.668	***	1.535
VAN	−0.095		−0.300	*******	0.540	***	2.586	*
FPN	0.122		−0.024		0.463	***	1.658
DMN	−0.216	*	−0.348	*******	0.638	***	1.916
SUB	−0.215	*	−0.447	*******	0.269	**	2.465	*
Between-network connectivity
VIS	−0.474	*******	−0.500	*******	0.743	***	0.496
SMN	−0.312	*******	−0.547	*******	0.375	***	2.872	*
DAN	−0.351	*******	−0.453	*******	0.593	***	1.479
VAN	−0.355	*******	−0.519	*******	0.494	***	2.207
FPN	−0.416	*******	−0.519	*******	0.507	***	1.429
DMN	−0.496	*******	−0.584	*******	0.607	***	1.445
SUB	−0.550	*******	−0.614	*******	0.722	***	1.286

Open in a new tab

Networks are abbreviated as visual (VIS), sensorimotor (SMN), dorsal attention (DAN), ventral attention (VAN), frontoparietal (FPN), default mode (DMN), and sub-cortical (SUB). Asterisks correspond to the false discovery rate corrected P-value for the correlation coefficient presented in the preceding column. All correlations were covaried for scanner setup. Significant effects are presented in bold. *** = P_FDR < 0.001, ** = P_FDR < 0.01, * = P_FDR < 0.05.

To test whether these similarities were affected by age, we repeated the correlations using age as an additional covariate. The correlations between acquisitions generally remained the same for within-network connectivity, with correlation coefficients ranging from 0.197 to 0.659 (Supplementary Table 1). For between-network connectivity, the correlations generally decreased after included age as a covariate, but nonetheless remained strongly correlated, with correlation coefficients ranging from 0.257 to 0.663 (Supplementary Table 1).

Age-related differences

To assess age-related differences in structural connectivity, we conducted partial correlations between age and within- or between-network connectivity, separately for each acquisition type, while covarying for scanner setup and applying FDR correction for comparisons across seven networks. For within-network connectivity (Fig. 2), the high-resolution data yielded significant age-related decreases for all networks except the dorsal attention and frontoparietal networks (Table 2; r High-Res—Age), and the standard resolution data yielded significant age-related decreases for all networks except the ventral attention and frontoparietal networks (Table 2; r Standard—Age). For between-network connectivity (Fig. 3), both the high-resolution (Table 2; r High-Res—Age) and standard resolution (Table 2; r Standard—Age) data yielded significant age-related decreases for all networks.

**Age-related differences in between-network structural connectivity.** Scatterplots display associations between age and between-network connectivity for each network, shown in Panels (A-G), separately for the high-resolution (black triangles) and standard resolution (coloured circles) data (n = 140). The shaded grey areas around the regression lines represent 95% confidence intervals. Significant effects from partial correlation analyses are presented in bold. The reported correlation coefficients were covaried for the type of scanner setup used. Each data point corresponds to an individual participant’s measure of structural connectivity for that particular network. FDR = false discovery rate.

To determine whether the magnitude of age-related decreases in structural connectivity significantly differed between the standard resolution and high-resolution acquisitions, we compared the correlated correlation coefficients using two-tailed Steiger’s z tests.^82,83 These results indicated that the age-related correlations were significantly larger for the high-resolution relative to standard resolution data for within-network connectivity in the sensorimotor, ventral attention and sub-cortical networks, and between-network connectivity in the sensorimotor network (Table 2; Steiger’s z).

Functional connectivity

To assess age-related differences in functional connectivity, we conducted separate partial correlations between age and within- or between-network connectivity, while covarying for scanner setup and head motion and applying FDR correction for comparisons across seven networks. For within-network connectivity, results indicated that increased age was significantly associated with lower connectivity in each of the individual networks (Fig. 4). However, there were no significant associations between age and between-network connectivity for any of the individual networks, r ≤ 0.120, P_FDR ≥ 0.562.

Based on prior reports of larger age-related differences in graph theoretical measures of functional connectivity for association relative to primary sensory networks,^32,35 we created an average measure of within-network connectivity for association networks (frontoparietal, default mode, dorsal attention and ventral attention) and compared age-related differences in this measure to age-related differences within-network connectivity for the visual and sensorimotor networks. We used the same two-tailed Steiger’s z approach as above and applied FDR correction for two comparisons. Results indicated that the age-related difference was significantly larger for the association networks relative to the sensorimotor network, z = 2.226, P_FDR = 0.046, but was not significantly different from the age-related difference in the visual network, z = 1.664, P_FDR = 0.096.

Structural versus functional connectivity

Next, we sought to test whether structural connectivity mediated age-related differences in functional connectivity. We limited these analyses to measures of within-network connectivity that were significantly related to age for the functional, high-resolution structural and standard resolution structural data (i.e. visual, default mode, and sub-cortical). For each of the three networks, we constructed mediation models with age as the predictor variable, within-network functional connectivity as the outcome variable, and the high-resolution within-network structural connectivity measure and the corresponding standard resolution version of it as parallel mediators. This parallel approach provides a direct comparison of the mediation effects for the measures derived from the two DWI acquisitions because the measures are covaried for each other.

Results indicated that high-resolution structural connectivity within the default mode network was a significant mediator of age-related decreases in functional connectivity within the default mode network (Table 3 and Supplementary Fig. 1). The beta coefficient for the indirect effect was negative (β = −0.0003), reflecting the product of one negative and one positive effect: the decrease in functional connectivity with increasing age (the a path), and the increase in within-network functional connectivity with increasing within-network structural connectivity (the b path). Thus, age influenced the decrease in default mode functional connectivity indirectly, through the decrease in structural connectivity, which was detectable with the higher resolution MUSE sequence.

Table 3.

Mediation of default mode functional connectivity by structural connectivity

	Effect	SE	t	P	Lower CI	Upper CI
Age effect (a path)
High-resolution	−0.0001	0.00003	−4.041	< 0.001	−0.00021	−0.00007
Standard resolution	−0.0001	0.00004	−2.921	0.004	−0.00017	−0.00003
Mediator to outcome (b path)
High-resolution	2.173	0.945	2.300	0.023	0.305	4.042
Standard resolution	−1.836	0.938	−1.957	0.052	−3.692	0.020
Total effect for age (c path)
	−0.0009	0.0003	−2.904	0.004	−0.0015	−0.0002
Direct effect for age (c′ path)
	−0.0008	0.0003	−2.407	0.017	−0.0014	−0.0001
Mediation effect (a × b path interaction)
Total	−0.0001	0.00012	−	−	−0.00037	0.00010
High-resolution	−0.0003	0.00016	−	−	−0.00064	−0.00004
Standard resolution	0.0002	0.00011	−	−	−0.000001	0.00044

Open in a new tab

Parallel mediation models with age as the predictor variable (x), functional connectivity within the default mode network as the outcome variable (y), and high-resolution and standard resolution structural connectivity within the default mode network as parallel mediators (m); a = predictor to mediator pathway; b = mediator to outcome pathway, while controlling for a path; c = total effect of predictor; c’ = direct effect of predictor, controlling for mediators; a x b = interaction of a and b paths representing indirect influence of x on y as mediated by m; effect = unstandardized beta coefficient; Lower/Upper CI = lower/upper bounds of bias-corrected 95% confidence intervals. Significant effects are presented in bold.

Connectivity versus cognition

Finally, we tested whether the connectivity measures mediated age-related differences in cognitive performance, specifically focusing on measures that were significantly correlated with age. For between-network connectivity, these were structural connectivity for each of the seven individual networks. For within connectivity, the relevant measures were structural connectivity from all networks except for the frontoparietal network, and functional connectivity from all seven networks. Similarly, we focused on the general fluid cognition factor as the outcome measure as it exhibited the largest correlation with age.

For between-network connectivity, we constructed separate mediation models for each of the seven networks with general fluid cognition performance as the outcome variable, high-resolution and standard resolution structural connectivity as parallel mediators, and scanner setup as a covariate of no interest. Results indicated that that the individual indirect effects for the high-resolution and standard resolution connectivity measures were not significant for any network (Table 4 and Supplementary Figs 2–4). However, the total indirect effect was significant for the models using structural connectivity from the sensorimotor, dorsal attention, or frontoparietal networks as mediators, indicating that the combined effect of the standard resolution and high-resolution connectivity measures significantly mediated age-related differences in general fluid cognition (Table 4). The individual or total indirect effects were not significant for the other four networks.

Table 4.

Mediation of fluid cognition by between-network structural connectivity

	Effect	SE	t	P	Lower CI	Upper CI
Model 1: Sensorimotor between-network connectivity
Age effect (a path)
High-resolution	−0.00008	0.00001	−7.647	< 0.001	−0.00010	−0.00006
Standard resolution	−0.00004	0.00001	− 3.844	< 0.001	−0.00006	−0.00002
Mediator to outcome (b path)
High-resolution	32.181	23.109	1.393	0.166	−13.522	77.884
Standard resolution	40.707	23.702	1.717	0.088	−6.169	87.582
Total effect for age (c path)
	−0.039	0.003	−14.325	< 0.001	−0.044	−0.033
Direct effect for age (c’ path)
	−0.035	0.003	−10.810	< 0.001	−0.041	−0.028
Mediation effect (a x b path interaction)
Total	−0.004	0.002	−	−	−0.0080	−0.0006
High-resolution	−0.003	0.002	−	−	−0.0067	0.0011
Standard resolution	−0.002	0.001	−	−	−0.0036	0.0002
Model 2: Dorsal attention between-network connectivity
Age effect (a path)
High-resolution	−0.00007	0.00001	−5.942	< 0.001	−0.00010	−0.00005
Standard resolution	−0.00006	0.00001	−4.399	< 0.001	−0.00008	−0.00003
Mediator to outcome (b path)
High-resolution	31.355	21.795	1.439	0.153	−11.749	74.459
Standard resolution	28.107	20.471	1.373	0.172	−12.377	68.592
Total effect for age (c path)
	−0.039	0.003	−14.325	< 0.001	−0.044	−0.033
Direct effect for age (c’ path)
	−0.035	0.003	−11.680	< 0.001	−0.041	−0.029
Mediation effect (a x b path interaction)
Total	−0.0039	0.0016	−	−	−0.0073	−0.0009
High-resolution	−0.0023	0.0017	−	−	−0.0059	0.0007
Standard resolution	−0.0016	0.0012	−	–	−0.0041	0.0005
Model 3: Frontoparietal between-network connectivity
Age effect (a path)
High-resolution	−0.00007	0.00001	−7.100	< 0.001	−0.00009	−0.00005
Standard resolution	−0.00005	0.00001	−5.348	< 0.001	−0.00007	−0.00003
Mediator to outcome (b path)
High-resolution	37.148	24.059	1.544	0.125	−10.434	84.729
Standard resolution	37.659	24.600	1.531	0.128	−10.992	86.309
Total effect for age (c path)
	−0.039	0.003	−14.325	< 0.001	−0.044	−0.033
Direct effect for age (c’ path)
	−0.034	0.003	−10.747	< 0.001	−0.040	−0.028
Mediation effect (a x b path interaction)
Total	−0.005	0.002	–	–	−0.0092	−0.0010
High-resolution	−0.003	0.002	–	–	−0.0070	0.0007
Standard resolution	−0.002	0.001	–	–	−0.0048	0.0004

Open in a new tab

Mediation models with age as the predictor variable (x), fluid cognition as the outcome variable (y), and high-resolution and standard DWI between-connectivity measures as parallel mediators (m); a = predictor to mediator pathway; b = mediator to outcome pathway, while controlling for a path; c = total effect of predictor; c’ = direct effect of predictor, controlling for mediators; a x b = interaction of a and b paths representing indirect influence of x on y as mediated by m; effect = unstandardized beta coefficient; Lower/Upper CI = lower/upper bounds of bias-corrected 95% confidence intervals. Significant effects are presented in bold.

For within-network connectivity, we constructed similar mediation models with standard and high-resolution structural connectivity, plus resting-state functional connectivity, as parallel mediators, with resting-state motion as an additional covariate. Results indicated a significant individual indirect effect for high-resolution structural connectivity from the sensorimotor network, indicating that age-related decline in structural connectivity within the sensorimotor network, detected by the high-resolution acquisition, mediated age-related differences in fluid cognition (Table 5 and Supplementary Fig. 5). The total indirect effect from the sensorimotor model was also significant, reflecting the combined contribution of the structural and functional connectivity measures to age-related differences in fluid cognition. The total or individual indirect effects were not significant for the remaining models assessed.

Table 5.

Mediation of fluid cognition by within-network sensorimotor connectivity

	Effect	SE	t	P	Lower CI	Upper CI
Age effect (a path)
High-resolution	−0.0005	0.0001	−6.000	< 0.001	−0.00068	−0.00034
Standard resolution	−0.0002	0.0001	−2.182	0.031	−0.00042	−0.00002
Resting-state	−0.0011	0.0005	−2.070	0.040	−0.00212	−0.00005
Mediator to outcome (b path)
High-resolution	6.568	3.026	2.171	0.032	0.583	12.552
Standard resolution	3.502	2.542	1.378	0.171	−1.525	8.530
Resting-state	−0.203	0.443	−0.457	0.648	−1.080	0.675
Total effect for age (c path)
	−0.038	0.003	−13.613	< 0.001	−0.044	−0.032
Direct effect for age (c′ path)
	−0.034	0.003	−10.984	< 0.001	−0.040	−0.028
Mediation effect (a × b path interaction)
Total	−0.0039	0.0017	–	–	−0.0077	−0.0009
High-resolution	−0.0034	0.0016	–	–	−0.0069	−0.0006
Standard resolution	−0.0008	0.0008	–	–	−0.0028	0.0003
Resting-state	0.0002	0.0005	–	–	−0.0008	0.0013

Open in a new tab

Mediation models with age as the predictor variable (x), fluid cognition as the outcome variable (y), and high-resolution DWI, standard resolution DWI, and resting-state within-connectivity measures from the sensorimotor network as parallel mediators (m); a = predictor to mediator pathway; b = mediator to outcome pathway, while controlling for a path; c = total effect of predictor; c′ = direct effect of predictor, controlling for mediators; a × b = interaction of a and b paths representing indirect influence of x on y as mediated by m; effect = unstandardized beta coefficient; Lower/Upper CI = lower/upper bounds of bias-corrected 95% confidence intervals. Significant effects are presented in bold.

Discussion

Previous studies report that higher spatial resolution DWI better estimates the structural connectome relative standard resolution DWI. We translated these findings to healthy neurocognitive aging across the adult lifespan using a high-resolution DWI protocol that achieved a 3-fold increase in spatial resolution relative to a standard DWI protocol. We found that high-resolution DWI identified larger age-related decreases in WM structural connectivity than would be detected using standard resolution DWI alone. High-resolution structural connectivity also helped better explain age-related decreases in functional connectivity and fluid cognition, when directly compared to standard resolution DWI. These findings illustrate the utility of this high-resolution MRI protocol for detecting age-related alterations in brain function and cognitive performance, which will help set the stage for future studies wishing to capitalize on this methodology for further translational aging research.

Structural connectivity and healthy aging

Increasing adult age was significantly associated with decreased WM structural connectivity across cortical and sub-cortical networks when assessed by both DWI acquisitions. When assessed by both acquisitions, age-related decreases in between-network connectivity were significant for all networks, but the difference in within-network connectivity was only present for the visual, default mode, and sub-cortical networks. However, the high-resolution acquisition identified significantly larger age-related differences in connectivity within the sensorimotor, ventral attention and sub-cortical networks, and for sensorimotor between-network connectivity. Finding larger, more widespread age-related differences in between- than within-network structural connectivity replicates prior graph theoretical reports.^17,85,86 Extending this work, our findings suggest that age-related differences in structural connectivity may be better detected by high-resolution relative to standard resolution DWI acquisitions, in line with studies of animal models and healthy younger adults.^24-26

Connectivity measures between the two acquisitions were highly correlated, similar to our previous study using a subset of these data (n = 61) that did not assess relations to functional connectivity or age-related differences in structural connectivity into the ninth decade of life.³¹ Interestingly, our prior study observed larger age-related correlations for structural connectivity when assessed by standard relative to high-resolution DWI, but the current study instead found larger age-related correlations for a few networks when assessed by high-resolution DWI. This divergence could be attributable to several potential differences, including the larger sample size, the combination of single-band and multi-band high-resolution data (the latter of which also had a greater number of diffusion-weighted directions and therefore, higher angular resolution), and the inclusion of adults beyond 80 years of age.

Structural connectivity and its relation to functional connectivity

Beyond structural connectivity, we also identified significant age-related differences in within-network functional connectivity for all networks, whereas between-network connectivity was not significantly related to age for any network. This finding replicates and extends prior work^17,32 by demonstrating that this pattern persists even when including ‘oldest-old’ adults beyond 80 years of age, who may exhibit more magnified effects of age on brain function.³³ Although some studies report age-related increases in between-network functional connectivity,^{18,21,35,87,88} our current finding from a high-functioning sample of community-dwelling volunteers instead supports recent evidence suggesting that the pattern of increased between-network connectivity may be more characteristic of pathological aging, such as Alzheimer’s disease,⁸⁹ rather than healthy aging. Between-network connectivity, by definition, averages connectivity strength between nodes from one network and nodes outside of that network. As such, it is possible that age-related increases in connectivity between one pair of networks and age-related decreases connectivity between another pair of networks results in a net zero effect of age on between-network connectivity, as observed here. This possibility could be tested in future studies disaggregating the pairwise associations of connectivity between each combination of network. Some studies further report larger age-related effects on functional connectivity for association relative to sensory networks,^32,35,89 a pattern that was replicated here for association networks relative to the sensorimotor network.

Theoretical accounts would propose that the larger age-related alterations in association network functional connectivity should relate to the underlying structural architecture of association networks,^12,13 which also exhibit relatively greater degradation than primary sensory cortices in healthy aging.^90,91 However, empirical studies report mixed results on the association between measures of brain structure and function in adults across the lifespan.^14-16 We suspected that these mixed findings may partly reflect the lower spatial resolution of standard DWI acquisitions limiting the ability to assess connectivity in fine-grained WM regions.^92,93 We found that age-related differences in functional connectivity within the DMN (an association network) were significantly mediated by structural connectivity within the DMN, but only when assessed by high-resolution DWI. This finding suggests that the contribution of structural degradation to age-related alterations in brain function may be better detected by high-resolution DWI, and that measurement error in standard DWI may attenuate the ability to detect significant associations between WM connectivity and fMRI measures of brain function. This finding extends our previous work in a subset of the current sample that was limited only to analyses of structural connectivity, and to our knowledge, represents the first direct comparison between two different DWI measures of structural connectivity and their relation to functional connectivity.

Interestingly, the DMN comprises widely distributed anterior (e.g. superior and inferior frontal cortex) and posterior (e.g. precuneus, inferior parietal lobule) sub-regions that are uniquely impacted by Alzheimer’s-disease neuropathologies (e.g. amyloid burden), even years before the manifestation of cognitive impairment.^94,95 However, posterior sub-regions of this network are among the first to accumulate amyloid and have been linked to aberrant functional connectivity even in healthy adults.⁹⁶ The network partition employed here combined anterior and posterior sub-divisions of the default mode network²⁹ and biomarker measures of neuropathologies were not collected from the current sample. Although the current sample was well-characterized with respect to general cognitive function (e.g. more than 12 psychometric tests in addition to standard tests of general cognition and vocabulary), it is important for future work to repeat these analyses in a sample with measures of biomarker status (e.g. amyloid and tau positivity) so that the effect of neuropathologies on functional and structural connectivity can be directly assessed, especially within the distinct anatomical sub-divisions of the DMN.

Previous studies have further demonstrated the utility of graph theory for assessing relations between structural and functional connectivity.^17,18,20,28 Our current results are in line with prior suggestions that structural and functional connectivity may be relatively independent,^17,20 and that these relations may specifically be network dependent. Studies of healthy younger adults report stronger relations between structural and functional connectivity in sensory than association networks,³⁴ but the current finding is in line with prior studies reporting stronger relations for association than sensory networks in healthy aging.^36,97 Together, these findings suggest that DMN functional connectivity is related to underlying age-related degradations in structural connectivity, which may be better detected by high-resolution DWI.

Structural connectivity, functional connectivity, and fluid cognition

The age-related decrease in structural connectivity mediated age-related decline in general fluid cognition, a measure that exhibited a strong negative relation with age (with some degree of unique decline in memory and perceptual-motor speed). Parallel mediation analyses indicated that the combination of high-resolution and standard resolution DWI measures of between-network connectivity from sensorimotor, dorsal attention and frontoparietal networks helped explain age-related decreases in fluid cognition, together supporting the notion of cortical disconnection as a mechanism underlying normal age-related cognitive decline.^6,8-10 Importantly, high-resolution structural connectivity within the sensorimotor network was a significant individual mediator of age-related differences in fluid cognition, and this network also exhibited a significantly larger age-related correlation between assessed by high-resolution than standard DWI. However, a significant total effect for this sensorimotor model indicated that standard resolution and functional within-network connectivity also exerted some degree of combined influence on age-related differences in cognitive performance. Together, these findings suggest that WM plays a crucial role in fluid cognition across the adult lifespan, and that high-resolution DWI may better detect this relation, at least for measures of within-network connectivity that likely comprise smaller, more fine-grained fibres.⁹³

Limitations and future research

This study is strengthened by the inclusion of adults across the lifespan, including adults beyond 80 years of age who are underrepresented in MRI studies of brain and cognitive aging. The high-resolution DWI acquisitions used here hold promise for bridging clinical and translational research as it requires a substantially shorter scan time than most high-resolution protocols but can nonetheless be acquired from animal models. However, comparisons between the standard and high-resolution data here may be confounded by the higher b-value in the standard data may better capture hindered and restricted sources of diffusion,^51,52 which is why we excluded standard resolution data from the b = 3000 s/mm² shell. Our analyses are also limited by the change in scanner setup that occurred during data collection. Although we included scanner setup as a covariate of no interest, we cannot fully rule out its impact on the current results. This change is scanner setup was also highly confounded with the upgrade from the single-band to multi-band high-resolution DWI data acquisition. Ninety-two percent of participants with single-band data were also scanned using the original scanner setup and 64% of participants with multi-band data were scanned using the upgraded scanner setup. By including scanner setup as a covariate of no interest in all analyses, this feature of the study also provides some confidence that the improved ability of high-resolution DWI to explain age-related differences in functional connectivity and fluid cognition cannot be solely attributed to differences between the high-resolution protocols. This feature does, however, limit our ability to disentangle potential differences between the two protocols here, representing an important direction for future studies in this line of work. Lastly, the present analyses were limited to observations from a single point in time and must be replicated in future work employing longitudinal designs to ensure that these patterns reflect true age-related effects instead of factors such as cohort effects.^98,99

Conclusions

Higher resolution MRI protocols may better capture the complex intricacies of neuroanatomy in vivo, but often at the expense of extensive scanning time. Here, we demonstrate the utility of a clinically feasible high-resolution DWI acquisition for understanding interactions among brain structure, brain function, and cognition in healthy aging. Compared to standard resolution DWI, we found that high-resolution structural connectivity better explained age-related variance in functional connectivity within the DMN and improved the ability to explain age-related variance in fluid cognition. Our results should inform future intervention work aimed at identifying adults who are at heightened risk for cognitive decline and should be informative for future studies that may wish to use this high-resolution DWI protocol in both human and animal models of neurocognitive aging.

Supplementary Material

fcaf376_Supplementary_Data

fcaf376_supplementary_data.docx^{(384.5KB, docx)}

Acknowledgements

We thank Amalia Desir, Jiayi Zhao and Hollie Mullin for their assistance.

Contributor Information

Jenna L Merenstein, Department of Psychology, University of Utah, Salt Lake City, UT 84112, USA; Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC 27710, USA.

Allen W Song, Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC 27710, USA; Department of Radiology, Duke University Medical Center, Durham, NC 27710, USA.

David J Madden, Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC 27710, USA; Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC 27710, USA.

Supplementary material

Supplementary material is available at Brain Communications online.

Funding

This work was supported by R01 AG039684, P30 AG072958 and F32 AG084167 from the National Institute on Aging.

Competing interests

The authors declare no competing financial interests.

Data availability

Because the community-dwelling adult participants did not consent to sharing their data, we are unable to make the raw data presented in this paper freely available online. The code that was used for data analyses is available on the Open Science Framework at https://osf.io/ck7nd/.

References

1. Salthouse TA. The processing-speed theory of adult age differences in cognition. Psychol Rev. 1996;103(3):403–428. [DOI] [PubMed] [Google Scholar]
2. Craik FIM, Bialystok E. Cognition through the lifespan: Mechanisms of change. Trends Cogn Sci. 2006;10(3):131–138. [DOI] [PubMed] [Google Scholar]
3. Madden DJ. Speed and timing of behavioral processes, eds. Handbook of the psychology of aging. 5 ed. Academic Press; 2001:288–312. [Google Scholar]
4. Salthouse TA. Speed of behavior and its implications for cognition. In: Birren JE, Schaie KW, eds. Handbook of the psychology of aging. 2nd edn. Van Nostrand Reinhold; 1985:400–426. [Google Scholar]
5. Madden DJ. Aging and visual attention. Curr Dir Psychol Sci. 2007;16(2):70–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
6. O'Sullivan M, Jones DK, Summers PE, Morris RG, Williams SCR, Markus HS. Evidence for cortical “disconnection” as a mechanism of age-related cognitive decline. Neurology. 2001;57(4):632–638. [DOI] [PubMed] [Google Scholar]
7. Salat DH. The declining infrastructure of the aging brain. Brain Connect. 2011;1(4):279–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Bennett IJ, Madden DJ. Disconnected aging: Cerebral white matter integrity and age-related differences in cognition. Neuroscience. 2014;276:187–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Madden DJ, Parks EL, Tallman CW, et al. Sources of disconnection in neurocognitive aging: Cerebral white-matter integrity, resting-state functional connectivity, and white-matter hyperintensity volume. Neurobiol Aging. 2017;54:199–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Bartzokis G. Age-related myelin breakdown: A developmental model of cognitive decline and Alzheimer's disease. Neurobiol Aging. 2004;25(1):5–18. [DOI] [PubMed] [Google Scholar]
11. Waxman SG. Determinants of conduction velocity in myelinated nerve fibers. Muscle Nerve. 1980;3(2):141–150. [DOI] [PubMed] [Google Scholar]
12. Honey CJ, Thivierge JP, Sporns O. Can structure predict function in the human brain? Neuroimage. 2010;52(3):766–776. [DOI] [PubMed] [Google Scholar]
13. Sepulcre J, Liu H, Talukdar T, Martincorena I, Yeo BTT, Buckner RL. The organization of local and distant functional connectivity in the human brain. PLoS Comput Biol. 2010;6(6):e1000808–e1000808. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Bennett IJ, Rypma B. Advances in functional neuroanatomy: A review of combined DTI and fMRI studies in healthy younger and older adults. Neurosci Biobehav Rev. 2013;37(7):1201–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Warbrick T, Rosenberg J, Shah NJ. The relationship between BOLD fMRI response and the underlying white matter as measured by fractional anisotropy (FA): A systematic review. Neuroimage. 2017;153:369–381. [DOI] [PubMed] [Google Scholar]
16. Fjell AM, Sneve MH, Grydeland H, et al. Relationship between structural and functional connectivity change across the adult lifespan: A longitudinal investigation. Hum Brain Mapp. 2017;38(1):561–573. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Madden DJ, Merenstein JL, Mullin HA, Jain S, Rudolph MD, Cohen JR. Age-related differences in resting-state, task-related, and structural brain connectivity: Graph theoretical analyses and visual search performance. Brain Struct Funct. 2024;229(7):1533–1559. [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Betzel RF, Byrge L, He Y, Goñi J, Zuo X-N, Sporns O. Changes in structural and functional connectivity among resting-state networks across the human lifespan. Neuroimage. 2014;102:345–357. [DOI] [PubMed] [Google Scholar]
19. Zamani Esfahlani F, Faskowitz J, Slack J, Mišić B, Betzel RF. Local structure-function relationships in human brain networks across the lifespan. Nat Commun. 2022;13(1):2053. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Madden DJ, Jain S, Monge ZA, et al. Influence of structural and functional brain connectivity on age-related differences in fluid cognition. Neurobiol Aging. 2020;96(1):205–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Pedersen R, Geerligs L, Andersson M, et al. When functional blurring becomes deleterious: Reduced system segregation is associated with less white matter integrity and cognitive decline in aging. Neuroimage. 2021;242:118449–118449. [DOI] [PubMed] [Google Scholar]
22. Burzynska AZ, Garrett DD, Preuschhof C, et al. A scaffold for efficiency in the human brain. J Neurosci. 2013;33(43):17150–17159. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Webb CE, Rodrigue KM, Hoagey DA, Foster CM, Kennedy KM. Contributions of white matter connectivity and BOLD modulation to cognitive aging: A lifespan structure-function association study. Cereb Cortex. 2020;30(3):1649–1661. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Crater S, Maharjan S, Qi Y, et al. Resolution and b value dependent structural connectome in ex vivo mouse brain. Neuroimage. 2022;255:119199. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Anderson RJ, Long CM, Calabrese ED, et al. Optimizing diffusion imaging protocols for structural connectomics in mouse models of neurological conditions. Front Phys. 2020;8:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Chang H-C, Sundman M, Petit L, et al. Human brain diffusion tensor imaging at submillimeter isotropic resolution on a 3 Tesla clinical MRI scanner. Neuroimage. 2015;118:667–675. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Rubinov M, Sporns O. Complex network measures of brain connectivity: Uses and interpretations. Neuroimage. 2010;52(3):1059–1069. [DOI] [PubMed] [Google Scholar]
28. Wu D, Li X. Graph propagation network captures individual specificity of the relationship between functional and structural connectivity. Hum Brain Mapp. 2023;44(9):3885–3896. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Yeo BT, Krienen FM, Sepulcre J, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Fan L, Li H, Zhuo J, et al. The human brainnetome atlas: A new brain atlas based on connectional architecture. Cereb Cortex. 2016;26(8):3508–3526. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Merenstein JL, Zhao J, Mullin HA, Rudolph MD, Song AW, Madden DJ. High-resolution multi-shot diffusion imaging of structural networks in healthy neurocognitive aging. Neuroimage. 2023;275(1):120191. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Deery HA, Di Paolo R, Moran C, Egan GF, Jamadar SD. The older adult brain is less modular, more integrated, and less efficient at rest: A systematic review of large-scale resting-state functional brain networks in aging. Psychophysiology. 2023;60(1):e14159. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Merenstein JL, Bennett IJ. Bridging patterns of neurocognitive aging across the older adult lifespan. Neurosci Biobehav Rev. 2022;135(1):104594. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Gu Z, Jamison KW, Sabuncu MR, Kuceyeski A. Heritability and interindividual variability of regional structure-function coupling. Nat Commun. 2021;12(1):4894. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Chan MY, Park DC, Savalia NK, Petersen SE, Wig GS. Decreased segregation of brain systems across the healthy adult lifespan. Proc Natl Acad Sci U S A. 2014;111(46):E4997–E5006. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Billaud CHA, Yu J; Alzheimer’s Disease Neuroimaging Initiative . Structure-function coupling using fixel-based analysis and functional magnetic resonance imaging in Alzheimer’s disease and mild cognitive impairment. Netw Neurosci. 2025;9:1–17.40161990 [Google Scholar]
37. Christensen KJ, Moye J, Armson RR, Kern TM. Health screening and random recruitment for cognitive aging research. Psychol Aging. 1992;7(2):204–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Folstein MF, Folstein SE, McHugh PR. Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–198. [DOI] [PubMed] [Google Scholar]
39. Nasreddine ZS, Phillips NA, Bédirian V, et al. The Montreal cognitive assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699. [DOI] [PubMed] [Google Scholar]
40. Wechsler D. WAIS-III: Wechsler adult intelligence scale, eds. Administration and scoring manual. 3rd ed. The Psychological Corporation; 1997:217–217. [Google Scholar]
41. Beck AT. The beck depression inventory. Psychological Corporation; 1978. [Google Scholar]
42. Bach M. The Freiburg visual acuity test—Automatic measurement of visual acuity. Optom Vis Sci. 1996;73(1):49–53. [DOI] [PubMed] [Google Scholar]
43. Dvorine I. Dvorine pseudo-isochromatic plates. 2nd ed. Psychological Corporation (a Harcourt Company); 1963. [Google Scholar]
44. Howard CM, Jain S, Cook AD, et al. Cortical iron mediates age-related decline in fluid cognition. Hum Brain Mapp. 2022;43(3):1047–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Merenstein JL, Mullin HA, Madden DJ. Age-related differences in frontoparietal activation for target and distractor singletons during visual search. Atten Percept Psychophys. 2023;85(3):749–768. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Merenstein JL, Zhao J, Madden DJ. Depthwise cortical iron relates to functional connectivity and fluid cognition in healthy aging. Neurobiol Aging. 2025;148(1):27–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Gershon RC, Wagster MV, Hendrie HC, Fox NA, Cook KF, Nowinski CJ. NIH toolbox for assessment of neurological and behavioral function. Neurology. 2013;80(11 Suppl 3):S2–S6. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Reitan RM. Trail making test results for normal and brain-damaged children. Percept Mot Skills. 1971;33(2):575–581. [DOI] [PubMed] [Google Scholar]
49. Saults JS, Cowan N. A central capacity limit to the simultaneous storage of visual and auditory arrays in working memory. J Exp Psychol Gen. 2007;136(4):663–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Rudolph MD, Cohen JR, Madden DJ. Distributed associations among white matter hyperintensities and structural brain networks with fluid cognition in healthy aging. Cogn Affect Behav Neurosci. 2024;24(6):1121–1140. [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Assaf Y, Freidlin RZ, Rohde GK, Basser PJ. New modeling and experimental framework to characterize hindered and restricted water diffusion in brain white matter. Magn Reson Med. 2004;52(5):965–978. [DOI] [PubMed] [Google Scholar]
52. Assaf Y, Basser PJ. Composite hindered and restricted model of diffusion (CHARMED) MR imaging of the human brain. Neuroimage. 2005;27(1):48–58. [DOI] [PubMed] [Google Scholar]
53. Tournier J, Smith R, Raffelt D, et al. MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation. Neuroimage. 2019;202:116137. [DOI] [PubMed] [Google Scholar]
54. Chen NK, Guidon A, Chang HC, Song AW. A robust multi-shot scan strategy for high-resolution diffusion weighted MRI enabled by multiplexed sensitivity-encoding (MUSE). Neuroimage. 2013;72(1):41–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Dhollander T, Mito R, Raffelt D, Connelly A. Improved white matter response function estimation for 3-tissue constrained spherical deconvolution. 2019.
56. Tournier JD, Calamante F, Gadian DG, Connelly A. Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution. Neuroimage. 2004;23(3):1176–1185. [DOI] [PubMed] [Google Scholar]
57. Jeurissen B, Tournier JD, Dhollander T, Connelly A, Sijbers J. Multi-tissue constrained spherical deconvolution for improved analysis of multi-shell diffusion MRI data. Neuroimage. 2014;103(1):411–426. [DOI] [PubMed] [Google Scholar]
58. Jenkinson M, Smith S. A global optimisation method for robust affine registration of brain images. Med Image Anal. 2001;5(2):143–156. [DOI] [PubMed] [Google Scholar]
59. Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage. 2002;17(2):825–841. [DOI] [PubMed] [Google Scholar]
60. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage. 2004;23(S1):S208–S219. [DOI] [PubMed] [Google Scholar]
61. Smith RE, Tournier J-D, Calamante F, Connelly A. Anatomically-constrained tractography: Improved diffusion MRI streamlines tractography through effective use of anatomical information. Neuroimage. 2012;62(3):1924–1938. [DOI] [PubMed] [Google Scholar]
62. Smith RE, Tournier JD, Calamante F, Connelly A. SIFT2: Enabling dense quantitative assessment of brain white matter connectivity using streamlines tractography. Neuroimage. 2015;119(1):338–351. [DOI] [PubMed] [Google Scholar]
63. Smith RE, Tournier JD, Calamante F, Connelly A. SIFT: Spherical-deconvolution informed filtering of tractograms. Neuroimage. 2013;67(1):298–312. [DOI] [PubMed] [Google Scholar]
64. Esteban O, Markiewicz CJ, Blair RW, et al. fMRIPrep: A robust preprocessing pipeline for functional MRI. Nat Methods. 2019;16(1):111–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
65. Alakörkkö T, Saarimäki H, Glerean E, Saramäki J, Korhonen O. Effects of spatial smoothing on functional brain networks. Eur J Neurosci. 2017;46(9):2471–2480. [DOI] [PMC free article] [PubMed] [Google Scholar]
66. Greve DN, Fischl B. Accurate and robust brain image alignment using boundary-based registration. Neuroimage. 2009;48(1):63–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
67. Cox RW. AFNI: Software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res. 1996;29(3):162–173. [DOI] [PubMed] [Google Scholar]
68. Satterthwaite TD, Elliott MA, Gerraty RT, et al. An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data. Neuroimage. 2013;64:240–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
69. Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage. 2012;59(3):2142–2154. [DOI] [PMC free article] [PubMed] [Google Scholar]
70. Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage. 2014;84:320–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
71. Ciric R, Wolf DH, Power JD, et al. Benchmarking of participant-level confound regression strategies for the control of motion artifact in studies of functional connectivity. Neuroimage. 2017;154:174–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
72. Parkes L, Fulcher B, Yücel M, Fornito A. An evaluation of the efficacy, reliability, and sensitivity of motion correction strategies for resting-state functional MRI. Neuroimage. 2018;171:415–436. [DOI] [PubMed] [Google Scholar]
73. Capogna E, Sneve MH, Raud L, et al. Whole-brain connectivity during encoding: Age-related differences and associations with cognitive and brain structural decline. Cereb Cortex. 2022;33(1):68–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
74. Luo N, Sui J, Abrol A, et al. Age-related structural and functional variations in 5,967 individuals across the adult lifespan. Hum Brain Mapp. 2020;41(7):1725–1737. [DOI] [PMC free article] [PubMed] [Google Scholar]
75. Girn M, Setton R, Turner GR, Spreng RN. The “limbic network,” comprising orbitofrontal and anterior temporal cortex, is part of an extended default network: Evidence from multi-echo fMRI. Netw Neurosci. 2024;8(3):860–882. [DOI] [PMC free article] [PubMed] [Google Scholar]
76. Civier O, Smith RE, Yeh C-H, Connelly A, Calamante F. Is removal of weak connections necessary for graph-theoretical analysis of dense weighted structural connectomes from diffusion MRI? Neuroimage. 2019;194:68–81. [DOI] [PubMed] [Google Scholar]
77. Hagmann P, Cammoun L, Gigandet X, et al. Mapping the structural core of human cerebral cortex. PLoS Biol. 2008;6(7):e159. [DOI] [PMC free article] [PubMed] [Google Scholar]
78. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J Roy Stat Soc Ser B (Stat Method). 1995;57(1):289–300. [Google Scholar]
79. Hedden T, Schultz AP, Rieckmann A, et al. Multiple brain markers are linked to age-related variation in cognition. Cereb Cortex. 2016;26(4):1388–1400. [DOI] [PMC free article] [PubMed] [Google Scholar]
80. Hedden T, Van Dijk KR, Shire EH, Sperling RA, Johnson KA, Buckner RL. Failure to modulate attentional control in advanced aging linked to white matter pathology. Cereb Cortex. 2012;22(5):1038–1051. [DOI] [PMC free article] [PubMed] [Google Scholar]
81. Salthouse TA, Habeck C, Razlighi Q, Barulli D, Gazes Y, Stern Y. Breadth and age-dependency of relations between cortical thickness and cognition. Neurobiol Aging. 2015;36(11):3020–3028. [DOI] [PMC free article] [PubMed] [Google Scholar]
82. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull. 1980;87(2):245–251. [Google Scholar]
83. Weiss BA. Hotelling's t Test and Steiger's Z test calculator. 2011. Accessed 8 February 2025. https://blogs.gwu.edu/weissba/teaching/calculators/hotellings-t-and-steigers-z-tests/
84. Hayes AF, Rockwood NJ. Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. Behav Res Ther. 2017;98(1):39–57. [DOI] [PubMed] [Google Scholar]
85. Wu K, Taki Y, Sato K, et al. Age-related changes in topological organization of structural brain networks in healthy individuals. Hum Brain Mapp. 2012;33(3):552–568. [DOI] [PMC free article] [PubMed] [Google Scholar]
86. Puxeddu MG, Faskowitz J, Betzel RF, Petti M, Astolfi L, Sporns O. The modular organization of brain cortical connectivity across the human lifespan. Neuroimage. 2020;218:116974. [DOI] [PubMed] [Google Scholar]
87. Liem F, Geerligs L, Damoiseaux JS, Margulies DS. Functional connectivity in aging. In: Schaie KW, Willis S, eds. Handbook of the psychology of Aging. 9th ed. Academic Press; 2019. [Google Scholar]
88. Damoiseaux JS. Effects of aging on functional and structural brain connectivity. Neuroimage. 2017;160:32–40. [DOI] [PubMed] [Google Scholar]
89. Zhang Z, Chan MY, Han L, et al. Dissociable effects of Alzheimer's disease-related cognitive dysfunction and aging on functional brain network segregation. J Neurosci. 2023;43(46):7879–7892. [DOI] [PMC free article] [PubMed] [Google Scholar]
90. Brickman AM, Meier IB, Korgaonkar MS, et al. Testing the white matter retrogenesis hypothesis of cognitive aging. Neurobiol Aging. 2012;33(8):1699–1715. [DOI] [PMC free article] [PubMed] [Google Scholar]
91. Hoagey DA, Rieck JR, Rodrigue KM, Kennedy KM. Joint contributions of cortical morphometry and white matter microstructure in healthy brain aging: A partial least squares correlation analysis. Hum Brain Mapp. 2019;40(18):5315–5329. [DOI] [PMC free article] [PubMed] [Google Scholar]
92. Metzler-Baddeley C, O'Sullivan MJ, Bells S, Pasternak O, Jones DK. How and how not to correct for CSF-contamination in diffusion MRI. Neuroimage. 2012;59(2):1394–1403. [DOI] [PubMed] [Google Scholar]
93. Guevara M, Guevara P, Román C, Mangin J-F. Superficial white matter: A review on the dMRI analysis methods and applications. Neuroimage. 2020;212:116673. [DOI] [PubMed] [Google Scholar]
94. Corriveau-Lecavalier N, Gunter JL, Kamykowski M, et al. Default mode network failure and neurodegeneration across aging and amnestic and dysexecutive Alzheimer's disease. Brain Commun. 2023;5(2):fcad058. [DOI] [PMC free article] [PubMed] [Google Scholar]
95. Han F, Liu X, Mailman RB, Huang X, Liu X. Resting-state global brain activity affects early β-amyloid accumulation in default mode network. Nat Commun. 2023;14(1):7788. [DOI] [PMC free article] [PubMed] [Google Scholar]
96. Ingala S, Tomassen J, Collij LE, et al. Amyloid-driven disruption of default mode network connectivity in cognitively healthy individuals. Brain Commun. 2021;3(4):fcab201. [DOI] [PMC free article] [PubMed] [Google Scholar]
97. Liu X, Tyler LK, Cam-Can, Davis SW, Rowe JB, Tsvetanov KA. Cognition’s dependence on functional network integrity with age is conditional on structural network integrity. Neurobiol Aging. 2023;129:195–208. [DOI] [PubMed] [Google Scholar]
98. Rönnlund M, Nyberg L, Bäckman L, Nilsson LG. Stability, growth, and decline in adult life span development of declarative memory: Cross-sectional and longitudinal data from a population-based study. Psychol Aging. 2005;20(1):3–18. [DOI] [PubMed] [Google Scholar]
99. Vidal-Pineiro D, Wang Y, Krogsrud SK, et al. Individual variations in ‘brain age’ relate to early-life factors more than to longitudinal brain change. Elife. 2021;10:e69995. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

fcaf376_Supplementary_Data

fcaf376_supplementary_data.docx^{(384.5KB, docx)}

Data Availability Statement

[fcaf376-B1] 1. Salthouse TA. The processing-speed theory of adult age differences in cognition. Psychol Rev. 1996;103(3):403–428. [DOI] [PubMed] [Google Scholar]

[fcaf376-B2] 2. Craik FIM, Bialystok E. Cognition through the lifespan: Mechanisms of change. Trends Cogn Sci. 2006;10(3):131–138. [DOI] [PubMed] [Google Scholar]

[fcaf376-B3] 3. Madden DJ. Speed and timing of behavioral processes, eds. Handbook of the psychology of aging. 5 ed. Academic Press; 2001:288–312. [Google Scholar]

[fcaf376-B4] 4. Salthouse TA. Speed of behavior and its implications for cognition. In: Birren JE, Schaie KW, eds. Handbook of the psychology of aging. 2nd edn. Van Nostrand Reinhold; 1985:400–426. [Google Scholar]

[fcaf376-B5] 5. Madden DJ. Aging and visual attention. Curr Dir Psychol Sci. 2007;16(2):70–74. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B6] 6. O'Sullivan M, Jones DK, Summers PE, Morris RG, Williams SCR, Markus HS. Evidence for cortical “disconnection” as a mechanism of age-related cognitive decline. Neurology. 2001;57(4):632–638. [DOI] [PubMed] [Google Scholar]

[fcaf376-B7] 7. Salat DH. The declining infrastructure of the aging brain. Brain Connect. 2011;1(4):279–293. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B8] 8. Bennett IJ, Madden DJ. Disconnected aging: Cerebral white matter integrity and age-related differences in cognition. Neuroscience. 2014;276:187–205. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B9] 9. Madden DJ, Parks EL, Tallman CW, et al. Sources of disconnection in neurocognitive aging: Cerebral white-matter integrity, resting-state functional connectivity, and white-matter hyperintensity volume. Neurobiol Aging. 2017;54:199–213. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B10] 10. Bartzokis G. Age-related myelin breakdown: A developmental model of cognitive decline and Alzheimer's disease. Neurobiol Aging. 2004;25(1):5–18. [DOI] [PubMed] [Google Scholar]

[fcaf376-B11] 11. Waxman SG. Determinants of conduction velocity in myelinated nerve fibers. Muscle Nerve. 1980;3(2):141–150. [DOI] [PubMed] [Google Scholar]

[fcaf376-B12] 12. Honey CJ, Thivierge JP, Sporns O. Can structure predict function in the human brain? Neuroimage. 2010;52(3):766–776. [DOI] [PubMed] [Google Scholar]

[fcaf376-B13] 13. Sepulcre J, Liu H, Talukdar T, Martincorena I, Yeo BTT, Buckner RL. The organization of local and distant functional connectivity in the human brain. PLoS Comput Biol. 2010;6(6):e1000808–e1000808. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B14] 14. Bennett IJ, Rypma B. Advances in functional neuroanatomy: A review of combined DTI and fMRI studies in healthy younger and older adults. Neurosci Biobehav Rev. 2013;37(7):1201–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B15] 15. Warbrick T, Rosenberg J, Shah NJ. The relationship between BOLD fMRI response and the underlying white matter as measured by fractional anisotropy (FA): A systematic review. Neuroimage. 2017;153:369–381. [DOI] [PubMed] [Google Scholar]

[fcaf376-B16] 16. Fjell AM, Sneve MH, Grydeland H, et al. Relationship between structural and functional connectivity change across the adult lifespan: A longitudinal investigation. Hum Brain Mapp. 2017;38(1):561–573. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B17] 17. Madden DJ, Merenstein JL, Mullin HA, Jain S, Rudolph MD, Cohen JR. Age-related differences in resting-state, task-related, and structural brain connectivity: Graph theoretical analyses and visual search performance. Brain Struct Funct. 2024;229(7):1533–1559. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B18] 18. Betzel RF, Byrge L, He Y, Goñi J, Zuo X-N, Sporns O. Changes in structural and functional connectivity among resting-state networks across the human lifespan. Neuroimage. 2014;102:345–357. [DOI] [PubMed] [Google Scholar]

[fcaf376-B19] 19. Zamani Esfahlani F, Faskowitz J, Slack J, Mišić B, Betzel RF. Local structure-function relationships in human brain networks across the lifespan. Nat Commun. 2022;13(1):2053. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B20] 20. Madden DJ, Jain S, Monge ZA, et al. Influence of structural and functional brain connectivity on age-related differences in fluid cognition. Neurobiol Aging. 2020;96(1):205–222. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B21] 21. Pedersen R, Geerligs L, Andersson M, et al. When functional blurring becomes deleterious: Reduced system segregation is associated with less white matter integrity and cognitive decline in aging. Neuroimage. 2021;242:118449–118449. [DOI] [PubMed] [Google Scholar]

[fcaf376-B22] 22. Burzynska AZ, Garrett DD, Preuschhof C, et al. A scaffold for efficiency in the human brain. J Neurosci. 2013;33(43):17150–17159. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B23] 23. Webb CE, Rodrigue KM, Hoagey DA, Foster CM, Kennedy KM. Contributions of white matter connectivity and BOLD modulation to cognitive aging: A lifespan structure-function association study. Cereb Cortex. 2020;30(3):1649–1661. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B24] 24. Crater S, Maharjan S, Qi Y, et al. Resolution and b value dependent structural connectome in ex vivo mouse brain. Neuroimage. 2022;255:119199. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B25] 25. Anderson RJ, Long CM, Calabrese ED, et al. Optimizing diffusion imaging protocols for structural connectomics in mouse models of neurological conditions. Front Phys. 2020;8:88. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B26] 26. Chang H-C, Sundman M, Petit L, et al. Human brain diffusion tensor imaging at submillimeter isotropic resolution on a 3 Tesla clinical MRI scanner. Neuroimage. 2015;118:667–675. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B27] 27. Rubinov M, Sporns O. Complex network measures of brain connectivity: Uses and interpretations. Neuroimage. 2010;52(3):1059–1069. [DOI] [PubMed] [Google Scholar]

[fcaf376-B28] 28. Wu D, Li X. Graph propagation network captures individual specificity of the relationship between functional and structural connectivity. Hum Brain Mapp. 2023;44(9):3885–3896. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B29] 29. Yeo BT, Krienen FM, Sepulcre J, et al. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol. 2011;106(3):1125–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B30] 30. Fan L, Li H, Zhuo J, et al. The human brainnetome atlas: A new brain atlas based on connectional architecture. Cereb Cortex. 2016;26(8):3508–3526. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B31] 31. Merenstein JL, Zhao J, Mullin HA, Rudolph MD, Song AW, Madden DJ. High-resolution multi-shot diffusion imaging of structural networks in healthy neurocognitive aging. Neuroimage. 2023;275(1):120191. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B32] 32. Deery HA, Di Paolo R, Moran C, Egan GF, Jamadar SD. The older adult brain is less modular, more integrated, and less efficient at rest: A systematic review of large-scale resting-state functional brain networks in aging. Psychophysiology. 2023;60(1):e14159. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B33] 33. Merenstein JL, Bennett IJ. Bridging patterns of neurocognitive aging across the older adult lifespan. Neurosci Biobehav Rev. 2022;135(1):104594. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B34] 34. Gu Z, Jamison KW, Sabuncu MR, Kuceyeski A. Heritability and interindividual variability of regional structure-function coupling. Nat Commun. 2021;12(1):4894. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B35] 35. Chan MY, Park DC, Savalia NK, Petersen SE, Wig GS. Decreased segregation of brain systems across the healthy adult lifespan. Proc Natl Acad Sci U S A. 2014;111(46):E4997–E5006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B36] 36. Billaud CHA, Yu J; Alzheimer’s Disease Neuroimaging Initiative . Structure-function coupling using fixel-based analysis and functional magnetic resonance imaging in Alzheimer’s disease and mild cognitive impairment. Netw Neurosci. 2025;9:1–17.40161990 [Google Scholar]

[fcaf376-B37] 37. Christensen KJ, Moye J, Armson RR, Kern TM. Health screening and random recruitment for cognitive aging research. Psychol Aging. 1992;7(2):204–208. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B38] 38. Folstein MF, Folstein SE, McHugh PR. Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–198. [DOI] [PubMed] [Google Scholar]

[fcaf376-B39] 39. Nasreddine ZS, Phillips NA, Bédirian V, et al. The Montreal cognitive assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699. [DOI] [PubMed] [Google Scholar]

[fcaf376-B40] 40. Wechsler D. WAIS-III: Wechsler adult intelligence scale, eds. Administration and scoring manual. 3rd ed. The Psychological Corporation; 1997:217–217. [Google Scholar]

[fcaf376-B41] 41. Beck AT. The beck depression inventory. Psychological Corporation; 1978. [Google Scholar]

[fcaf376-B42] 42. Bach M. The Freiburg visual acuity test—Automatic measurement of visual acuity. Optom Vis Sci. 1996;73(1):49–53. [DOI] [PubMed] [Google Scholar]

[fcaf376-B43] 43. Dvorine I. Dvorine pseudo-isochromatic plates. 2nd ed. Psychological Corporation (a Harcourt Company); 1963. [Google Scholar]

[fcaf376-B44] 44. Howard CM, Jain S, Cook AD, et al. Cortical iron mediates age-related decline in fluid cognition. Hum Brain Mapp. 2022;43(3):1047–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B45] 45. Merenstein JL, Mullin HA, Madden DJ. Age-related differences in frontoparietal activation for target and distractor singletons during visual search. Atten Percept Psychophys. 2023;85(3):749–768. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B46] 46. Merenstein JL, Zhao J, Madden DJ. Depthwise cortical iron relates to functional connectivity and fluid cognition in healthy aging. Neurobiol Aging. 2025;148(1):27–40. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B47] 47. Gershon RC, Wagster MV, Hendrie HC, Fox NA, Cook KF, Nowinski CJ. NIH toolbox for assessment of neurological and behavioral function. Neurology. 2013;80(11 Suppl 3):S2–S6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B48] 48. Reitan RM. Trail making test results for normal and brain-damaged children. Percept Mot Skills. 1971;33(2):575–581. [DOI] [PubMed] [Google Scholar]

[fcaf376-B49] 49. Saults JS, Cowan N. A central capacity limit to the simultaneous storage of visual and auditory arrays in working memory. J Exp Psychol Gen. 2007;136(4):663–684. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B50] 50. Rudolph MD, Cohen JR, Madden DJ. Distributed associations among white matter hyperintensities and structural brain networks with fluid cognition in healthy aging. Cogn Affect Behav Neurosci. 2024;24(6):1121–1140. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B51] 51. Assaf Y, Freidlin RZ, Rohde GK, Basser PJ. New modeling and experimental framework to characterize hindered and restricted water diffusion in brain white matter. Magn Reson Med. 2004;52(5):965–978. [DOI] [PubMed] [Google Scholar]

[fcaf376-B52] 52. Assaf Y, Basser PJ. Composite hindered and restricted model of diffusion (CHARMED) MR imaging of the human brain. Neuroimage. 2005;27(1):48–58. [DOI] [PubMed] [Google Scholar]

[fcaf376-B53] 53. Tournier J, Smith R, Raffelt D, et al. MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation. Neuroimage. 2019;202:116137. [DOI] [PubMed] [Google Scholar]

[fcaf376-B54] 54. Chen NK, Guidon A, Chang HC, Song AW. A robust multi-shot scan strategy for high-resolution diffusion weighted MRI enabled by multiplexed sensitivity-encoding (MUSE). Neuroimage. 2013;72(1):41–47. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B55] 55. Dhollander T, Mito R, Raffelt D, Connelly A. Improved white matter response function estimation for 3-tissue constrained spherical deconvolution. 2019.

[fcaf376-B56] 56. Tournier JD, Calamante F, Gadian DG, Connelly A. Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution. Neuroimage. 2004;23(3):1176–1185. [DOI] [PubMed] [Google Scholar]

[fcaf376-B57] 57. Jeurissen B, Tournier JD, Dhollander T, Connelly A, Sijbers J. Multi-tissue constrained spherical deconvolution for improved analysis of multi-shell diffusion MRI data. Neuroimage. 2014;103(1):411–426. [DOI] [PubMed] [Google Scholar]

[fcaf376-B58] 58. Jenkinson M, Smith S. A global optimisation method for robust affine registration of brain images. Med Image Anal. 2001;5(2):143–156. [DOI] [PubMed] [Google Scholar]

[fcaf376-B59] 59. Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage. 2002;17(2):825–841. [DOI] [PubMed] [Google Scholar]

[fcaf376-B60] 60. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage. 2004;23(S1):S208–S219. [DOI] [PubMed] [Google Scholar]

[fcaf376-B61] 61. Smith RE, Tournier J-D, Calamante F, Connelly A. Anatomically-constrained tractography: Improved diffusion MRI streamlines tractography through effective use of anatomical information. Neuroimage. 2012;62(3):1924–1938. [DOI] [PubMed] [Google Scholar]

[fcaf376-B62] 62. Smith RE, Tournier JD, Calamante F, Connelly A. SIFT2: Enabling dense quantitative assessment of brain white matter connectivity using streamlines tractography. Neuroimage. 2015;119(1):338–351. [DOI] [PubMed] [Google Scholar]

[fcaf376-B63] 63. Smith RE, Tournier JD, Calamante F, Connelly A. SIFT: Spherical-deconvolution informed filtering of tractograms. Neuroimage. 2013;67(1):298–312. [DOI] [PubMed] [Google Scholar]

[fcaf376-B64] 64. Esteban O, Markiewicz CJ, Blair RW, et al. fMRIPrep: A robust preprocessing pipeline for functional MRI. Nat Methods. 2019;16(1):111–116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B65] 65. Alakörkkö T, Saarimäki H, Glerean E, Saramäki J, Korhonen O. Effects of spatial smoothing on functional brain networks. Eur J Neurosci. 2017;46(9):2471–2480. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B66] 66. Greve DN, Fischl B. Accurate and robust brain image alignment using boundary-based registration. Neuroimage. 2009;48(1):63–63. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B67] 67. Cox RW. AFNI: Software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res. 1996;29(3):162–173. [DOI] [PubMed] [Google Scholar]

[fcaf376-B68] 68. Satterthwaite TD, Elliott MA, Gerraty RT, et al. An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data. Neuroimage. 2013;64:240–256. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B69] 69. Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage. 2012;59(3):2142–2154. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B70] 70. Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage. 2014;84:320–341. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B71] 71. Ciric R, Wolf DH, Power JD, et al. Benchmarking of participant-level confound regression strategies for the control of motion artifact in studies of functional connectivity. Neuroimage. 2017;154:174–187. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B72] 72. Parkes L, Fulcher B, Yücel M, Fornito A. An evaluation of the efficacy, reliability, and sensitivity of motion correction strategies for resting-state functional MRI. Neuroimage. 2018;171:415–436. [DOI] [PubMed] [Google Scholar]

[fcaf376-B73] 73. Capogna E, Sneve MH, Raud L, et al. Whole-brain connectivity during encoding: Age-related differences and associations with cognitive and brain structural decline. Cereb Cortex. 2022;33(1):68–82. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B74] 74. Luo N, Sui J, Abrol A, et al. Age-related structural and functional variations in 5,967 individuals across the adult lifespan. Hum Brain Mapp. 2020;41(7):1725–1737. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B75] 75. Girn M, Setton R, Turner GR, Spreng RN. The “limbic network,” comprising orbitofrontal and anterior temporal cortex, is part of an extended default network: Evidence from multi-echo fMRI. Netw Neurosci. 2024;8(3):860–882. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B76] 76. Civier O, Smith RE, Yeh C-H, Connelly A, Calamante F. Is removal of weak connections necessary for graph-theoretical analysis of dense weighted structural connectomes from diffusion MRI? Neuroimage. 2019;194:68–81. [DOI] [PubMed] [Google Scholar]

[fcaf376-B77] 77. Hagmann P, Cammoun L, Gigandet X, et al. Mapping the structural core of human cerebral cortex. PLoS Biol. 2008;6(7):e159. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B78] 78. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J Roy Stat Soc Ser B (Stat Method). 1995;57(1):289–300. [Google Scholar]

[fcaf376-B79] 79. Hedden T, Schultz AP, Rieckmann A, et al. Multiple brain markers are linked to age-related variation in cognition. Cereb Cortex. 2016;26(4):1388–1400. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B80] 80. Hedden T, Van Dijk KR, Shire EH, Sperling RA, Johnson KA, Buckner RL. Failure to modulate attentional control in advanced aging linked to white matter pathology. Cereb Cortex. 2012;22(5):1038–1051. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B81] 81. Salthouse TA, Habeck C, Razlighi Q, Barulli D, Gazes Y, Stern Y. Breadth and age-dependency of relations between cortical thickness and cognition. Neurobiol Aging. 2015;36(11):3020–3028. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B82] 82. Steiger JH. Tests for comparing elements of a correlation matrix. Psychol Bull. 1980;87(2):245–251. [Google Scholar]

[fcaf376-B83] 83. Weiss BA. Hotelling's t Test and Steiger's Z test calculator. 2011. Accessed 8 February 2025. https://blogs.gwu.edu/weissba/teaching/calculators/hotellings-t-and-steigers-z-tests/

[fcaf376-B84] 84. Hayes AF, Rockwood NJ. Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. Behav Res Ther. 2017;98(1):39–57. [DOI] [PubMed] [Google Scholar]

[fcaf376-B85] 85. Wu K, Taki Y, Sato K, et al. Age-related changes in topological organization of structural brain networks in healthy individuals. Hum Brain Mapp. 2012;33(3):552–568. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B86] 86. Puxeddu MG, Faskowitz J, Betzel RF, Petti M, Astolfi L, Sporns O. The modular organization of brain cortical connectivity across the human lifespan. Neuroimage. 2020;218:116974. [DOI] [PubMed] [Google Scholar]

[fcaf376-B87] 87. Liem F, Geerligs L, Damoiseaux JS, Margulies DS. Functional connectivity in aging. In: Schaie KW, Willis S, eds. Handbook of the psychology of Aging. 9th ed. Academic Press; 2019. [Google Scholar]

[fcaf376-B88] 88. Damoiseaux JS. Effects of aging on functional and structural brain connectivity. Neuroimage. 2017;160:32–40. [DOI] [PubMed] [Google Scholar]

[fcaf376-B89] 89. Zhang Z, Chan MY, Han L, et al. Dissociable effects of Alzheimer's disease-related cognitive dysfunction and aging on functional brain network segregation. J Neurosci. 2023;43(46):7879–7892. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B90] 90. Brickman AM, Meier IB, Korgaonkar MS, et al. Testing the white matter retrogenesis hypothesis of cognitive aging. Neurobiol Aging. 2012;33(8):1699–1715. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B91] 91. Hoagey DA, Rieck JR, Rodrigue KM, Kennedy KM. Joint contributions of cortical morphometry and white matter microstructure in healthy brain aging: A partial least squares correlation analysis. Hum Brain Mapp. 2019;40(18):5315–5329. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B92] 92. Metzler-Baddeley C, O'Sullivan MJ, Bells S, Pasternak O, Jones DK. How and how not to correct for CSF-contamination in diffusion MRI. Neuroimage. 2012;59(2):1394–1403. [DOI] [PubMed] [Google Scholar]

[fcaf376-B93] 93. Guevara M, Guevara P, Román C, Mangin J-F. Superficial white matter: A review on the dMRI analysis methods and applications. Neuroimage. 2020;212:116673. [DOI] [PubMed] [Google Scholar]

[fcaf376-B94] 94. Corriveau-Lecavalier N, Gunter JL, Kamykowski M, et al. Default mode network failure and neurodegeneration across aging and amnestic and dysexecutive Alzheimer's disease. Brain Commun. 2023;5(2):fcad058. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B95] 95. Han F, Liu X, Mailman RB, Huang X, Liu X. Resting-state global brain activity affects early β-amyloid accumulation in default mode network. Nat Commun. 2023;14(1):7788. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B96] 96. Ingala S, Tomassen J, Collij LE, et al. Amyloid-driven disruption of default mode network connectivity in cognitively healthy individuals. Brain Commun. 2021;3(4):fcab201. [DOI] [PMC free article] [PubMed] [Google Scholar]

[fcaf376-B97] 97. Liu X, Tyler LK, Cam-Can, Davis SW, Rowe JB, Tsvetanov KA. Cognition’s dependence on functional network integrity with age is conditional on structural network integrity. Neurobiol Aging. 2023;129:195–208. [DOI] [PubMed] [Google Scholar]

[fcaf376-B98] 98. Rönnlund M, Nyberg L, Bäckman L, Nilsson LG. Stability, growth, and decline in adult life span development of declarative memory: Cross-sectional and longitudinal data from a population-based study. Psychol Aging. 2005;20(1):3–18. [DOI] [PubMed] [Google Scholar]

[fcaf376-B99] 99. Vidal-Pineiro D, Wang Y, Krogsrud SK, et al. Individual variations in ‘brain age’ relate to early-life factors more than to longitudinal brain change. Elife. 2021;10:e69995. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

High-resolution structural connectivity mediates age-related differences in functional connectivity and fluid cognition

Jenna L Merenstein

Allen W Song

David J Madden

Abstract

Graphical Abstract

Graphical Abstract.

Introduction

Materials and methods

Participants

Table 1.

Psychometric and cognitive tasks

Imaging data acquisition

Processing of DWI data

Processing of rsfMRI data

Graph theoretical analyses

Experimental design and statistical analysis

Results

Age-related differences in cognitive performance

Figure 1.

Structural connectivity

Correlations Between DWI Acquisitions

Table 2.

Age-related differences

Figure 2.

Figure 3.

Functional connectivity

Figure 4.

Structural versus functional connectivity

Table 3.

Connectivity versus cognition

Table 4.

Table 5.

Discussion

Structural connectivity and healthy aging

Structural connectivity and its relation to functional connectivity

Structural connectivity, functional connectivity, and fluid cognition

Limitations and future research

Conclusions

Supplementary Material

Acknowledgements

Contributor Information

Supplementary material

Funding

Competing interests

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases