Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Mar 11.
Published in final edited form as: Psychosom Med. 2018 Jan;80(1):69–77. doi: 10.1097/PSY.0000000000000526

THE ROLE OF BRAIN STRUCTURE IN PREDICTING ADHERENCE TO A PHYSICAL ACTIVITY REGIMEN

Swathi Gujral 1,2, Edward McAuley 3,4, Lauren E Oberlin 1,2, Arthur F Kramer 4, Kirk I Erickson 1,2
PMCID: PMC6411299  NIHMSID: NIHMS904552  PMID: 28914724

Abstract

Objective:

Physical activity is important for maintaining health throughout the lifespan. However, adherence to PA regimens is poor with approximately 50% of older adults terminating activity intervention programs within 6 months. In this study we tested whether gray matter volume and white matter microstructural integrity prior to the initiation of a PA intervention predicts PA adherence.

Methods:

159 adults aged 60–80 years were randomly assigned to a moderate-intensity aerobic walking condition or a non-aerobic stretching and toning condition. Participants engaged in supervised exercise 3 times per week for 12 months. Data were collected over a period of one year. Voxel based morphometry (VBM) and Tract-Based Spatial Statistics (TBSS) protocols were used to process neuroimaging data, and OLS regression models with bootstrapping were used to analyze voxelwise neural predictors of PA adherence.

Results:

Greater volume in several regions predicted greater PA adherence, including prefrontal, motor, somatosensory, temporal, and parietal regions (p< .01). We also found that higher fractional anisotropy (FA) in several white matter tracts predicted greater PA adherence (PFDR-corrected < .05), including the superior longitudinal fasciculus, anterior thalamic radiation, forceps minor, and body of the corpus callosum.

Conclusion:

These findings provide preliminary support for macro- and micro-structural neural predictors of PA adherence, and may translate to other health behaviors and behavioral goal-pursuit more broadly.

Keywords: physical activity, adherence, gray matter, white matter

INTRODUCTION

Physical activity (PA) is important for maintaining health and wellbeing throughout the lifespan (1–3). Yet, despite the awareness of the long-term health benefits of an active lifestyle, only ~10% of adults meet recommended PA guidelines and ~50% of people starting an exercise regimen stop within 6 months (4, 5). Identifying consistent determinants of adherence are essential for designing effective strategies, incentives, and interventions to enhance continued participation in PA.

Research on adherence to PA has focused on contextual and psychological factors, with little emphasis on neurobiological factors. Social-cognitive theory is the most widely used framework for studying psychological motivations for PA adherence, of which self-efficacy is a key construct (6, 7). Self-efficacy refers to one’s beliefs about his or her capability to successfully perform a specified behavior (8). Indices of exercise self-efficacy have been shown to consistently predict adherence to a PA regimen (7, 9).

Importantly, self-efficacy and other social cognitive factors that influence PA adherence are likely to depend on the integrity of neural networks involved in goal-directed behavior, self-reflection, and self-regulatory capacity, although this has not yet been tested (10–13). Further, biobehavioral models of PA, including the reflective impulsive model (RIM) and temporal self-regulation theory (TST), emphasize the importance of neurobiological factors for explaining PA adherence (11, 14–19). Examining structural neural markers of PA adherence may in fact capture the collective variance in PA adherence explained by self-report instruments of social cognitive constructs (i.e., self-efficacy, self-regulation). Recent studies using the novel neuroimaging research methodology known as ‘brain as a predictor’ approach, have shown that measures of neural structure and function can predict long-term health and clinical outcomes, such as smoking cessation, relapse in illicit drug use, and responsiveness to therapy in depressed patients (20). An important assumption underlying this approach is that neural markers serve as objective summary measures of psychological constructs predicting behavioral outcomes. To our knowledge, only one study has applied this ‘brain as a predictor’ approach to examine whether neuroimaging measures of neural integrity predict PA adherence in the context of structured exercise programming, and found that lateral prefrontal cortex volume predicts better exercise adherence among older women (21).

The aim of the present study was to replicate and extend the findings of Best et al. (21) by applying the ‘brain as a predictor’ approach to examine whether markers of gray matter and white matter integrity obtained at baseline of a 12-month randomized exercise intervention, predicted better intervention adherence in cognitively healthy older men and women. Specifically, we tested whether 1) greater gray matter volume and 2) higher fractional anisotropy (FA), a measure of white matter microstructural integrity, predicted better PA adherence. In a sensitivity analysis, we also tested the extent to which these markers of neural integrity predicted additional variance in PA adherence after accounting for exercise self-efficacy, a consistent predictor of PA adherence. Gray matter consists of neuronal cell bodies, unmyelinated axons, as well as glial cells and capillaries, which support neural function. In contrast, white matter consists of long-range myelinated axons that convey signals between various gray matter regions. We reasoned that gray matter volume and white matter microstructural integrity would be important markers of overall brain health in older adults that may predict PA adherence, given that both gray matter and white matter show age-related degeneration and have both been linked to superior cognitive functioning in older adults (13, 14, 16, 22, 23). Specifically, we predicted that greater gray matter volume and white matter microstructural integrity in a broad range of regions relevant to executive function, self-regulation, self-reflection, and PA engagement may be predictive of PA adherence over the 12-month intervention; these may include prefrontal, parietal and motor regions.

MATERIALS AND METHODS

Participants

One hundred and fifty-nine participants between the ages of 60 and 81 (mean age = 66.6 years; standard deviation = 5.6 years) were recruited to participate in a 1 year randomized exercise intervention examining the effects of aerobic fitness training on brain and cognitive health. This study was approved by the University of Illinois Institutional Review Board. Participants were recruited through community advertisements and physician referrals. Potential participants were initially screened over the phone for inclusion and exclusion criteria (see below for details). Upon passing the initial phone screening, participants were invited to a group orientation to receive study details and ask questions regarding the program. All participants provided informed consent. Three subsequent baseline sessions were performed after the group orientation. The current study focused on the high-resolution structural anatomical Magnetic Resonance images (MPRAGE) and Diffusion Tensor Imaging (DTI) data collected at baseline before randomization to the intervention and the exercise self-efficacy questionnaires described below.

Only a subsample (N= 105) of the original sample (N=159) with MPRAGE, self-efficacy, and adherence data had valid diffusion tensor imaging data that could be used to analyze the relationship between white matter integrity and PA adherence. Detailed characteristics of the sample used for DTI data analysis can be found in Oberlin and colleagues (24). The subsample used in the DTI data analysis did not differ from the original sample on demographic characteristics, self-efficacy, or PA adherence.

Investigations of the full sample and sub-samples of this trial have also been described in several prior papers (e.g. (25–29)).

Inclusion criteria

Inclusion criteria for entry into the trial were: 60+ years of age, capable of performing physical exercise, physician consent to perform physical exercise, successful completion of a graded maximal exercise test, and be physically inactive at baseline. An inactive lifestyle was defined as participating in no more than one 20-minute bout of physical activity per week for the past 6 months.

Exclusion criteria

Individuals with possible cognitive impairment, as indicated by a score below 51 on the modified Mini Mental Status Examination, clinical depression, as indicated by a score ≥ 2 on the Geriatric Depression Scale (GDS-5) (30), or impaired vision, as indicated by acuity greater than 20/40 were excluded from the study. Also, participants who did not meet safety criteria for participating in an MRI study were excluded from the intervention. These criteria include no previous history of head trauma, head or neck surgery, diabetes, neuropsychiatric or neurological conditions including brain tumors, or having any ferrous metallic implants that could cause injury due to the magnetic field.

Measures

Self-efficacy.

Participant’s perceptions of their ability to adhere to an exercise regimen, in the face of barriers, and to engage in PA were assessed using the three self-efficacy scales described below. These self-efficacy scales are commonly used measures of self-efficacy in the PA literature, and all have good internal consistency (α ≥ 0.93) (9, 17). All self-efficacy scales were administered to participants at the end of the third week of the exercise intervention to ensure accurate assessments of efficacy judgments.

Exercise Self-Efficacy Scale:

8-item scale that assesses individuals’ belief that they can exercise at moderate intensities three times per week for 40+ minutes at 1-week increments over the next 8-week period. This scale is scored on a 100-point percentage scale comprised of 10-point increments, ranging from 0% (not at all confident) to 100% (highly confident) (31). A total scale score is derived by summing the responses to each item and dividing by the total number of items in the scale. This measure has been used widely in the social cognitive literature in understanding PA and has demonstrated outstanding internal consistency (α = .99) (e.g., (32, 33)).

Barriers Self-Efficacy Scale:

13-item scale used to assess individuals’ perceived capabilities to exercise three times per week for 40 minutes over the next two months in the face of commonly identified barriers to participation. This scale is scored on a 100-point percentage scale comprised of 10-point increments, ranging from 0% (not at all confident) to 100% (highly confident). Responses to each item are summed, and divided by the total number of items to achieve an overall efficacy strength score ranging from 0 to 100. This scale has good internal consistency (α ≥ .93) (34).

Lifestyle Self-Efficacy Scale:

12-item scale used to assess individuals’ confidence in their ability to accumulate 30 min of PA on 5 or more days of the week for incremental monthly periods. The scale is scored on a 100-point percentage scale comprised of 10-point increments, 0–100 scale, ranging from 0% (not at all confident) to 100% (highly confident). Responses to each item are summed, and divided by the total number of items to achieve an overall efficacy strength score ranging from 0 to 100. The items in this scale have good internal consistency (α ≥ .95) (35).

Physical Activity Adherence.

Adherence reflects the percentage of attendance at exercise classes over the last 11 months of the intervention, given that self-efficacy (a predictor of adherence) was assessed 3 weeks after enrolling in the intervention. Attendance data were recorded each day by staff, aggregated, and divided by the total possible number of sessions to calculate PA adherence.

Physical Activity Scale for the Elderly.

The PASE is a 10-item self-report instrument designed to assess PA levels in large samples of older adults (over the age of sixty-five) over a one week time period. The PASE combines information from several domains including leisure, household, and occupational functioning. Participants indicate the frequency with which they participated in leisure activities (e.g., outdoor walking, light, moderate, and strenuous sports and recreation, muscle strengthening). The validity of the PASE has been established by studies (36, 37) showing an association between PASE scores and several physiological performance indicators, including: a sickness impact profile score, grip and leg strength, resting heart rate, age, peak oxygen uptake, percent body fat, and balance.

Self-Efficacy Composite Score

A composite self-efficacy score was created by standardizing and then averaging the self-efficacy scores from each of the three self-efficacy scales: exercise self-efficacy, barriers self-efficacy, and lifestyle self-efficacy (9, 17). The three self-efficacy scales were moderately correlated (r > 0.45 for all scales). See Table 1 for correlations between self-efficacy measures. The composite self-efficacy score was the final variable included in the regression models as a covariate. The composite self-efficacy score was used as a covariate in the regression model in order to assess the extent to which brain morphology predicts additional variance in PA adherence after accounting for self-efficacy.

Table 1.

Participant Characteristics

GM-Adherence Sample
N=159		 WM-Adherence Sample
N= 105
Mean Standard Deviation Mean Standard Deviation
Age (Years) 66.7 5.7 66.6 5.7
Years of Education 15.8 2.9 15.2 2.9
Sex (%Female) 66% -- 63% --
Exercise Self-Efficacy 84.1% 18.2% 76.9% 21.9%
Barriers Self-Efficacy 72.7% 19.8% 68.8% 21.1%
Lifestyle Self-Efficacy 79.0% 21.5% 73.3% 23.4%
Attendance 74.9% 17.4% 74.3% 18.2%
Open in a new tab
Structural magnetic resonance imaging (MRI).

MRI scanning was conducted prior to the start of the intervention. All participants underwent structural MRI scanning on a 3 Tesla Siemens Allegra scanner. High-resolution (1.3 mm × 1.3 mm × 1.3 mm) T1-weighted brain images were acquired using a 3D magnetization-prepared rapid gradient echo imaging protocol with 144 contiguous slices collected in an ascending fashion.

Diffusion tensor imaging.

Diffusion weighted images were acquired using a 3 T Siemens Allegra head-only scanner. The echo time (TE) was 94 ms, with repetition time (TR) = 4200 ms. Twenty-eight 4 mm slices positioned according to the AC-PC line were obtained along the anterior-posterior commissural plane. The protocol involved a T2-weighted acquisition followed by a 12-direction diffusion-weighted echo planar imaging scan (b-value = 1000 s/mm2), which was repeated six times.

Procedure

Participants came to the lab for a 2-hour baseline MRI session within one month prior to the start of the intervention. Structural MR images were collected during this session. During the intervention, participants reported to a university recreation facility 3 times a week for 40-minute sessions to either walk or participate in stretching and toning (control condition). In the walking condition, participants started off by walking for 10 minutes and increased walking duration by 5-min increments on a weekly basis until a duration of 40 minutes was achieved at week 7. Participants walked for 40 min per session for the remainder of the program. All walking was conducted on an indoor track. In the stretching condition, participants engaged in muscle-toning exercises using dumbbells or resistance bands, two exercises designed to improve balance, one yoga sequence, and one exercise of their choice. To keep participants interested, a new group of exercises was introduced every 3 weeks. Three weeks after the start of the intervention, participants were asked to complete exercise self-efficacy questionnaires. Participants then continued to participate in the intervention for 11 more months, at which time total adherence was assessed as the percentage of classes attended during this period.

Analysis

MRI Data Analysis: Gray Matter Volume

MR data was analyzed to determine the extent to which gray matter volume predicts PA adherence over the 11-month interval. MR data was processed using tools in the FMRIB Software Library (Image Analysis Group, FMRIB, Oxford, UK; http://www.fmrib.ox.ac.uk/fsl/; (38)). An optimized voxel based morphometry (VBM) protocol was used to analyze structural MRI data (FSL-VBM). An advantage of VBM is that it permits a whole-brain volumetric analysis in a semi-automated manner, making it easy to replicate for researchers with different levels of familiarity with neuroanatomy and does not limit analyses to particular regions-of-interest. A VBM analysis computes the probability that each voxel in a structural MR image is cerebrospinal fluid, gray matter, or white matter and yields statistical maps for each voxel type (see Ashburner and Friston (39) for a detailed description of VBM methods). Voxels are then classified into the structural category with the highest probability and can be statistically analyzed between subjects.

All images were processed using the following steps: (1) non-brain matter was removed using the brain extraction technique in FSL (40). (2) All brain-extracted images were visually inspected for any residual non-brain matter, and any residual matter was then manually removed from the image. (3) Next, these brain-extracted images were segmented into gray matter, white matter, and cerebrospinal fluid using FSL’s automated segmentation technique (41). Values were thresholded at >.2 to eliminate voxels that are of questionable tissue type. (4) Next, the partial volume estimate maps of gray matter were registered to the Montreal Neurological Institute template (42) and followed by non-linear registration (43) to a study-specific template created from the 159 participants with both MRI and self-efficacy data. (5) Each voxel of each registered gray matter image was modulated by applying the Jacobian determinant from the transformation matrix (44). (6) These modulated images were then concatenated into a 4D image, which was then smoothed using a 3 mm Gaussian kernel. Statistical analyses were then conducted on these segmented, registered, modulated, and smoothed gray matter images. A voxelwise threshold of p < 0.01 and a cluster-based threshold of p < 0.05 was used to determine statistical significance of the associations found in the regression models.

Diffusion Tensor Imaging Data Analysis: Fractional Anisotropy (FA)

Diffusion data was processed using FMRIB’s Diffusion Toolbox (v.3.0; http://fmrib.ox.ac.uk/fsl/fdt/index.html in the FMRIB Software Library (FSL v5.0.1) (Image Analysis Group, FMRIB, Oxford, UK; http://www.fmrib.ox.ac.uk/fsl/; (38). Voxelwise eigenvalues and eigenvectors of the diffusion tensor from each participant’s image were computed, calculating various diffusion parameters, including fractional anisotropy (FA). FA is a commonly used measure of white matter derived from DTI, and represents overall anisotropy within a voxel (45). FA values fall between 0 and 1, indicating the degree of microstructural organization, with higher values indicating greater directionality of diffusion.

FA data was fed into the FSL (v4.1.8) tract-based spatial statistics toolbox (TBSS; v1.2, http://www.fmrib.ox.ac.uk/fsl/tbss/index.html; (46). TBSS is used frequently in DTI processing (46). First, FA images were eroded to remove likely outliers. Then, FA images were normalized to MNI152 standard space. Next, a study-specific template was created and was used as the target for registration. To create the study specific template, we first registered all native-space FA images to the FA template in MNI space using an affine warp, then averaged the registered images across subjects to generate the study-specific template. Registration to the study-specific template is done by combining two transformations: 1) a non-linear transformation of each subject’s FA image to the study specific template and 2) an affine registration of the template to MNI152 standard space. Following registration, a mean FA image was computed and an average skeleton was generated that represented major tracts common across participants. The skeleton was thresholded at an FA value of 0.2 (46) to ensure that major white matter tracts were included. Then, in order to account for any residual misalignments not corrected for during registration, each participant’s normalized FA image was projected onto the mean FA skeleton. These images were then used in the statistical analyses described below.

Statistical Analysis: Bootstrap Regression Models

After obtaining the final voxel-wise partial volume estimates (PVE) of gray matter and final FA images projected onto the mean FA skeleton, we tested the association between gray matter volume and PA adherence and white matter microstructural integrity and PA adherence in older adults using the bootstrap regression tool within the Bootstrap Regression Analysis of Voxelwise Observations (BRAVO) toolbox (47, 48). Documentation and tutorials for this toolbox are available at https://sites.google.com/site/bravotoolbox. A key benefit of this toolbox is its flexibility to allow for the use of neural data as a predictor, rather than as just an outcome variable. We used identical regression models to test the association of gray matter volume and white matter microstructural integrity with PA adherence. First, we tested whether voxel-wise values of gray matter volume (PVE) and FA would separately predict PA adherence after adjusting for age, sex, education, and self-reported PASE score of PA at baseline using the bootstrap permutation test approach (49, 50). For each regression model, 1000 permutation tests were performed per voxel, and in each permutation test, the values in the variable vectors (covariates, gray matter volume, and PA adherence) were independently scrambled. The significance of the association was determined by comparing the distribution of bootstrapped values with the distribution of the original values using a bias-corrected and accelerated method (51).

Clusters of gray matter voxels showing significant associations with PA adherence were identified by using a conservative voxelwise threshold of p < 0.01 and a cluster-based threshold of p < 0.05. This approach uses random field theory and family-wise error to correct for multiple comparisons by determining the probability that the cluster of voxels could occur by chance given the smoothness of the data. To determine significant associations between FA data and PA adherence, we controlled for multiple comparisons ensexed by voxelwise testing using the False-Discovery Rate method (FDR). The FDR approach used the p-value distributions from our bootstrap regression models and yielded a q-value of 0.045. Thus, the significance threshold for all subsequent analyses with FA data was set as pFDR< 0.045.

As a sensitivity analysis, a similar bootstrap permutation approach was used to test the significance of the association between gray matter volume/FA and PA adherence after controlling for self-efficacy. The analysis was conducted to test the extent to which brain-PA adherence relationships were accounted for by self-efficacy, given that it is a key predictor of PA adherence.

RESULTS

Self-Efficacy Predicts Physical Activity Adherence

Characteristics of the 159 participants are shown in Table 1. As reported in previous studies using this sample (29), exercise self-efficacy ratings on each of the three self-efficacy scales were independently associated with adherence (all p’s < 0.05). See Table 2 for correlations between covariates (age and education), self-efficacy scales, and adherence. The association between self-efficacy and adherence did not vary by sex or years of education (all p’s > 0.05). Age was modestly correlated with adherence (r=0.16, p = 0.04), such that older participants had higher attendance rates during the intervention. Adherence rates did not significantly differ between men and women (χ2= 144.88, p = 0.25). After accounting for variance in adherence associated with age, sex, and education in a linear regression model, a composite score of the 3 self-efficacy scales explained 6% of the variance in adherence (Adjusted R2 Covariates: 0.017 Adjusted R2 change Self-efficacy= 0.056 Beta= 0.25 p = 0.002). The association between self-efficacy and adherence did not differ by intervention group when modeling the interaction term (walking vs. stretching) (Self-efficacy x Group interaction Beta = −0.08 p = 0.54). Thus, self-efficacy explains adherence to the PA intervention regardless of whether the intervention was moderate-intensity PA or stretching-and-toning PA. Importantly, adherence rates did not differ by intervention group.

Table 2.

Correlations between Self-Efficacy and PA Adherence in Total Sample (N=159)

1 2 3 4 5 6
1. Age -- −0.086 −0.038 −0.095 0.012 0.160*
2. Years of Education -- −0.081 −0.085 −0.069 −0.088
3. Exercise Self-efficacy -- 0.454** 0.589** 0.216**
4. Barriers Self-Efficacy -- 0.455** 0.224**
5. Lifestyle Self-Efficacy -- 0.169*
6. Attendance --
Open in a new tab

Pearson Correlations (2-tailed)

*

p < 0.05

**

p < 0.01

Gray Matter Volume Predicts Adherence to the Physical Activity Intervention

We used whole-brain voxelwise regression models with permutation testing in the BRAVO Matlab toolbox to test our hypothesis that gray matter volume in regions supporting self-regulatory, self-efficacy, and other executive processes would predict PA adherence after adjusting for age, education, baseline PA, and sex. Consistent with our hypothesis, greater gray matter volume in a broad range of regions predicted PA adherence, including bilateral precentral and postcentral gyrus, inferior temporal gyrus, temporoparietal junction, and superior parietal lobule (See Table 3). After controlling for self-efficacy, volume in many of these regions remained predictive of PA adherence, although a smaller percentage of voxels within each region was associated with PA adherence. After extracting partial volume estimates of regions significantly predictive of PA adherence, we found that gray matter volume explained 18% of variance in PA adherence, even after accounting for self-efficacy (R2 = 0.18). See Figure 1 for a visual comparison of gray matter regions associated with adherence with and without adjusting for self-efficacy. The intervention group assignment did not moderate these associations, so the interaction term modeling group assignment was subsequently dropped from the statistical model.

Table 3.

Gray Matter Clusters predicting PA Adherence

MNI Coordinates (COG)
Gray Matter Region Laterality Cluster Size (Voxels) Peak Z X Y Z
Gray Matter Volume
Predicting PA Adherence
Precentral/Postcentral Gyrus Left 4563 5.03 −40 −30 52
Inferior Temporal Gyrus Right 3050 5.16 18 −42 −30
Inferior Temporal Gyrus Left 1328 3.66 −38 −2 −34
Precentral/Postcentral Gyrus Right 1068 3.98 50 −18 44
Gray Matter Volume
Predicting PA Adherence Independent of SE
Inferior Temporal Gyrus Right 1383 4.7 30 −34 −30
Middle Frontal/Precental/
Postcentral Gyrus		 Left 1325 3.64 −32 −16 62
Inferior Temporal Gyrus Left 1149 3.62 −36 0 −36
Open in a new tab

Voxel-based morphometry (VBM) used to estimate gray matter clusters predictive of PA Adherence using a voxel-wise threshold p< 0.01, cluster threshold p<0.05

Figure 1. Gray Matter Regions Predictive of Physical Activity Adherence.

Figure 1.

Open in a new tab

(Top Row: Axial Slices, Bottom Row: 3D Sagittal View)

PrG= Precentral Gyrus, PoG= Postcentral gyrus, TPJ= Temporoparietal Junction, ITG= Inferior Temporal Gyrus.

RED represents GM regions predicting PA adherence that no longer remained significant (voxel-wise threshold p< 0.01, cluster threshold p<0.05) after controlling for SE.

BLUE represents GM regions that remained statistically significant (voxel-wise threshold p< 0.01, cluster threshold p<0.05) predictors of PA adherence after controlling for SE. Regions predicting PA adherence were left lateralized (See Table 3).

White Matter Microstructure Predicts Adherence to the Physical Activity Intervention

A voxelwise analysis revealed that greater FA in multiple white matter tracts predicted higher attendance rates/better adherence, even after adjusting for age, sex, baseline PA, and years of education. Regions containing clusters predictive of PA adherence included the body of the corpus callosum, the left forceps minor, right external capsule, and bilateral segments of the superior longitudinal fasciculus (SLF), and anterior thalamic radiation (ATR) (pFDRcorrected < 0.05). Associations between adherence and FA in the SLF, ATR, and body of the corpus callosum persisted after controlling for self-efficacy, although a smaller percentage of voxels within each predicted PA adherence. In contrast, the association between the left forceps minor and PA adherence was no longer significant after including self-efficacy in the model. Figure 2 shows the spatial distribution of statistically significant clusters within the white matter skeleton that predicted PA adherence both prior to and after adjusting for self-efficacy.

Figure 2. White mater microstructural Integrity predicts Physical Activity Adherence.

Figure 2.

Open in a new tab

RED represents regions in which integrity of white matter microstructure was predictive of PA adherence BEFORE controlling for Self-Efficacy (PFDR-corrected < .05)

BLUE represents regions in which integrity of white matter microstructure was predictive of PA adherence AFTER controlling for Self-Efficacy (PFDR-corrected < .05)

DISCUSSION

We tested the hypothesis that greater gray matter volume and white matter microstructural integrity would predict better adherence to a 12-month exercise intervention in older adults. Consistent with this prediction, greater gray matter volume in frontal, temporal, and parietal regions were predictive of better adherence to the intervention, irrespective of intervention group. Gray matter volume in many of these regions remained predictive of adherence after accounting for self-efficacy, although the percentage of voxels predictive of adherence declined. Higher white matter microstructural integrity in a wide array of tracts was also predictive of PA adherence, including the body of the corpus callosum, anterior thalamic radiation, superior longitudinal fasiculus, and forceps minor. Most of these associations remained significant after controlling for self-efficacy.

The aim of this study was to explore whether there is a neurobiological basis for continued participation in PA. While only one prior study has examined neural predictors of PA (21), recent efforts (11, 52) have explored cognitive predictors of PA, and have shown that cognitive control and self-regulatory processes that are critical for initiating and maintaining PA are supported by widely distributed brain networks spanning across the frontal, temporal, and parietal lobe (i.e., frontal-parietal network, cingulo-opercular network, and default-mode network) (See (11) for Review). The present findings contribute to this literature by showing that regions that predict PA adherence significantly overlap with several networks associated with cognitive control and self-regulatory processes.

The gray matter regions we found to predict PA adherence, including the regions around the intraparietal sulcus, the precentral gyrus, and lateral prefrontal cortex have been functionally linked as part of a ‘multiple demand’ system by Duncan and colleagues (53, 54) through resting state and task-evoked fMRI studies. These studies show that these regions may collectively represent a domain general network linked to cognitive control processes that are involved in a variety of behaviors, such as selective attention, maintenance of goals, and performance monitoring. The overlap between our gray matter findings and prior work identifying the ‘multiple demand’ system suggests that preservation of gray matter in these regions may not only be particularly beneficial for PA adherence, but may be more broadly implicated in a number of processes involved in behavioral goal-pursuit. The attenuation of some of these gray matter associations with adherence after accounting for self-efficacy is striking, given that self-efficacy only explained a modest amount of variance in PA adherence; this pattern may suggest that the self-efficacy-PA adherence relationship may partially mediate the relationship between gray matter volume and adherence, although this needs to be tested in future studies.

We also found that integrity of white matter microstructure in widely distributed white matter tracts was predictive of PA adherence. These regions included the temporal portion of the superior longitudinal fasiculus, which partially overlapped with gray matter regions we found to be predictive of PA adherence in the superior temporal cortex (See Z= 28 in Figures 1 and 2), suggesting that structural integrity of both gray matter and white matter in this region predicts PA adherence. Other white matter tracts that we found to predict PA adherence include the body of the corpus callosum, which facilitates inter-hemispheric communication, the anterior thalamic radiation, which is involved in reciprocal communication of limbic regions with prefrontal and anterior cingulate cortex, and the forceps minor, which is involved in communication between lateral and medial prefrontal regions. Interestingly, white matter integrity in the forceps minor became non-significant after co-varying for self-efficacy, suggesting that the self-efficacy-PA adherence relationship may be supported by structural connectivity between lateral and medial prefrontal regions, although this needs to be further explored. Importantly, integrity of all of these tracts is critical for the preservation of cognitive control processes in late life (55–57). In turn, these cognitive control processes are involved in self-regulation of a variety of goal-directed behaviors. Similar to the gray matter findings, the breadth of white matter tracts in which microstructural integrity was predictive of PA suggests that these findings may extend beyond PA to goal pursuit more generally.

These findings also support the ‘brain as a predictor’ approach to understanding real-world behavioral phenomenon (20). The aim of this methodological approach is to leverage objective measures of neural structure and function to predict long-term, ecologically valid outcomes that extend beyond laboratory testing. The advent of neuroimaging technology affords the possibility to link objective neurobiological markers to behavior in a variety of domains, including cognitive function, health (i.e. smoking cessation), economic decision-making, and clinical and neurological outcomes (20). Our findings contribute to this literature by showing that macro- and microstructural integrity of neural architecture in widely distributed networks predict PA adherence, many of which support cognitive control and self-regulatory processes.

The findings from the present study have shown that older adults with greater gray matter volume and white matter microstructure in regions supporting cognitive control and self-regulation, show better adherence to a yearlong structured PA intervention. These associations were also found to be statistically independent of randomization, such that gray matter volume and white matter integrity were predictive of better adherence regardless of whether participants were in the walking group or the stretching-and-toning group. The implications of these associations may also extend beyond PA adherence, to include the adoption and maintenance of other healthy lifestyle behaviors that are protective against physical and cognitive health decline. In turn, macro- and micro-structural integrity in these regions may broadly influence quality of life and participation in health behaviors more generally.

Future research can extend these findings by examining the extent to which gray matter and white matter regions predictive of adherence show PA-induced volumetric changes and whether such brain changes positively influence cognition. This will help us understand whether the relationship between brain health and adherence affects exercise-induced improvements in brain health. One possibility is that older adults with greater gray matter atrophy and white matter degeneration in regions supporting self-efficacy and self-regulatory strategies may not show as much exercise-related improvements in gray matter and white matter because of poor adherence. To address this, interventions could be tailored to focus on improving self-efficacy during the initial phases of the intervention and by improving self-regulatory skills, such as planning and goal setting. On the other hand, individuals with greater gray matter atrophy and white matter degeneration in these regions may show similar levels of improvement in brain health as those with less atrophy. This could indicate that those with poorer brain health have ‘more to gain’ from the exercise intervention, relative to those with better brain health. Future research can also expand on this study by examining the relationship between structural neural markers of adherence and other psychological predictors of PA adherence (i.e., self-regulatory strategies, executive functions). Examining which cognitive variables that are most related to neural markers of PA adherence may highlight efficient and cost-effective behavioral measures that investigators could use to identify those participants most likely to adhere, or not adhere, to a long-term structured PA intervention.

Limitations and summary

There are several limitations to the present study. First, the brain regions identified here are related to executive function, self-regulation, emotional control, reward, and other psychosocial, affective, and cognitive processes and we cannot determine the psychological constructs that are linked both to these regions and to adherence. Moreover, we did not account for cardiometabolic risk factors that are associated with both reduced gray and white matter integrity and sustained physical inactivity in this study. We see the outcomes from this study as an important first step in characterizing the neural correlates of PA adherence, but more research is needed before definitive conclusions can be made. Next, this was a 12-month intervention, and it is unclear whether these same associations would occur for trials of a different type, duration, or intensity (e.g., resistance training). This study was also conducted using a mostly white sample of highly educated healthy older adults from a midwestern city; therefore, these results may not be easily generalizable to more culturally diverse, younger, and clinical populations. There are a number of additional limitations related to the MRI analysis methods used in this study. First, voxel-based morphometry only provides estimates of tissue type, and does not provide absolute values of volume. Additionally, the diffusion data was collected using only 12 gradient directions, which limits the precision of the tensor estimation. Further, a FLAIR sequence was not run, so we could not control for white matter hyperintensities.

In summary, we found that gray matter volume and white matter microstructural integrity in a broad range of frontal, temporal, and parietal regions predicted adherence to a yearlong structured PA intervention in older adults. Many of these regions support executive control, self-regulatory processes, and voluntary motor function. These findings provide preliminary support for neural substrates underlying PA adherence in the context of structured exercise programming. Future research will need to expand on these findings by examining how these associations affect exercise-related improvements in brain health.

Table 4.

White Matter Regions predicting PA Adherence

MNI Coordinates (COG)
White Matter Tract Laterality Cluster Size (Voxels) Max P-Value X Y Z
Predicting PA Adherence
Posterior Corona Radiata Left 1564 0.999 −40 −30 52
Anterior Thalamic Radiation Right 1024 0.996 18 −42 −30
Anterior Thalamic Radiation Left 608 0.999 −33 −50 24
Body of Corpus Callosum Right 553 0.99 14 −10 3
Superior Longitudinal Fasiculus Left 550 0.998 −8 −14 8
Superior Longitudinal Fasiculus Left 484 0.999 −4 −9 27
Inferior Fronto-Occipital Fasiculus Left 479 0.999 −34 2 20
Superior Longitudinal Fasiculus Right 466 0.998 39 −40 26
Superior Longitudinal Fasiculus Right 315 0.99 32 27 1
Inferior Fronto-Occipital Fasiculus/
Uncinate Fasiculus
Right
302		 0.998 39 12 15
Forceps Minor Right 268 0.998 44 −9 23
Predicting PA Adherence Independent of SE
Superior Longitudinal Fasiculus Left 894 0.999 −35 −40 24
Anterior Thalamic Radiation Left 571 0.999 −8 −14 8
Superior Longitudinal Fasiculus Right 487 0.999 39 −39 25
Superior Longitudinal Fasiculus Right 480 0.994 43 −10 23
Superior Longitudinal Fasiculus Left 453 0.998 −35 −2 22
Superior Longitudinal Fasiculus Right 337 0.993 39 12 15
Anterior Thalamic Radiation/
Corticospinal Tract		 Right 282 0.995 20 −9 −4
Open in a new tab

FSL Tract-Based Spatial Statistics and BRAVO toolbox used to identify clusters of white matter regions in which FA values were predictive of adherence (pFDR= 0.05). Table is limited to clusters including > 200 voxels.

Funding Acknowledgement:

This work was supported by the National Institute on Aging at the National Institutes of Health (RO1 AG25667, RO1 AG25032) awarded to AK and EM. SG was supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. 2014192810. KIE was supported by research grants (R01 DK095172, P30 AG024827, P30 MH90333)

ACRONYMS:

PA

Physical Activity

VBM

Voxel based Morphometry

FA

Fractional Anisotropy

Footnotes

None of the authors have any conflicts of interest.

References

  • 1.Fagard RH. Exercise characteristics and the blood pressure response to dynamic physical training. Med Sci Sports Exerc. 2001;33:S484–92. [DOI] [PubMed] [Google Scholar]
  • 2.Colcombe S, Kramer AF. Fitness effects on the cognitive function of older adults: a meta-analytic study. Psychol Sci. [Meta-Analysis Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S.]. 2003;14:125–30. [DOI] [PubMed] [Google Scholar]
  • 3.Lavretsky H, Ballmaier M, Pham D, Toga A, Kumar A. Neuroanatomical characteristics of geriatric apathy and depression: a magnetic resonance imaging study. The American journal of geriatric psychiatry : official journal of the American Association for Geriatric Psychiatry. [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t]. 2007;15:386–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Conn VS, Minor MA, Burks KJ, Rantz MJ, Pomeroy SH. Integrative review of physical activity intervention research with aging adults. J Am Geriatr Soc. [Research Support, U.S. Gov’t, P.H.S. Review]. 2003;51:1159–68. [DOI] [PubMed] [Google Scholar]
  • 5.Dishman R Introduction: consensus, problems, and prospects Human Kinetics. Champaign, IL1994. p. 1–27. [Google Scholar]
  • 6.Young MD, Plotnikoff RC, Collins CE, Callister R, Morgan PJ. Social cognitive theory and physical activity: a systematic review and meta-analysis. Obes Rev. 2014;15:983–95. [DOI] [PubMed] [Google Scholar]
  • 7.McAuley E, Blissmer B. Self-efficacy determinants and consequences of physical activity. Exerc Sport Sci Rev. [Research Support, U.S. Gov’t, P.H.S. Review]. 2000;28:85–8. [PubMed] [Google Scholar]
  • 8.Bandura A The anatomy of stages of change. Am J Health Promot. [Editorial]. 1997;12:8–10. [DOI] [PubMed] [Google Scholar]
  • 9.McAuley E, Morris KS, Motl RW, Hu L, Konopack JF, Elavsky S. Long-term follow-up of physical activity behavior in older adults. Health Psychol. [Randomized Controlled Trial Research Support, N.I.H., Extramural]. 2007;26:375–80. [DOI] [PubMed] [Google Scholar]
  • 10.Amodio DM, Frith CD. Meeting of minds: the medial frontal cortex and social cognition. Nat Rev Neurosci. [Review]. 2006;7:268–77. [DOI] [PubMed] [Google Scholar]
  • 11.Hare TA, Hakimi S, Rangel A. Activity in dlPFC and its effective connectivity to vmPFC are associated with temporal discounting. Front Neurosci. 2014;8:50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Braver TS, Krug MK, Chiew KS, Kool W, Westbrook JA, Clement NJ, Adcock RA, Barch DM, Botvinick MM, Carver CS, Cools R, Custers R, Dickinson A, Dweck CS, Fishbach A, Gollwitzer PM, Hess TM, Isaacowitz DM, Mather M, Murayama K, Pessoa L, Samanez-Larkin GR, Somerville LH, group M. Mechanisms of motivation-cognition interaction: challenges and opportunities. Cogn Affect Behav Neurosci. 2014;14:443–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Buckley J, Cohen JD, Kramer AF, McAuley E, Mullen SP. Cognitive control in the self-regulation of physical activity and sedentary behavior. Frontiers in human neuroscience. [Review]. 2014;8:747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Best JR, Nagamatsu LS, Liu-Ambrose T. Improvements to executive function during exercise training predict maintenance of physical activity over the following year. Front Hum Neurosci. 2014;8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Hall PA, Fong GT, Epp LJ, Elias LJ. Executive function moderates the intention-behavior link for physical activity and dietary behavior. Psychol Health. [Clinical Trial Research Support, Non-U.S. Gov’t]. 2008;23:309–26. [DOI] [PubMed] [Google Scholar]
  • 16.Hall PA, Fong GT. Conscientiousness Versus Executive Function as Predictors of Health Behaviors and Health Trajectories. Annals of Behavioral Medicine. 2013;45:398–9. [DOI] [PubMed] [Google Scholar]
  • 17.McAuley E, Mullen SP, Szabo AN, White SM, Wojcicki TR, Mailey EL, Gothe NP, Olson EA, Voss M, Erickson K, Prakash R, Kramer AF. Self-regulatory processes and exercise adherence in older adults: executive function and self-efficacy effects. Am J Prev Med. [Randomized Controlled Trial Research Support, N.I.H., Extramural]. 2011;41:284–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Hall PA, Fong GT. Temporal self-regulation theory: a neurobiologically informed model for physical activity behavior. Frontiers in human neuroscience. 2015;9:117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hofmann W, Friese M, Wiers RW. Impulsive versus reflective influences on health behavior: A theoretical framework and empirical review. Health Psychology Review. 2008;2:111–37. [Google Scholar]
  • 20.Berkman ET, Falk EB. Beyond Brain Mapping: Using Neural Measures to Predict Real-World Outcomes. Current Directions in Psychological Science. 2013;22:45–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Best JR, Chiu BK, Hall PA, Liu-Ambrose T. Larger Lateral Prefrontal Cortex Volume Predicts Better Exercise Adherence Among Older Women: Evidence From Two Exercise Training Studies. The Journals of Gerontology: Series A. 2017;72:804–10. [DOI] [PubMed] [Google Scholar]
  • 22.Raz N, Gunning-Dixon FM, Head D, Dupuis JH, Acker JD. Neuroanatomical correlates of cognitive aging: evidence from structural magnetic resonance imaging. Neuropsychology. [Clinical Trial Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S.]. 1998;12:95–114. [DOI] [PubMed] [Google Scholar]
  • 23.Raz N, Lindenberger U, Rodrigue KM, Kennedy KM, Head D, Williamson A, Dahle C, Gerstorf D, Acker JD. Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cereb Cortex. [Clinical Trial Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov’t Research Support, U.S. Gov’t, P.H.S.]. 2005;15:1676–89. [DOI] [PubMed] [Google Scholar]
  • 24.Oberlin LE, Verstynen TD, Burzynska AZ, Voss MW, Prakash RS, Chaddock-Heyman L, Wong C, Fanning J, Awick E, Gothe N, Phillips SM, Mailey E, Ehlers D, Olson E, Wojcicki T, McAuley E, Kramer AF, Erickson KI. White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial working memory in older adults. Neuroimage. 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Erickson KI, Prakash RS, Voss MW, Chaddock L, Hu L, Morris KS, White SM, Wojcicki TR, McAuley E, Kramer AF. Aerobic fitness is associated with hippocampal volume in elderly humans. Hippocampus. [Research Support, N.I.H., Extramural]. 2009;19:1030–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Erickson KI, Voss MW, Prakash RS, Basak C, Szabo A, Chaddock L, Kim JS, Heo S, Alves H, White SM, Wojcicki TR, Mailey E, Vieira VJ, Martin SA, Pence BD, Woods JA, McAuley E, Kramer AF. Exercise training increases size of hippocampus and improves memory. Proc Natl Acad Sci U S A. [Randomized Controlled Trial Research Support, N.I.H., Extramural]. 2011;108:3017–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Prakash RS, Voss MW, Erickson KI, Lewis JM, Chaddock L, Malkowski E, Alves H, Kim J, Szabo A, White SM, Wojcicki TR, Klamm EL, McAuley E, Kramer AF. Cardiorespiratory fitness and attentional control in the aging brain. Frontiers in human neuroscience. 2011;4:229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Voss MW, Erickson KI, Prakash RS, Chaddock L, Kim JS, Alves H, Szabo A, Phillips SM, Wojcicki TR, Mailey EL, Olson EA, Gothe N, Vieira-Potter VJ, Martin SA, Pence BD, Cook MD, Woods JA, McAuley E, Kramer AF. Neurobiological markers of exercise-related brain plasticity in older adults. Brain Behav Immun. [Randomized Controlled Trial Research Support, N.I.H., Extramural]. 2013;28:90–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.McAuley E, Mailey EL, Mullen SP, Szabo AN, Wojcicki TR, White SM, Gothe N, Olson EA, Kramer AF. Growth trajectories of exercise self-efficacy in older adults: influence of measures and initial status. Health Psychol. [Randomized Controlled Trial Research Support, N.I.H., Extramural]. 2011;30:75–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Sheikh JI, Yesavage JA Geriatric depression scale (GDS): Recent evidence and development of a shorter version In: Brink TL, editor. Clinical Gerontology: A Guide to Assessment and Intervention. New York: The Haworth Press; 1986. p. 165–73. [Google Scholar]
  • 31.McAuley E. Self-efficacy and the maintenance of exercise participation in older adults. J Behav Med. [Research Support, U.S. Gov’t, P.H.S.]. 1993;16:103–13. [DOI] [PubMed] [Google Scholar]
  • 32.Duncan TE, McAuley E. Social support and efficacy cognitions in exercise adherence: a latent growth curve analysis. J Behav Med. [Research Support, U.S. Gov’t, P.H.S.]. 1993;16:199–218. [DOI] [PubMed] [Google Scholar]
  • 33.McAuley E, Jerome GJ, Elavsky S, Marquez DX, Ramsey SN. Predicting long-term maintenance of physical activity in older adults. Prev Med. [Clinical Trial Randomized Controlled Trial Research Support, U.S. Gov’t, P.H.S.]. 2003;37:110–8. [DOI] [PubMed] [Google Scholar]
  • 34.McAuley E The role of efficacy cognitions in the prediction of exercise behavior in middle-aged adults. J Behav Med. 1992;15:65–88. [DOI] [PubMed] [Google Scholar]
  • 35.McAuley E, Hall KS, Motl RW, White SM, Wojcicki TR, Hu L, Doerksen SE. Trajectory of declines in physical activity in community-dwelling older women: social cognitive influences. J Gerontol B Psychol Sci Soc Sci. 2009;64:543–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activity Scale for the Elderly (PASE): development and evaluation. Journal of clinical epidemiology. 1993;46:153–62. [DOI] [PubMed] [Google Scholar]
  • 37.Washburn RA, McAuley E, Katula J, Mihalko SL, Boileau RA. The physical activity scale for the elderly (PASE): evidence for validity. Journal of clinical epidemiology. 1999;52:643–51. [DOI] [PubMed] [Google Scholar]
  • 38.Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang Y, De Stefano N, Brady JM, Matthews PM. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage. [Research Support, Non-U.S. Gov’t Review]. 2004;23 Suppl 1:S208–19. [DOI] [PubMed] [Google Scholar]
  • 39.Ashburner J, Friston KJ. Voxel-based morphometry--the methods. Neuroimage. [Research Support, Non-U.S. Gov’t Review]. 2000;11:805–21. [DOI] [PubMed] [Google Scholar]
  • 40.Smith SM, Nichols TE. Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage. [Research Support, Non-U.S. Gov’t]. 2009;44:83–98. [DOI] [PubMed] [Google Scholar]
  • 41.Zhang Y, Brady M, Smith S. Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging. [Research Support, Non-U.S. Gov’t]. 2001;20:45–57. [DOI] [PubMed] [Google Scholar]
  • 42.Jenkinson M, Smith S. A global optimisation method for robust affine registration of brain images. Med Image Anal. [Comparative Study Research Support, Non-U.S. Gov’t]. 2001;5:143–56. [DOI] [PubMed] [Google Scholar]
  • 43.Andersson J, Jenkinson M, Smith S Non-linear optimisation, aka spatial normalisation. 2007; FMRIB technical report TR07JA1; ]. Available from: www.fmrib.ox.ac.uk/analysis/techrep. [Google Scholar]
  • 44.Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS. A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage. [Research Support, Non-U.S. Gov’t]. 2001;14:21–36. [DOI] [PubMed] [Google Scholar]
  • 45.Jones DK, Knosche TR, Turner R. White matter integrity, fiber count, and other fallacies: the do’s and don’ts of diffusion MRI. Neuroimage. 2013;73:239–54. [DOI] [PubMed] [Google Scholar]
  • 46.Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE, Watkins KE, Ciccarelli O, Cader MZ, Matthews PM, Behrens TE. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage. 2006;31:1487–505. [DOI] [PubMed] [Google Scholar]
  • 47.Verstynen TD, Lynch B, Miller DL, Voss MW, Prakash RS, Chaddock L, Basak C, Szabo A, Olson EA, Wojcicki TR, Fanning J, Gothe NP, McAuley E, Kramer AF, Erickson KI. Caudate Nucleus Volume Mediates the Link between Cardiorespiratory Fitness and Cognitive Flexibility in Older Adults. Journal of aging research. 2012;2012:939285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Gianaros PJ, Marsland AL, Sheu LK, Erickson KI, Verstynen TD. Inflammatory pathways link socioeconomic inequalities to white matter architecture. Cerebral cortex. [Research Support, N.I.H., Extramural]. 2013;23:2058–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Manly B Randomization, bootstrap, and monte carlo methods in biology. London: Chapman & Hall; 1997. [Google Scholar]
  • 50.Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods. 2008;40:879–91. [DOI] [PubMed] [Google Scholar]
  • 51.DiCiccio TJ, Efron B. Bootstrap confidence intervals. Statistical Science. 1996;11:189–228. [Google Scholar]
  • 52.Hall PA, Fong GT. Temporal self-regulation theory: looking forward. Health Psychol Rev. 2010;4:83–92. [Google Scholar]
  • 53.Jasinska AJ, Ramamoorthy A, Crew CM. Toward a neurobiological model of cue-induced self-control in decision making: relevance to addiction and obesity. The Journal of neuroscience : the official journal of the Society for Neuroscience. [Comment]. 2011;31:16139–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Hare TA, Malmaud J, Rangel A. Focusing attention on the health aspects of foods changes value signals in vmPFC and improves dietary choice. The Journal of neuroscience : the official journal of the Society for Neuroscience. [Research Support, U.S. Gov’t, Non-P.H.S.]. 2011;31:11077–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Kelly AMC, Uddin LQ, Biswal BB, Castellanos FX, Milham MP. Competition between functional brain networks mediates behavioral variability. NeuroImage. 2008;39:527–37. [DOI] [PubMed] [Google Scholar]
  • 56.Hutcherson CA, Plassmann H, Gross JJ, Rangel A. Cognitive Regulation during Decision Making Shifts Behavioral Control between Ventromedial and Dorsolateral Prefrontal Value Systems. Journal of Neuroscience. 2012;32:13543–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Koechlin E, Basso G, Pietrini P, Panzer S, Grafman J. The role of the anterior prefrontal cortex in human cognition. Nature. 1999;399:148–51. [DOI] [PubMed] [Google Scholar]

RESOURCES