Toward prognostic sensitivity of DTI to cognitive outcomes following physical exercise interventions.
Anna Rita Egbert1, Ryan Falck1, John Best1, and Teresa Liu-Ambrose1,2

1Aging, Mobility and Cognitive Neuroscience Laboratory, University of British Columbia, Vancouver, BC, Canada, 2Centre for Hip Health and Mobility, Vancouver Coastal Health Research Institute and University of British Columbia, Vancouver, BC, Canada


Diffusion Tensor Imaging (DTI) is widely implemented in clinical research, yet, its prognostic value in brain and cognitive health remains uncertain. Prospective estimation of the effectiveness of interventions, such as physical exercise, can increase cost-effectiveness of treatment, thus, maximizing the impact of accessible modifiable preventive factors in improving health outcomes in the general population. We examined DTI in relationship to cognitive outcomes of physical activity intervention. This is the first study to show that white matter integrity related to sleep efficiency can be an early predictor of the cognitive outcomes of physical exercise intervention.


There is a growing evidence that physical exercise benefits brain health1 and augments cognitive function2. However, there is much variability in the degree of benefit observed between individuals. Thus, it is critical to define key predictors of exercise efficacy. The purpose of this study was to examine white matter (WM) tractography as a metric of prospective estimation of physical exercise intervention effects on increasing cognitive function in older adults with osteoarthritis. Our hypotheses were that: 1) brain WM integrity is related to physical activity and sleep efficiency; 2) physical exercise enhances cognitive function; 3) brain WM integrity at baseline is related to cognitive outcomes following physical exercise intervention.


We analyzed a subset of data acquired from 16 older adults with osteoarthritis (age M=60, SD=7.7 years) enrolled in a randomized controlled trial (RCT); all 16 participants were enrolled in the intervention which consisted of: 1) standardized group education session about benefits of moderate-to-vigorous physical activity (MVPA) and sedentary behavior (SB); 2) a Fitbit® FlexTM; and 3) individual activity counselling with a physiotherapist. Throughout 2-month intervention, physical activity goals were assessed and modified during biweekly over-the-phone counselling with a physiotherapist. Baseline assessment included: DTI; extensive cognitive test battery (i.e., NIH Toolbox Cognition Battery3; Stroop Colour-Word Test); and objective measure of physical activity and sleep (i.e., SenseWear Mini4). Follow-up cognitive assessment was conducted at the end of the 2-month intervention. Additional cognitive examination took place at a 2-month post-intervention. We performed DTI preprocessing in DiffusionKit toolbox, and conducted tractography with the tbss scripts implemented in FSL. Cognitive scores were computed as variables reflecting the change in cognitive scores between the three time points, i.e., Change A = post intervention minus baseline; Change B = 2-month post-intervention minus post intervention. The following scores are reported here: NIH Toolbox PSM Mean Score (higher score means better outcomes); Stroop Test Mean RT to Incongruent Cues (higher scores signify decline in reaction time, i.e., worsening of performance); and Stroop Test Standard Deviation RT to Incongruent Cues (higher score means more stable performance; i.e., better outcomes). Data on physical activity (i.e., MVPA, SB), sleep total sleep efficiency (i.e., time in sleep phase vs. time trying to sleep, and sleep total time) were analyzed with Camntech software. Group-level statistics were performed with randomise command, where the model was performed on the individual whole-brain WM tracts, and the regressors were indicated as follows: age, MVPA, SB, sleep efficiency, sleep duration. Analyses were performed with Bonferroni correction for multiple comparisons and cluster size above 100. Clusters and the WM tracts they belong to were visualized using TrackVis. Next, for each participant, we extracted the mean FA values from all clusters showing significant effects of interest. These values were then analyzed in relationship to variables describing cognitive change due to the intervention.


At baseline, WM integrity was related to sleep efficiency and duration but not MVPA, SB, or age. For sleep efficiency, there were 13 significant clusters, and 5 clusters for sleep duration. Table 1 presents data from the obtained significant clusters (i.e., peak coordinates, FA values and cluster size). Out of those, two clusters survived Sperman's Rho correlation with cognitive change variables. Figure 1 shows those two significant clusters with the WM pathways that they belong to. Data in Table 2 show that higher mean FA values in those two clusters were related to better executive and memory outcomes following physical activity intervention, except for more disrupted FA in Cluster_3' being related to higher increases in Picture Sequence Memory Test score immediately after intervention.


Our data show DTI to be sensitive to sleep efficiency and duration in adults with no neurological conditions. Importantly, WM integrity (in tracts sensitive to sleep efficiency), was related to the degree of cognitive change following physical exercise intervention. Generally, higher WM integrity at baseline was related to more increases in executive function and working memory immediately after intervention as well as at 2-month delay. Noted Cluster_8' falls within WM tracts connecting temporal, frontal, parietal, occipital regions and hypothalamus. Such structural connectivity degradation may have widespread effects disabling the interplay between brain regions related to higher order cognitive processes and downstream physiological functions. This findings allows to articulate hypotheses on the hypothalamic involvement in regulating the effects of physical exercise. Furthermore, Cluster_3' is integral to WM tracts which has substantial amount of the projections leading between cortical regions and cerebellum. This allows to speculate about the functional role of the potential back projections from the cerebellum to cortical regions in the effectiveness of physical exercise interventions.


Funding for this project was provided by a Canadian Institutes of Health Research Operational Grant (F14-03974), and by the Jack Brown & Family Alzheimer’s Research Foundation. RSF is funded by the University of British Columbia Rehabilitation Sciences Scholarship, the Omer H. Patrick II Memorial Prize, and the Louise McGregor Memorial Award in Neurorehabilitation. TLA is a Canada Research Chair.


1. Negash S, Bennett DA, Wilson RS, Schneider JA, Arnold SE. Cognition and neuropathology in aging: multidemensional perspectives from the Rush Religious Orders Study and Rush Memory and Aging Project. Curr Alzheimer Res 2011;8(4):336-40.

2. Stern Y. Cognitive reserve: Implications for Assessment and Intervention. Folia Phoniatr Logop. 2013;65(2):49-54.

3. Dikmen SS, Bauer PJ, Weintraub S, Mungas D, Slotkin J, Beaumont JL, et al. Measuring episodic memory across the lifespan: NIH toolbox picture sequence memory test. JINS. 2014;20(6):611-9.

4. Johannsen DL, Calabro MA, Stewart J, Franke W, Rood JC, Welk GJ. Accuracy of armband monitors for Measuring Daily Energy Expenditure in Health Adults. Med Sci Sports Exerc. 2011;42(11):2134-40.


Figure 1. Two regions are presented together with the white matter tracts that they belong to. The significant disruption of white matter integrity in these two regions was significantly related to reduced sleep efficiency (controlling for age, physical activity and sedentary behavior), as well as lower cognitive outcomes following physical exercise intervention (i.e., immediately after and at 2-month follow-up). Row A shows the directional staining of white matter tracts where Cluster_4' dorsal-ventral, and Cluster_9' anterior-posterior and left-right. Row B shows separate pathways with solid staining for Cluster_4' – red, and Cluster_9' – yellow.

Table 1. Characteristics of the white matter clusters showing significant effects of sleep quality.

Table 2. Sperman's Rho correlation between mean values of WM integrity in cluster sensitive to sleep effects at baseline and cognitive outcomes following physical exercise intervention.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)