YNIMG-12612; No.

of pages: 11; 4C: 5, 6, 8

NeuroImage xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

NeuroImage

journal homepage: www.elsevier.com/locate/ynimg

1Q1 White matter microstructure mediates the relationship between

2 cardiorespiratory fitness and spatial working memory in older adults☆

F
3Q2 Lauren E. Oberlin a,b,⁎, Timothy D. Verstynen b,c, Agnieszka Z. Burzynska d,g, Michelle W. Voss e,
Ruchika Shaurya Prakash f, Laura Chaddock-Heyman g, Chelsea Wong g, Jason Fanning h, Elizabeth Awick h,

O
4
5 Neha Gothe i, Siobhan M. Phillips j, Emily Mailey k, Diane Ehlers h, Erin Olson l, Thomas Wojcicki m,
Edward McAuley h, Arthur F. Kramer g, Kirk I. Erickson a,b

O
6

7Q3 a
Department of Psychology, University of Pittsburgh, USA
8 b
Center for the Neural Basis of Cognition, University of Pittsburgh, USA

R
9 c
Department of Psychology, Carnegie Mellon University, USA
10 d
Department of Human Development and Family Studies, Colorado State University — Fort Collins, USA
11 e
Department of Psychological and Brain Sciences, University of Iowa, USA

P
12 f
Department of Psychology, Ohio State University, USA
13 g
Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, USA
14 h
Department of Kinesiology and Community Health, University of Illinois at Urbana-Champaign, USA
15 i
Department of Kinesiology, Wayne State University, USA D
16 j
Department of Preventative Medicine, Northwestern University, USA
17 k
Department of Kinesiology, Kansas State University, USA
18 l
Harvard Medical School, USA
E
19 m
Exercise Science Department, Bellarmine University, USA
20
T

2 1 a r t i c l e i n f o a b s t r a c t
C

22 Article history: White matter structure declines with advancing age and has been associated with a decline in memory and exec- 30
23 Accepted 22 September 2015 utive processes in older adulthood. Yet, recent research suggests that higher physical activity and fitness levels may 31
24 Available online xxxx
E

be associated with less white matter degeneration in late life, although the tract-specificity of this relationship is not 32
well understood. In addition, these prior studies infrequently associate measures of white matter microstructure to 33
25 Keywords:
cognitive outcomes, so the behavioral importance of higher levels of white matter microstructural organization 34
R

26 Aging
27 Memory
with greater fitness levels remains a matter of speculation. Here we tested whether cardiorespiratory fitness 35
28 White matter (VO2max) levels were associated with white matter microstructure and whether this relationship constituted an 36
29 indirect pathway between cardiorespiratory fitness and spatial working memory in two large, cognitively and 37
R

Fitness
neurologically healthy older adult samples. Diffusion tensor imaging was used to determine white matter micro- 38
structure in two separate groups: Experiment 1, N = 113 (mean age = 66.61) and Experiment 2, N = 154 39
O

(mean age = 65.66). Using a voxel-based regression approach, we found that higher VO2max was associated with 40
higher fractional anisotropy (FA), a measure of white matter microstructure, in a diverse network of white 41
matter tracts, including the anterior corona radiata, anterior internal capsule, fornix, cingulum, and corpus callosum 42
C

(PFDR-corrected b .05). This effect was consistent across both samples even after controlling for age, gender, and edu- 43
cation. Further, a statistical mediation analysis revealed that white matter microstructure within these regions, 44
among others, constituted a significant indirect path between VO2max and spatial working memory performance. 45
N

These results suggest that greater aerobic fitness levels are associated with higher levels of white matter microstruc- 46
tural organization, which may, in turn, preserve spatial memory performance in older adulthood. 47
© 2015 Published by Elsevier Inc. 48
U

52
50
49
51
Introduction 53

The aging brain experiences both macro- and microstructural 54

changes, including gray matter atrophy and degeneration of white 55
matter tracts. Older adults are particularly susceptible to precipitous 56
☆ Author note: All authors report no conflicts of interest.
declines in white matter, with anterior tracts showing the most 57
⁎ Corresponding author at: Department of Psychology, University of Pittsburgh, 210 S.
Bouquet St., 3211 Sennott Square, Pittsburgh, PA 15260, USA. pronounced degradation (Burzynska et al., 2010; Pfefferbaum and 58
E-mail address: leo11@pitt.edu (L.E. Oberlin). Sullivan, 2003; Salat et al., 2005; Westlye et al., 2009). Age-related 59

http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
1053-8119/© 2015 Published by Elsevier Inc.

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 2 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx

60 declines in white matter microstructure may lead to disruptions in performance (Prakash et al., 2010; Voss et al., 2013b). Following a 126
61 neural communication, which, in turn, could lead to consequent one-year aerobic exercise intervention (n = 70), Voss et al. (2013b) 127
62 declines in cognitive function. This notion is supported by studies found that greater gains in CRF were associated with greater FA in 128
63 combining behavioral measures of cognitive performance and diffusion prefrontal and temporal white matter. However, the increases in 129
64 tensor imaging (DTI), which demonstrate that age-related decline in ce- white matter FA post-intervention were not associated with memory 130
65 rebral white matter may be related to cognitive deficits associated with improvement, although this may be a consequence of lack of statistical 131
66 aging. In particular, white matter degeneration in older adults is associ- power for the cognitive measure employed (backward digit span) (Voss 132
67 ated with impaired performance on memory, executive function, and et al., 2013b). Similar patterns emerged in a study of multiple sclerosis 133
68 processing speed tasks (Bennett and Madden, 2014; Charlton et al., (MS) (MS n = 21; Healthy controls n = 15), but in this study higher 134
69 2006; Gold et al., 2010; Grieve et al., 2007; Kennedy and Raz, 2009; FA was associated with both greater fitness levels and faster processing 135
70 Madden et al., 2009; Vernooij et al., 2009; Voineskos et al., 2012). speed (Prakash et al., 2010). Thus, while only a small number of studies 136
71 Fortunately, moderate intensity physical activity (PA) may prevent have examined the link between fitness and white matter microstruc- 137
72 or reverse age-related changes in neural structure and function. For ture, even fewer have investigated the cognitive implications of this 138
73 example, randomized controlled trials (RCTs) have demonstrated that relationship. 139

F
74 6–12 months of aerobic exercise improves cognitive performance and Our primary aim was to further examine the relationship between 140
75 alters gray matter structural morphology and function in older CRF and white matter microstructure in older adulthood. Using diffu- 141

O
76 adulthood (Colcombe and Kramer, 2003; Colcombe et al., 2006; sion imaging, we examined the relationship between CRF and tensor- 142
77 Erickson et al., 2011; Niemann et al., 2014; Ruscheweyh et al., 2011; based models of white matter structure in two large samples of healthy 143
78 Voelcker-Rehage et al., 2011; Voss et al., 2010b). Despite the wealth of older adults (each with n N 110). Importantly, the two samples include 144

O
79 data on associations between PA, fitness, and exercise on gray matter an objective measure of aerobic fitness (CRF), collected using identical 145
80 volume in older adults (Erickson et al., 2014) considerably less is assessment methods, employed similar spatial memory tasks, and had 146
81 known about these relationships with white matter microstructure. A similar demographic characteristics. Such similarities between two 147

R
82 handful of recent cross-sectional and prospective longitudinal studies large, and independent, samples provided us with a unique opportunity 148
83 provide preliminary evidence that self-reported regular PA may to examine associations between fitness and white matter microstruc- 149

P
84 preserve white matter in older adulthood (Gons et al., 2013; Gow ture on a voxelwise basis to better characterize the tract-specificity of 150
85 et al., 2012; Tian et al., 2014a). In addition, positive relationships these associations. Based on studies of grey matter, fitness and exercise 151
86 between objectively measured PA using accelerometry and fractional are most consistently associated with volume of the hippocampus and
D 152
87 anisotropy (FA), a measure of white matter microstructure, have been prefrontal cortex (Erickson et al., 2009, 2011, 2014; Weinstein et al., 153
88 shown (Burzynska et al., 2014; Tian et al., 2015). Recent work has also 2012). Therefore, we predicted that higher CRF would be associated 154
89 examined associations between white matter microstructure and with greater FA, particularly in tracts that facilitate communication be- 155
E
90 cardiorespiratory fitness (CRF) in older adulthood. CRF, often measured tween subcortical, hippocampal, and prefrontal regions. A secondary 156
91 by VO2max, a quantitative estimate of oxygen capacity and utilization aim was to examine whether fitness and white matter associations 157
T

92 during an exercise test, can be improved (increased) by aerobic would constitute significant indirect pathways to spatial working 158
93 exercise. Higher CRF levels have been associated with greater white memory performance, a cognitive domain that is sensitive to aging 159
C

94 matter microstructural integrity in several tracts including the cingu- (Schmiedek et al., 2009), and fitness effects (Erickson et al., 2009). In ad- 160
95 lum, corpus callosum, superior corona radiata, and inferior longitudinal dition, performance on the spatial working memory paradigm used in 161
96 fasciculus (Hayes et al., 2015; Johnson et al., 2012; Marks et al., 2010; the present investigation has been linked in prior studies to fitness 162
E

97 Tseng et al., 2013); however see Burzynska et al., 2014). Yet, despite and hippocampal volume (Erickson et al., 2009, 2011) and resting con- 163
98 this evidence for associations between PA, fitness, and white matter nectivity (Voss et al., 2010b), making this measure well-suited for the 164
R

99 microstructure, many of these studies have been encumbered by purposes of the present study. Also, similar spatial working memory 165
100 methodological limitations that restrict the scope of interpretation. paradigms were used in both experiments described here, which 166
101 Most prior studies have had relatively small sample sizes (n b 30 healthy allowed us to test associations between white matter and spatial 167
R

102 older adults) (Johnson et al., 2012; Marks et al., 2010; Tian et al., 2014b; memory performance in two different samples. To this end, quantitative 168
103 Tseng et al., 2013), or have employed subjective measures of PA that are mediation modeling was applied on a voxelwise basis to both samples 169
O

104 prone to social desirability bias and may not reflect actual PA patterns to investigate whether white matter microstructure statistically 170
105 (Gons et al., 2013; Gow et al., 2012; Tian et al., 2014a). In addition, mediated the relationship between fitness and spatial working memory 171
106 there is little agreement across studies on the particular white matter performance. We predicted that white matter microstructure would be 172
C

107 paths that correlate with PA or fitness. This may partially be due to the a significant indirect pathway by which higher CRF would be associated 173
108 analytical approaches used in previous work, which in some cases has with superior spatial working memory performance. Characterizing 174
been limited to particular fiber bundles at the expense of other brain
N

109 such pathways provides insight into the putative mechanisms and 175
110 areas (Marks et al., 2010; Tian et al., 2014a). Thus, while previous re- neurocognitive implications of fitness-related variation in white matter 176
111 search suggests a positive linear relationship between fitness and microstructure in older adults. 177
U

112 white matter microstructural integrity, regional specificity remains un-

113 clear, with small sample sizes and methodological limitations restricting Methods 178
114 interpretation.
115 There is also a dearth of knowledge on the role of white matter in the We tested our hypotheses in two separate samples that are 179
116 relationship between fitness and cognitive function. Higher levels of described below as Experiment 1 and 2. The data analysis and analytic 180
117 aerobic fitness are frequently associated with better cognitive procedures are described last as they were the same across both 181
118 performance and, as recent research suggests, greater white matter mi- experiments. 182
119 crostructural integrity (Hayes et al., 2015; Johnson et al., 2012; Tian
120 et al., 2014b; Tseng et al., 2013). Given the association between white Experiment 1 183
121 matter and cognition in healthy older adults, fitness-related variation
122 in white matter microstructure may partially mediate the relationship Participant characteristics 184
123 between fitness and cognitive function, although this remains a matter One hundred and seventy-three participants between the ages of 60 185
124 of speculation. Only two studies have examined whether microstructur- and 81 (mean age 66.6 years; standard deviation = 5.6 years) were re- 186
125 al changes associated with fitness are also linked to elevated cognitive cruited to take part in a one-year, single-blind randomized controlled 187

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx 3

188 exercise intervention. Subjects were recruited through community Following the 3-second delay, a red dot appeared on the screen. 252
189 advertisements and physician referrals in eastern Illinois. For the Subjects had to indicate whether the red dot displayed was in the 253
190 purposes of the present study, only data from baseline assessments same location (match) or a different location (non-match) than one of 254
191 were used. All subjects participated in three baseline sessions, which in- the previously presented black dot(s) by pressing a designated key on 255
192 cluded cognitive and CRF assessments and the collection of magnetic a computer keyboard (x = nonmatch; m = match). There were 40 trials 256
193 resonance imaging (MRI) data. Of the 173 participants, 16 were per set size (1, 2, and 3 black dots), with 20 match and 20 non-match 257
194 excluded due to incomplete information on relevant behavioral data trials in each, totaling to 120 trials. Prior to task administration, several 258
195 (cognitive, VO2, or demographic). An additional 12 subjects did not practice trials were conducted to familiarize the participant with the 259
196 complete the baseline MRI scan. Finally, 32 participants were excluded task, during which time they were directed to respond as quickly and 260
197 due to problems with their diffusion imaging data including 1) failure accurately as possible. Accuracy rates were recorded separately for the 261
198 to finish the entire scan (n = 2), 2) poor orientation during image acqui- 1, 2, and 3 dot conditions and were then averaged to create mean 262
199 sition (n = 4), and 3) excessive noise/field distortion (n = 24). Thus, the accuracy scores. 263
200 final sample consisted of 113 participants. Excluded participants did not
201 differ in age, gender, years of education, or fitness levels from those

F
202 included in the analysis (all p N .10). Diffusion tensor imaging 264
203 Participants were required to score ≥ 51 on the modified Mini Diffusion weighted images were acquired using a 3 T Siemens 265

O
204 Mental Status Examination (high score of 57) to rule out clinically pres- Allegra head-only scanner. The echo time (TE) was 94 ms, with repeti- 266
205 ent cognitive impairment (Stern et al., 1987). Additionally, to rule out tion time (TR) = 4200 ms. Twenty-eight 4 mm slices positioned accord- 267
206 depression, individuals that scored N3 on the Geriatric Depression ing to the AC-PC line were obtained along the anterior-posterior 268

O
207 Scale were excluded (Sheikh et al., 1986). Participants were required commissural plane. The protocol involved a T2-weighted acquisition 269
208 to have normal color vision, a visual acuity of at least 20/40, and no his- followed by a 12-direction diffusion-weighted echo planar imaging 270
scan (b-value = 1000 s/mm2), which was repeated six times.

R
209 tory of neuropsychiatric conditions or neurological diseases or infarcts 271
210 including Parkinson's disease, multiple sclerosis, Alzheimer's disease,
211 or stroke. Further, participants were excluded if they reported a history

P
212 of Type II diabetes and cardiovascular disease. Also, participants had to Experiment 2 272
213 be free of any metal implants or other contraindications for MRI. In
214 order to participate, subjects needed to obtain consent from their Participant characteristics
D 273
215 physician to engage in an exercise intervention, as well as a maximal The participants were a pre-intervention cross-sectional subsample 274
216 graded exercise test (VO2 max). Finally, participants were required to from a 6-month single-blind randomized controlled exercise 275
217 be 60+ years of age and physically inactive as defined by engaging in trial (ClinicalTrials.gov; NCT01472744). Two hundred forty seven 276
E
218 30 minutes or less each week within the 6 months prior to baseline ex- community-dwelling healthy, older adults were randomized into four 277
219 amination (Erickson et al., 2011). All participants signed an informed intervention groups; baseline assessments from this intervention are 278
T

220 consent approved by the University of Illinois. reported here. Eligible participants met the following criteria: (1) were 279
between the ages of 60 and 80 years old, (2) were free from psychiatric 280
C

221 Cardiorespiratory fitness assessment and neurological illness and had no history of stroke or transient ischemic 281
222 Maximal oxygen uptake (VO2max) was used as an objective measure attack, (3) scored ≥23 on the Mini-Mental State Exam (MMSE) and N21 282
223 of cardiorespiratory fitness. As detailed by Voss et al. (2010b), on a Telephone Interview of Cognitive Status (TICS-M) questionnaire, 283
E

224 assessment of CRF was conducted using graded maximum exercise (4) scored b10 on the geriatric depression scale (GDS-15), (5) scored 284
225 testing on a motor-driven treadmill with continuous monitoring of ≥75% right-handedness on the Edinburgh Handedness Questionnaire, 285
R

226 respiration, heart rate, and blood pressure by a cardiologist and nurse (6) demonstrated normal or corrected-to-normal vision of at least 286
227 (Voss et al., 2010b). During the assessment, subjects walked at a speed 20/40 and no color blindness, (7) cleared for suitability in the MRI 287
228 slightly faster than their normal walking pace with increasing graded environment, that is, no metallic implants that could interfere with 288
R

229 increments of 2% every 2 minutes. Oxygen uptake was measured at 30 the magnetic field or cause injury, no claustrophobia, and no history 289
230 second intervals until a max VO2 was attained or to the point of test ter- of head trauma, and (8) reported no participation in exercise-related 290
O

231 mination due to exhaustion. VO2max was defined as the highest record- physical activity (maximum of two moderate bouts per week) in the 291
232 ed VO2 value when two of three criteria were satisfied: 1) a plateau in past 6 months. Therefore, this sample is described as low active and 292
233 VO2 peak between two or more workloads, 2) a respiratory exchange low fit, although capable of performing exercise (i.e., no physical dis- 293
C

234 ratio N 1.00 or 3) a heart rate equivalent to their age predicted maximum ability that prohibits mobility). In addition, all participants needed to 294
235 (i.e., 220-age). be cleared by their physician to participate in an exercise intervention 295
N

236 VO2max scores are expressed in units of milliliters per kilogram per and CRF testing. Out of 247 participants, 216 completed the baseline 296
237 minute (ml/kg/min). VO2peak was used as a proxy for VO2max among MRI imaging session with successful collection of DTI data. Of the 216 297
238 the ~ 20% of participants in each sample that did not meet two of the participants, 32 were excluded due to missing cognitive, VO2max, or 298
U

239 three criteria for VO2max. demographic data. An additional 27 participants were excluded due to 299
insufficient diffusion data quality including 1) suboptimal placement 300
240 Spatial working memory paradigm of the field-of-view resulting in insufficient brain coverage (n = 23) 301
241 The spatial memory paradigm was administered as part of a larger and 2) image artifacts that affected more than 2 volumes/scan (n = 302
242 battery of cognitive tests. However, as mentioned in the introduction, 4). Finally, 3 subjects scored b 24 on the Modified Mini Mental Status 303
243 we focused our hypotheses on the spatial memory task as it has been as- Exam-II, and were therefore excluded. Thus, the final sample consisted 304
244 sociated in previous work to fitness and hippocampal volume (Erickson of 154 participants. Those excluded from analysis did not differ from in- 305
245 et al., 2009, 2011) and resting connectivity (Voss et al., 2010b), and a cluded participants on demographic factors or fitness levels (all p N .10).306
246 similar task was used for assessing cognitive performance in Experi-
247 ment 2. At the beginning of the task, participants were shown a fixation
248 crosshair for 1 second, followed by the appearance of one, two or three Cardiorespiratory fitness assessment 307
249 dots placed in random locations on the screen for 500 ms. Then a 3 sec- As in Experiment 1, maximal oxygen uptake was used as an objec- 308
250 ond fixation crosshair appeared, during which time participants were tive measure of CRF. Similar methods as described in Experiment 1 309
251 asked to try to remember where the previous dot(s) were located. were used in this study (see Section 2.2). 310

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 4 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx

311 Spatial working memory paradigm reflects an average of young to middle aged adults. Therefore, in each 375
312 Participants completed a spatial working memory task similar to experiment, a study-specific template was created and was used as 376
313 that described in Experiment 1. At the start of the task, participants the target for registration. To create the study specific templates, we 377
314 were shown a fixation crosshair for 1 second, followed by the appear- first registered all native-space FA images to the FA template in MNI 378
315 ance of 2, 3, or 4 dots placed in random locations on the screen. Thus, space using an affine warp, then averaged the registered images across 379
316 the set sizes differed slightly between Experiment 1 (1, 2, and 3 dot subjects to generate the study-specific templates. Registration to the 380
317 trials) and Experiment 2 (2, 3, and 4 dot trials). Two and 3-dot trials study-specific template is done by combining two transformations: 381
318 were displayed for 500 milliseconds, and sets of 4 dots were displayed 1) a non-linear transformation of each subject's FA image to the study 382
319 for 1 second. Then, a 3 second fixation crosshair appeared, followed by specific template and 2) an affine registration of the template to 383
320 the appearance of a red dot, which displayed for 2 seconds. Similar to MNI152 standard space. Following the MNI transformation for all 384
321 Experiment 1, participants had to indicate whether the red dot subjects, a mean FA image was computed and an average skeleton 385
322 displayed was in the same location (match) or a different location was generated that represented major tracts common across partici- 386
323 (non-match) than one of the previously presented black dots, by press- pants. The skeleton was thresholded at an FA value of 0.2 (Smith et al., 387
324 ing a designated key on a computer keyboard. Participants completed a 2007) to ensure that major white matter tracts were included and to 388

F
325 total of 120 trials, 40 trials per set size (2, 3, and 4 black dots), each exclude regions that may contain multiple types of tissue. Then, in 389
326 containing 20 match and 20 non-match trials. Participants completed order to account for any residual misalignments not corrected for 390

O
327 12 practice trials (6 match, 6 non-match) prior to task administration during registration, each participant's normalized FA image was 391
328 to familiarize them with the task. Accuracy rates were recorded projected onto the mean FA skeleton. These images were then used in 392
329 separately for the 2, 3, and 4 dot conditions and were averaged to create the statistical analyses described below. 393

O
330 a mean accuracy score.
Statistical analyses 394
331 Diffusion tensor imaging We tested for differences across samples in age, gender, years of 395

R
332 Diffusion images were acquired on a 3 T Siemens Trio Tim system education, CRF, and spatial working memory performance using 396
333 with 45 mT/m gradients and 200 T/m/s slew rates (Siemens, Erlangen, chi-squared tests and two-tailed t-tests. Within each study sample, 397

P
334 Germany) with a twice-refocused spin echo single-shot Echo Planar Im- the relationships between CRF and nuisance variables were explored 398
335 aging sequence (Reese et al., 2003) to minimize eddy current-induced using independent samples t-tests and bivariate correlations. In 399
336 image distortions. The protocol consisted of a set of 30 non-collinear addition, the relationship between CRF and average spatial working
D 400
337 diffusion-weighted acquisitions with b-value = 1000 s/mm2 and memory accuracy was assessed using linear regression analyses. These 401
338 two T2-weighted b-value = 0 s/mm2 acquisitions, repeated two times analyses were conducted using SPSS (v. 21) (Spss, 2012). Sex, age, and 402
339 (TR/TE = 5500/98 ms, 128 × 128 matrix, 1.7 × 1.7 mm2 in-plane years of education were entered as covariates in the regression models 403
E
340 resolution, FA = 90, GRAPPA acceleration factor 2, and bandwidth of due to their association with the cognitive outcome variable. 404
341 1698 Hz/Px, comprising 40 3-mm-thick slices), obtained parallel to We tested the association between CRF and white matter 405
T

342 the anterior-posterior commissure plane with no interslice gap. Visual microstructure using a permutation-based regression tool within the 406
343 checks were performed on every volume of the raw data of every Bootstrap Regression Analysis of Voxelwise Observations (BRAVO) 407
C

344 participant by AZB. If excessive noise/field distortions were detected, toolbox. Documentation and tutorials for this toolbox are available at 408
345 up to 2 volumes/scan were removed with the corresponding b-vectors https://sites.google.com/site/bravotoolbox. For each regression model, 409
346 and b-values before further processing. a permutation approach evaluated statistical significance at each 410
E

voxel. A total of 500 permutation tests were performed at each voxel, 411
347 Image processing and for each iteration, the values in the model variables (covariates, 412
R

348 Identical image processing procedures were used on diffusion data CRF, FA) were independently scrambled. The significance of the 413
349 in Experiment 1 and Experiment 2. In both studies, diffusion data was association between CRF and fractional FA at each voxel (a path) was 414
350 processed using tools in the FMRIB Software Library (FSL v5.0.1) determined by comparing the distribution of the bootstrapped values 415
R

351 (Image Analysis Group, FMRIB, Oxford, UK; http://www.fmrib.ox.ac. with the distribution of the original values using a bias-corrected and 416
352 uk/fsl/; (Smith et al., 2004). Using FMRIB's Diffusion Toolbox (v.3.0; accelerated method (DiCiccio and Efron, 1996) (see Verstynen et al., 417
O

353 http://fmrib.ox.ac.uk/fsl/fdt/index.html), each participant's data was 2013; Gianaros et al., 2012 for more details). Clusters that contained 418
354 eddy current corrected by affine registration to the first Bo image. This k ≥ 20 contiguous voxels with a p threshold of b 0.05 after controlling 419
355 was followed by the removal of non-brain tissue using the Brain Extrac- for voxelwise multiple comparisons (false-discovery-rate) were 420
C

356 tion Tool (BET). Next, DTIfit was used to calculate the diffusion tensor at considered significant. 421
357 each voxel. Specifically, this step computes the voxelwise eigenvalues In the mediation model, we examined whether white matter 422
N

358 and eigenvectors of the diffusion tensor from each participant's image, integrity served as an indirect pathway through which CRF predicted 423
359 calculating various diffusion parameters, including FA. FA is a commonly cognitive performance after adjusting for nuisance variables including 424
360 used measure of white matter derived from DTI, and represents overall age, gender, and years of education. A regression approach with 425
U

361 anisotropy within a voxel (Jones et al., 2013). FA values fall between 0 permutation-based confidence interval estimation was employed to 426
362 and 1, indicating the degree of microstructural organization, with evaluate the indirect effect, with 500 permutation tests performed at 427
363 higher values indicating greater directionality of diffusion. each voxel. This resulted in a statistical parametric map of the indirect 428
364 FA data was fed into the FSL (v4.1.8) tract-based spatial statistics relationship between CRF and cognitive performance through 429
365 toolbox (TBSS; v1.2, http://www.fmrib.ox.ac.uk/fsl/tbss/index.html; voxelwise estimates of white matter microstructure (FA). In particular, 430
366 (Smith et al., 2006) pipeline. TBSS is used frequently in DTI processing, mean spatial working memory accuracy rate, created by calculating 431
367 and its algorithms for alignment of FA images across multiple subjects the average accuracy scores across the 3 set sizes in each study, was 432
368 into a standard space have been tested and validated (Smith et al., our dependent variable. We chose to examine average accuracy rather 433
369 2006). First, FA images were eroded to remove likely outliers from the than accuracy rates within each set size to 1) reduce the number of 434
370 diffusion tensor-fitting step. Then, FA images were normalized to multiple comparisons and challenges with interpreting multiple 435
371 1 × 1 × 1 mm MNI152 standard space via alignment to a common mediator models, as this approach would have required at least 6 436
372 registration target. As considerable atrophy occurs in older adulthood, separate mediation models and 2) accuracy rates for set sizes were highly 437
373 it is standard to compute a study specific template when using older correlated with each other in both samples (Experiment 1: r N 0.730 for 438
374 adult populations, as the standard FSL FA template (FMRIB58_FA) correlations between set sizes; p b .001; Experiment 2: r N 0.754 for 439

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx 5

440 correlations between set sizes; p b .001). In order to correct for multiple 2. Identical analytical approaches were applied in Experiment 1 and 447
441 comparisons, we used a false-discovery-rate at the voxel-level threshold Experiment 2. 448
442 of 0.05 and isolated clusters using a spatial extent threshold of k ≥ 20 con-
443 tiguous voxels. Finally, we overlaid the Johns Hopkins University atlas to Results 449
444 the voxel-based results to provide anatomical labels and to extract FA
445 values for purposes of plotting. FA values for significant clusters within A series of two-tailed t-tests compared age, education, fitness, and 450
446 these regions were used to create the scatterplots shown in Figs. 1 and cognitive performance data across the two samples. These analyses 451

F
O
O
R
P
D
E
T
C
E
R
R
O
C
N
U

Fig. 1. Shown are significant associations between cardiorespiratory fitness (CRF), fractional anisotropy (FA), and mean spatial working memory performance from Experiment 1.
(A) Clusters of voxels where fitness was significantly associated with FA. Warm-colored voxels show positive associations, cool-colored voxels demonstrate negative associations. Side
panels indicate slice placements. Age, gender, and years of education were included as covariates. Clusters are thresholded to k ≥ 20 contiguous voxels and a false discovery rate of
0.05. The z-plane coordinates of each slice, in MNI space, are presented at the bottom. For visualization purposes, tbss_fill was used to dilate statistical maps. (B) For illustration purposes,
scatterplots and best-fit lines for the relationship between CRF and FA in selected regions. (C) Spatial distribution of indirect path voxel clusters. Cyan-colored voxel clusters show a positive
indirect effect, brown voxel clusters indicate a negative indirect effect. Ant, anterior; IC, internal capsule; CC, corpus callosum; Ext, external; FA, fractional anisotropy.

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 6 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx

F
O
O
R
P
D
E
T
C
E
R
R
O
C
N
U

Fig. 2. Shown are associations between cardiorespiratory fitness (CRF), fractional anisotropy (FA), and spatial working memory performance from Experiment 2, with the same plotting
conventions as Fig. 1. Ant, anterior; FA, fractional anisotropy; CC, corpus callosum; IC, internal capsule; CR, corona radiata.

452 revealed a significant difference in CRF between Experiment 1 Experiment 1 460

453 (VO2max mean = 21.40) and Experiment 2 (VO2max mean = 20.17)
454 (t = − 2.139; p = .033). There were no significant differences in Fitness and cognitive performance 461
455 age, years of education, or cognitive performance (average accuracy) Experiment 1 comprised of 113 low-active older adults (mean age = 462
456 between the two samples (all p N .10). In addition, there was a 66.61; 36.3% male). Demographic information and average accuracy for 463
457 similar distribution of males and females across samples (χ2 (1) = the spatial memory task are presented in Table 1. CRF differed by gender 464
458 0.767; p = 0.381). Results are presented separately for each sample, in this sample, with men (mean (SD) VO2max = 24.22 (5.27)) 465
459 below. demonstrating higher average fitness levels relative to women (mean 466

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx 7

t1:1 Table 1 CRF, white matter microstructure, and cognition 496

t1:2 Demographic characteristics, CRF, and task performance for Experiment 1 and Experiment To determine whether white matter microstructure served as an 497
t1:3 2.
indirect mediating pathway between CRF and spatial working memory 498
t1:4 Experiment 1 Experiment 2 performance, we employed a permutation-based regression approach. 499
(n = 113) (n = 154) Consistent with our prediction, mediation analysis showed significant 500
t1:5 Age (years) 66.61 (5.65) 65.66 (4.55) indirect associations between CRF and spatial working memory perfor- 501
t1:6 Gender (% male) 36.30% 31.20% mance through distributed white matter regions, highlighted in cyan in 502
t1:7 Education (years) 15.48 (2.92) 15.62 (2.85)
Fig. 1C. These regions included clusters in the genu and body of the cor- 503
t1:8 Cardiorespiratory fitness (VO2max, ml/kg/min) 21.40 (4.89)a 20.17 (4.49)
t1:9 Spatial working memory accuracy (%) 81.81 (12.86) 81.02 (11.66) pus callosum, bilateral anterior corona radiata, superior corona radiata, 504
and anterior internal capsule. Effects going in the positive direction 505
t1:10 Values are mean (standard deviation).
a indicate that higher levels of CRF were associated with greater FA in 506
t1:11 Indicates a significant difference across studies at p b .05
these clusters, which in turn, were associated with better performance 507
(fewer errors) on the spatial working memory task. Only one small 508
467 (SD) VO2max = 19.79 (3.85)) (t = −4.71; p b .01). Those with higher cluster in the right posterior white matter demonstrated a negative 509

F
468 fitness levels also tended to be younger (r = − 0.460; p b .01) and indirect effect. 510
469 have greater years of education (r = 0.291; p b .01) (Table 2). In

O
470 addition, older age was correlated with poorer spatial working memory
Experiment 2 511
471 accuracy, even after adjusting for gender and years of education
472 (r = −0.318; p = .001).

O
Fitness and cognitive performance 512
473 After controlling for age, gender, and years of education, there was
474 not a significant association between CRF and spatial working memory Experiment 2 included 154 low-active older adults (mean age = 513
65.66; 31.2% male). Subject demographics along with fitness and 514

R
475 performance in this sample (β = 0.171; p = 0.151) (Table 2). Due to
476 the differences in set size on the spatial working memory task across cognitive performance data can be found in Table 1. Whereas bivariate 515
correlations showed no association between CRF and years of education 516
477 samples, a second average accuracy score was calculated using the

P
478 2-dot and 3-dot conditions, excluding the 1-dot condition. Follow-up (r = 0.107; p = 0.187), age was inversely correlated with CRF in this 517
479 analysis with this measure revealed a trending, but non-significant, sample (r = − 0.247; p b .01). In addition, an independent samples 518
480 association between CRF and average accuracy on the 2-dot and 3-dot t-test revealed that females (mean (SD) VO2peak = 19.32 (3.97)) tended 519
to have lower fitness levels relative to males (mean (SD) VO2peak = 520
481 conditions (β = 0.228; p = 0.052).
D
22.03 (5.02)) (t = − 3.292; p b .01) (Table 2). In addition, age was 521
inversely correlated with accuracy rates (r = −0.210; p = .009) after 522
E
482 Cardiorespiratory fitness is associated with white matter microstructure adjusting for gender and years of education. 523
483 Consistent with our hypotheses, a voxelwise analysis revealed that After accounting for age, gender, and years of education, multilevel 524
T

484 higher CRF was associated with greater FA values across multiple linear regression analysis revealed a significant association between 525
485 regions of the white matter skeleton. These associations remained CRF and spatial working memory performance, such that higher CRF 526
levels predicted better accuracy on the task (β = 0.237; p = 0.006) 527
C

486 significant after adjusting for age, gender, and years of education.
487 Regions in which significant associations between CRF and FA were (Table 2). 528
488 located include the genu, body, and splenium of the corpus callosum;
E

489 the fornix; bilateral anterior corona radiata, and bilateral anterior inter- Cardiorespiratory fitness is associated with white matter microstructure 529
490 nal capsule (PFDR-corrected b 0.05). A statistical map of the voxels within Consistent with Experiment 1, we found that higher CRF was associ- 530
R

491 the white matter skeleton that were associated with CRF can be found ated with higher FA within multiple white matter regions, even after 531
492 in Fig. 1A. For visualization purposes, the relationship between CRF controlling for age, gender, and education. Fig. 2A shows the spatial 532
493 and FA within several post hoc ROIs is plotted in Fig. 1B. One small
R

distribution of statistically significant positive (red) and negative 533

494 cluster in the right posterior thalamic radiation demonstrated a nega- (blue) clusters. Among the regions containing significant clusters are 534
495 tive relationship between fitness and CRF. the genu, body, and splenium of the corpus callosum, the left cingulum, 535
O

bilateral segments of the superior longitudinal fasciculus, the 536

anterior internal capsule, and the superior and anterior corona radiata 537
(PFDR-corrected b 0.05). A few small clusters showed a negative relation- 538
C

t2:1 Table 2 ship between CRF and FA when controlling for covariates, with the 539
t2:2 Relationships between fitness, demographic characteristics, and task performance for Ex-
largest clusters found bilaterally in the posterior limb of the internal 540
N

t2:3 periments 1 and 2.

capsule. 541
t2:4 Experiment 1 Experiment 2
(n = 113) (n = 154)
U

t2:5 Agea −0.460d −0.247d Fitness, white matter microstructure, and cognition 542
t2:6 (r, p value) b .001 .002 After controlling for age, gender, and education, mediation analyses 543
t2:7 Genderb −4.710d −3.292d revealed a large number of clusters distributed throughout white 544
t2:8 (t, p value) b .001 .002 matter voxels that were indirect pathways between CRF and spatial 545
t2:9 Educationa 0.291d 0.1070.187
working memory performance. Clusters showing a positive indirect 546
t2:10 (r, p value) .002
t2:11 Spatial working memory accuracyc 0.1710.151 0.237d association are highlighted in cyan in Fig. 2C, and include clusters within 547
t2:12 (β; p value) 0.006 the genu of the corpus callosum and the left cingulum segment. The 548
t2:13 a
Results of bivariate correlations examining the association between fitness and each positive values observed in these regions suggest that CRF is associated 549
t2:14 variable. Values are Pearson's r; p-value. with greater FA in these clusters, which, in turn, are associated with 550
b
t2:15 Results from independent samples t-test demonstrating the difference in VO2max be- greater accuracy on the spatial working memory task. The distribution 551
t2:16 tween males and females. Values are t; p-value. Females were coded 0; men coded 1. of voxels showing significant negative indirect path effects is highlight- 552
c
t2:17 Results from multilevel linear regression analysis of the relationship between fitness
ed in brown in Fig. 2C. Aside from bilateral clusters in the most rostral 553
t2:18 and spatial working memory performance after adjusting for age, gender, and years of ed-
t2:19 ucation. Values are β; p-value. segment of the anterior thalamic radiation, negative indirect path 554
t2:20 d
Indicates significance at p b .01. voxels were primarily localized in the posterior white matter. 555

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 8 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx

556 Conjunction analyses right anterior corona radiata, right anterior internal capsule, and 569
557 A post-hoc conjunction analysis was used to identify overlapping re- bilateral sagittal stratum (Fig. 3B). Thus, in Experiment 1 and Experi- 570
558 gions associated with CRF across Experiment 1 and Experiment 2. For ment 2, higher CRF was related to greater FA in the aforementioned 571
559 this analysis, the results from the association with CRF in Experiment regions, which, in turn, were associated with better spatial memory 572
560 1 were overlaid with the results from the association between CRF performance. 573
561 and FA in Experiment 2. Then, a white matter atlas was used to identify
562 regions in which overlap was observed. Regions containing clusters Discussion 574
563 commonly associated with fitness across samples included the body
564 and splenium of the corpus callosum, the fornix, bilateral anterior coro- Higher CRF has been associated with greater gray matter volume in 575
565 na radiata, anterior internal capsule, and sagittal stratum (Fig. 3A). the prefrontal cortex (Colcombe et al., 2006; Weinstein et al., 2012) and 576
566 A similar method was employed to examine common indirect path hippocampus (Erickson et al., 2009, 2011) as well as greater task- 577
567 voxels across the two samples. Regions of significant overlap between evoked and intrinsic resting functional connectivity between 578
568 the two samples included the body of the corpus callosum, fornix, hippocampal and frontal regions (Burdette et al., 2010; Smith et al., 579

F
O
O
R
P
D
E
T
C
E
R
R
O
C
N
U

Fig. 3. A. Shown are results from the conjunction analysis of the relationship between fitness and FA in Experiment 1 and Experiment 2. B. Shown are the results from the conjunction
analysis of the indirect path voxels in Experiment 1 and Experiment 2.

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx 9

580 2012; Voss et al., 2010a) in older adults. Consistent with this literature, domains have been more strongly and consistently correlated with 646
581 here we found in two large, independent samples of older adults, that fiber architecture, including processing speed and executive function 647
582 greater CRF was associated with greater FA across a diverse network (Bennett and Madden, 2014; Grieve et al., 2007; Madden et al., 2009; 648
583 of white matter tracts. White matter that was significantly associated Vernooij et al., 2009). Although this is an important first step in explor- 649
584 with CRF across both samples included tracts that facilitate ing these relationships, our results may not reflect white matter tracts 650
585 intrahemispheric communication between the medial temporal lobe that may mediate the relationship between fitness and other cognitive 651
586 and the prefrontal cortex, but also included a distributed network of domains including processing speed, verbal memory, and attentional 652
587 regions throughout the brain. These results are important as they dem- control processes. Nonetheless, our results suggest widespread relation- 653
588 onstrate that the benefits of higher CRF may be relatively widespread ships between CRF and white matter microstructure, and that these re- 654
589 across many areas of white matter, but with some regional specificity lationships constitute important indirect associations with spatial 655
590 to areas that support communication between prefrontal and medial working memory performance. 656
591 temporal lobe regions. Further, consistent with our prediction, we In contrast with our hypotheses, as well as the results from Experi- 657
592 found that white matter microstructure constituted an indirect path be- ment 2, there was not a direct association between fitness and cognitive 658
593 tween CRF and spatial working memory performance. This indicates performance in Experiment 1. This may, in part, be due to subtle differ- 659

F
594 that white matter plays an important role in elevated spatial memory ences in the spatial working memory task across samples. Using an av- 660
595 performance in older adults. erage accuracy score including only the 2-dot and 3-dot conditions, 661

O
596 The associations we report here between CRF and white matter are which is more consistent with the set size in Experiment 2, strength- 662
597 consistent with a growing literature on the beneficial associations of ened the association, although it remained non-significant. In addition, 663
598 PA, exercise, and fitness with white matter structure. Prior studies Experiment 2 had a larger sample size and thus may have been more ad- 664

O
599 have reported that greater amounts of PA (Gons et al., 2013; Gow equately powered to detect an association. Notably, a direct relationship 665
600 et al., 2012; Tian et al., 2014a), greater fitness levels (Hayes et al., between the independent and predictor variable is not required for a 666
2015; Johnson et al., 2012; Marks et al., 2010), and changes in fitness

R
601 significant indirect pathway to be present (Hayes, 2009; Preacher and 667
602 levels after participation in an exercise intervention (Voss et al., Hayes, 2004). For instance, in Experiment 1, it may be that fitness is re- 668
603 2013b) are positively associated with white matter microstructure. lated to spatial working memory performance through its association 669

P
604 Results from the present study complement the associations observed with white matter microstructure, which in turn is associated with bet- 670
605 in earlier work, including some overlap in the regions identified. For in- ter task performance. 671
606 stance, consistent with previous reports, we observed associations be- D Also in contrast with our hypotheses, we found several negative re- 672
607 tween fitness and FA in the cingulum bundle (Marks et al., 2010; Tian lationships between CRF, FA, and spatial working memory performance. 673
608 et al., 2014b) and segments of the corpus callosum (Johnson et al., In both studies, negative associations between CRF and FA were present, 674
609 2012). Our findings also expand these associations to other white albeit negligible, and limited to a few small clusters. The bulk of the 675
E
610 matter tracts not previously identified including the anterior internal negative associations, however, emerged from the mediation model. 676
611 capsule, fornix, and anterior corona radiata. These differences may be As a majority of the CRF-FA associations were in the positive direction, 677
T

612 due to variations in sample characteristics, fitness measures, and analyt- the comparably larger presence of negative associations in the indirect 678
613 ical approaches between this work and others. For example, many of the paths suggests that this may be driven by negative associations between 679
C

614 prior studies examining associations with CRF have had small sample white matter integrity and spatial working memory performance. This 680
615 sizes (Hayes et al., 2015; Johnson et al., 2012; Marks et al., 2010; is an interesting trend, and we can only speculate about why this 681
616 Tseng et al., 2013) or have examined relationships using a priori regions might be occurring. It may be that greater white matter microstructural 682
E

617 of interest (Marks et al., 2010; Tian et al., 2014a), which provides only a integrity in some paths is associated with a particular strategy prefer- 683
618 limited understanding of these relationships across neural circuits ence that is suboptimal. Functional MRI studies have shown that older 684
R

619 throughout the brain. In addition, some previous studies have explored adults tend to differ both in terms of regional specificity and level of re- 685
620 these associations in older adult samples with various conditions gional recruitment relative to younger adults during neurocognitive 686
621 known to affect microstructural white matter integrity including tasks (Huang et al., 2012; Kennedy et al., 2015). Further, these age- 687
R

622 cerebral small vessel disease (Gons et al., 2013), multiple sclerosis related differences in neural recruitment patterns have been linked to 688
623 (Prakash et al., 2010), and chronic conditions including cardiovascular poorer task performance in older adults (Zhu et al., 2013), suggesting 689
O

624 disease and diabetes (Tian et al., 2014a, 2014b). Therefore, the relation- that effective versus non-effective strategies are related to both 690
625 ship between fitness and white matter in these populations may differ variation in performance and variation in neural circuits associated 691
626 in strength and regional-specificity relative to healthy older adult with performance. Similarly, greater white matter microstructure in 692
C

627 samples. In order to directly address these limitations, we conducted a certain brain areas may facilitate particular strategies that are associated 693
628 voxelwise analysis of white matter in two large samples (N = 113; with less effective task performance. 694
N

629 N = 154) of healthy older adults, without neurological or cardiovascular Despite these general conclusions, there were some important 695
630 conditions. qualitative differences between the results from the two samples in 696
631 Further, few studies have examined whether fitness-related our study. While both studies demonstrated widespread associations 697
U

632 variation in white matter microstructure translates to differences in between fitness and FA in multiple neural pathways, the results from 698
633 cognitive performance. Across the two samples, we found indirect asso- these associations in Experiment 1 were more focused in anterior 699
634 ciations between fitness and spatial working memory performance in segments (e.g., genu) than Experiment 2, which were more diffuse 700
635 various white matter tracts including the genu of the corpus callosum and consistently spread throughout white matter (see Figs. 1A and 701
636 and the anterior corona radiata. This is consistent with previous work 2A). Further, as seen in the scatter plots of Figs. 1B and 2B, the variability 702
637 demonstrating associations between performance on similar working in FA in Experiment 2 was greater than that found in Experiment 1 and 703
638 memory tasks and FA in the genu as well as the anterior corona radiate the mediation results with spatial memory performance were also more 704
639 (Kennedy and Raz, 2009). We selected to use spatial working memory diffuse in Experiment 2 than in Experiment 1. The differences in 705
640 as our outcome variable because 1) spatial working memory processes variability may be a consequence of the fact that diffusion data were 706
641 decline with age (Schmiedek et al., 2009), 2) this task has been found collected using 12 gradient directions in Experiment 1 and 30 gradient 707
642 to be sensitive to fitness (Erickson et al., 2009) and exercise effects directions in Experiment 2, with the latter providing a more reliable 708
643 (Erickson et al., 2011) and 3) to ensure consistency across samples, as tensor estimate. Although we did not conduct a quantitative compari- 709
644 this task was administered in both Experiments. While memory perfor- son between the two samples, there are several factors that could 710
645 mance has been linked to white matter fiber integrity, other cognitive explain some of these differences. For example, there were significant 711

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 10 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx

712 differences between the samples in fitness levels, with subjects in A final limitation is that our MR parameters between the two samples 778
713 Experiment 2 having a slightly lower mean VO2max than the participants were quite different which limits our ability to validly conduct direct 779
714 in Experiment 1. Although the effects we report appear linear, there quantitative comparisons between samples. As the field of white matter 780
715 could be regional differences in the degree of linearity with some imaging is quickly evolving, it will be important for studies to consider 781
716 areas becoming less associated with CRF at lower fitness levels. In addi- the consistency of the parameters to more accurately compare results 782
717 tion, the MR sequence parameters and the machine type were different across time and samples. 783
718 between studies as were some of the general demographic characteris- In summary, we find that higher CRF levels in two separate samples 784
719 tics (e.g., slightly younger sample in Experiment 2). Furthermore, the of older adults are positively related to white matter in widely distribut- 785
720 spatial memory task used in both experiments was very similar, but ed neural networks and that many of these associations statistically 786
721 differed in one key aspect: in Experiment 1 the task used a set size of mediate the association between CRF and spatial working memory 787
722 1, 2, or 3 to-be-remembered locations whereas in Experiment 2 the performance. Remaining aerobically fit might be an important approach 788
723 set sizes ranged from 2, 3, or 4 to-be-remembered locations. Such a for maintaining white matter health and working memory function into 789
724 difference was not associated with differences in performance, but late adulthood. 790
725 could have contributed to strategy differences or the extent to which

F
726 different brain areas supported task performance. Nonetheless, despite Funding 791
727 these qualitative differences between the results shown in Figs. 1

O
728 and 2, the results from both samples clearly demonstrate that higher This work was supported by the National Institute on Aging (NIA) at 792
729 CRF was associated with greater white matter microstructural integrity, the National Institutes of Health (NIH) grants RO1 AG25667, RO1 793
730 which was, in turn, related to better spatial working memory AG25032 and R37-AG025667 awarded to A.F.K. and E.M. In addition, 794

O
731 performance. this work was funded by NIH grants RO1 DK095172 and P30 795
732 We can only speculate about the physiological mechanisms underly- AG024827 awarded to K.I.E. This work was also funded by a grant 796
733 ing these associations. In studies of gray matter volume, changes in

R
from Abbott Nutrition through the Center of Nutrition, Learning and 797
734 hippocampal volume after participation in an exercise intervention Memory at the University of Illinois. This funder had no role in study 798
735 were correlated with changes in serum brain-derived neurotrophic design, data collection and analysis, decision to publish, or preparation 799

P
736 factor (BDNF) levels (Erickson et al., 2011). Other studies have also of the article. 800
737 identified BDNF as an important potential mediator of resting state
738 changes (Voss et al., 2013a), and these results in humans have been sup-
Acknowledgments 801
739 ported by non-human animal research showing increased expression of
D
740 BDNF in the hippocampus and cortex after several weeks of exercise
741 (Gómez-Pinilla et al., 2002; Vaynman et al., 2004). Rodent work has The authors thank Tammy Chen, Anamarta Lamoutte, Vineet 802
E
742 also identified associations between BDNF and brain white matter. For Agarwal, and Chanheng He for their assistance quality checking the 803
743 example, a recent study found that administration of BDNF predicted imaging data for Experiment 2. 804
T

744 remyelination following induced subcortical damage in a rodent

745 model of ischemic stroke (Ramos-Cejudo et al., 2015). These findings References 805
C

746 are complemented by recent advancements in the human literature, in-

Bennett, I.J., Madden, D.J., 2014. Disconnected aging: cerebral white matter integrity and 806
747 cluding a longitudinal study of healthy older adults, which found that age-related differences in cognition. Neuroscience 276, 187–205. 807
748 lower levels of plasma BDNF were associated with a steeper decline in Burdette, J.H., Laurienti, P.J., Espeland, M.A., Morgan, A., Telesford, Q., Vechlekar, C.D., 808
E

749 white matter volume among females (Driscoll et al., 2012). However, Hayasaka, S., Jennings, J.M., Katula, J.A., Kraft, R.A., 2010. Using network science to eval- 809
uate exercise-associated brain changes in older adults. Front. Aging Neurosci. 2, 23. 810
750 other mechanisms are also likely at play. For example, exercise reduces Burzynska, A.Z., Preuschhof, C., Bäckman, L., Nyberg, L., Li, S.-C., Lindenberger, U., 811
R

751 pro-inflammatory cascades, which have been shown to negatively Heekeren, H.R., 2010. Age-related differences in white matter microstructure: 812
752 impact white matter microstructure (Wersching et al., 2010). Fitness- region-specific patterns of diffusivity. NeuroImage 49, 2104–2112. 813
Burzynska, A.Z., Chaddock-Heyman, L., Voss, M.W., Wong, C.N., Gothe, N.P., Olson, E.A., 814
753 related associations with insulin pathways, gene expression patterns,
R

Knecht, A., Lewis, A., Monti, J.M., Cooke, G.E., 2014. Physical activity and cardiorespi- 815
754 or cerebrovasculature may also be underlying mechanisms for the ratory fitness are beneficial for white matter in low-fit older adults. PLoS One 9, 816
755 effects we report here. e107413. 817
818
O

Charlton, R., Barrick, T., McIntyre, D., Shen, Y., O'Sullivan, M., Howe, F.e., emsp14, al, Clark,
756 There are several limitations to our study. First, we report results
C., Morris, R., Markus, H., 2006. White matter damage on diffusion tensor imaging 819
757 from a cross-sectional analysis of fitness and white matter. Currently correlates with age-related cognitive decline. Neurology 66, 217–222. 820
758 there is an absence of studies examining white matter changes resulting Colcombe, S., Kramer, A.F., 2003. Fitness effects on the cognitive function of older adults a 821
C

759 from randomized trials of exercise (however see Voss et al., 2013b). meta-analytic study. Psychol. Sci. 14, 125–130. 822
Colcombe, S.J., Erickson, K.I., Scalf, P.E., Kim, J.S., Prakash, R., McAuley, E., Elavsky, S., 823
760 Thus, we cannot determine a causal association between increased Marquez, D.X., Hu, L., Kramer, A.F., 2006. Aerobic exercise training increases brain 824
aerobic fitness and white matter. Both of the studies reported here
N

761 volume in aging humans. J. Gerontol. Ser. A Biol. Med. Sci. 61, 1166–1170. 825
762 were baseline assessments from two randomized trials of exercise, DiCiccio, T.J., Efron, B., 1996. Bootstrap confidence intervals. Stat. Sci. 189–212. 826
Driscoll, I., Martin, B., An, Y., Maudsley, S., Ferrucci, L., Mattson, M.P., Resnick, S.M., 2012. 827
763 with the results from Experiment 2 still being collected and analyzed. Plasma BDNF is associated with age-related white matter atrophy but not with 828
U

764 Future reports will be able to more adequately describe the changes in cognitive function in older, non-demented adults. PLoS One 7 (e35217-e35217). 829
765 white matter microstructure resulting from the intervention. Both Erickson, K.I., Prakash, R.S., Voss, M.W., Chaddock, L., Hu, L., Morris, K.S., White, S.M., 830
Wójcicki, T.R., McAuley, E., Kramer, A.F., 2009. Aerobic fitness is associated with 831
766 samples consisted of healthy, cognitively intact older adults that were hippocampal volume in elderly humans. Hippocampus 19, 1030–1039. 832
767 physically inactive at the time of baseline assessment. Due to these Erickson, K.I., Voss, M.W., Prakash, R.S., Basak, C., Szabo, A., Chaddock, L., Kim, J.S., Heo, S., 833
768 inclusion criteria, it is unclear whether similar results would be found Alves, H., White, S.M., 2011. Exercise training increases size of hippocampus and 834
improves memory. Proc. Natl. Acad. Sci. 108, 3017–3022. 835
769 in either a more impaired or more fit older adult sample. These results
Erickson, K.I., Leckie, R.L., Weinstein, A.M., 2014. Physical activity, fitness, and gray matter 836
770 might be more robust in a more physically active sample with a greater volume. Neurobiol. Aging 35, S20–S28. 837
771 range of CRF. Further, genetic factors play a role in determining fitness Gianaros, P.J., Marsland, A.L., Sheu, L.K., Erickson, K.I., Verstynen, T.D., 2012. Inflammatory 838
772 levels, and it is impossible in our study to rule out the contribution of pathways link socioeconomic inequalities to white matter architecture. Cereb. Cortex 839
23, 2058–2071. 840
773 genetic variants compared to the contribution of physical activity in Gold, B.T., Powell, D.K., Xuan, L., Jicha, G.A., Smith, C.D., 2010. Age-related slowing of task 841
774 determining fitness. Additionally, the diffusion data in Experiment 1 switching is associated with decreased integrity of frontoparietal white matter. 842
775 was collected using only 12 gradient directions, which limits the preci- Neurobiol. Aging 31, 512–522. 843
Gómez-Pinilla, F., Ying, Z., Roy, R.R., Molteni, R., Edgerton, V.R., 2002. Voluntary exercise 844
776 sion of the tensor estimation. Further, a FLAIR sequence was not run in induces a BDNF-mediated mechanism that promotes neuroplasticity. J. Neurophysiol. 845
777 either sample so we could not control for white matter hyperintensities. 88, 2187–2195. 846

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053
 L.E. Oberlin et al. / NeuroImage xxx (2015) xxx–xxx 11

847 Gons, R.A., Tuladhar, A.M., de Laat, K.F., van Norden, A.G., van Dijk, E.J., Norris, D.G., Zwiers, Smith, S.M., Johansen-Berg, H., Jenkinson, M., Rueckert, D., Nichols, T.E., Miller, K.L., 920
848 M.P., de Leeuw, F.-E., 2013. Physical activity is related to the structural integrity of Robson, M.D., Jones, D.K., Klein, J.C., Bartsch, A.J., 2007. Acquisition and voxelwise 921
849 cerebral white matter. Neurology 81, 971–976. analysis of multi-subject diffusion data with tract-based spatial statistics. Nat. Protoc. 922
850 Gow, A.J., Bastin, M.E., Maniega, S.M., Hernández, M.C.V., Morris, Z., Murray, C., Royle, N.A., 2, 499–503. 923
851 Starr, J.M., Deary, I.J., Wardlaw, J.M., 2012. Neuroprotective lifestyles and the aging Smith, J.C., Nielson, K.A., Antuono, P., Lyons, J.-A., Hanson, R.J., Butts, A.M., Hantke, N.C., 924
852 brain Activity, atrophy, and white matter integrity. Neurology 79, 1802–1808. Verber, M.D., 2012. Semantic memory functional MRI and cognitive function after 925
853 Grieve, S., Williams, L., Paul, R., Clark, C., Gordon, E., 2007. Cognitive aging, executive exercise intervention in mild cognitive impairment. J. Alzheimers Dis. 37, 197–215. 926
854 function, and fractional anisotropy: a diffusion tensor MR imaging study. Am. Spss, I., 2012. IBM SPSS statistics version 21. International Business Machines Corp., 927
855 J. Neuroradiol. 28, 226–235. Boston, Mass. 928
856 Hayes, A.F., 2009. Beyond Baron and Kenny: statistical mediation analysis in the new Stern, Y., Sano, M., Paulson, J., Mayeux, R., 1987. Modified mini-mental state examination: 929
857 millennium. Commun. Monogr. 76, 408–420. validity and reliability. Neurology 37, 179. 930
858 Hayes, S.M., Salat, D.H., Forman, D.E., Sperling, R.A., Verfaellie, M., 2015. Cardiorespiratory Tian, Q., Erickson, K.I., Simonsick, E.M., Aizenstein, H.J., Glynn, N.W., Boudreau, R.M., 931
859 fitness is associated with white matter integrity in aging. Ann. Clin. Transl. Neurol. 2, Newman, A.B., Kritchevsky, S.B., Yaffe, K., Harris, T.B., 2014a. Physical activity predicts 932
860 688–698. microstructural integrity in memory-related networks in very old adults. J. Gerontol. 933
861 Huang, C.-M., Polk, T.A., Goh, J.O., Park, D.C., 2012. Both left and right posterior parietal Ser. A Biol. Med. Sci. 69, 1284–1290. 934
862 activations contribute to compensatory processes in normal aging. Neuropsychologia Tian, Q., Simonsick, E.M., Erickson, K.I., Aizenstein, H.J., Glynn, N.W., Boudreau, R.M., 935
863 50, 55–66. Newman, A.B., Kritchevsky, S.B., Yaffe, K., Harris, T., 2014b. Cardiorespiratory fitness 936
864 Johnson, N.F., Kim, C., Clasey, J.L., Bailey, A., Gold, B.T., 2012. Cardiorespiratory fitness is and brain diffusion tensor imaging in adults over 80 years of age. Brain Res. 1588, 937
865 938

F
positively correlated with cerebral white matter integrity in healthy seniors. 63–72.
866 NeuroImage 59, 1514–1523. Tian, Q., Glynn, N.W., Erickson, K.I., Aizenstein, H.J., Simonsick, E.M., Yaffe, K., Harris, T.B., 939
867 Jones, D.K., Knösche, T.R., Turner, R., 2013. White matter integrity, fiber count, and other Kritchevsky, S.B., Boudreau, R.M., Newman, A.B., 2015. Objective measures of physical 940

O
868 fallacies: the do's and don'ts of diffusion MRI. NeuroImage 73, 239–254. activity, white matter integrity and cognitive status in adults over age 80. Behav. 941
869 Kennedy, K.M., Raz, N., 2009. Aging white matter and cognition: differential effects of Brain Res. 284, 51–57. 942
870 regional variations in diffusion properties on memory, executive functions, and Tseng, B., Gundapuneedi, T., Khan, M., Diaz-Arrastia, R., Levine, B., Lu, H., Huang, H., Zhang, 943
871 speed. Neuropsychologia 47, 916–927. R., 2013. White matter integrity in physically fit older adults. NeuroImage 82, 944

O
872 Kennedy, K.M., Rodrigue, K.M., Bischof, G.N., Hebrank, A.C., Reuter-Lorenz, P.A., Park, D.C., 510–516. 945
873 2015. Age trajectories of functional activation under conditions of low and high Vaynman, S., Ying, Z., Gomez‐Pinilla, F., 2004. Hippocampal BDNF mediates the efficacy of 946
874 processing demands: an adult lifespan fMRI study of the aging brain. NeuroImage exercise on synaptic plasticity and cognition. Eur. J. Neurosci. 20, 2580–2590. 947

R
875 104, 21–34. Vernooij, M.W., Ikram, M.A., Vrooman, H.A., Wielopolski, P.A., Krestin, G.P., Hofman, A., 948
876 Madden, D.J., Spaniol, J., Costello, M.C., Bucur, B., White, L.E., Cabeza, R., Davis, S.W., Niessen, W.J., Van der Lugt, A., Breteler, M.M., 2009. White matter microstructural 949
877 Dennis, N.A., Provenzale, J.M., Huettel, S.A., 2009. Cerebral white matter integrity me- integrity and cognitive function in a general elderly population. Arch. Gen. Psychiatry 950
878 951

P
diates adult age differences in cognitive performance. J. Cogn. Neurosci. 21, 289–302. 66, 545–553.
879 Marks, B., Katz, L., Styner, M., Smith, J., 2010. Aerobic fitness and obesity: relationship to Verstynen, T.D., Weinstein, A., Erickson, K.I., Sheu, L.K., Marsland, A.L., Gianaros, P.J., 2013. 952
880 cerebral white matter integrity in the brain of active and sedentary older adults. Br. Competing physiological pathways link individual differences in weight and 953
881
Q4 J. Sports Med. abdominal adiposity to white matter microstructure. NeuroImage 79, 129–137. 954
882 Niemann, C., Godde, B., Voelcker-Rehage, C., 2014. Not only cardiovascular, but also coor- Voelcker-Rehage, C., Godde, B., Staudinger, U.M., 2011. Cardiovascular and coordination
D 955
883 dinative exercise increases hippocampal volume in older adults. Front. Aging training differentially improve cognitive performance and neural processing in 956
884 Neurosci. 6, 170. older adults. Front. Hum. Neurosci. 5, 26. 957
885 Pfefferbaum, A., Sullivan, E.V., 2003. Increased brain white matter diffusivity in normal Voineskos, A.N., Rajji, T.K., Lobaugh, N.J., Miranda, D., Shenton, M.E., Kennedy, J.L., Pollock, 958
E
886 adult aging: relationship to anisotropy and partial voluming. Magn. Reson. Med. 49, B.G., Mulsant, B.H., 2012. Age-related decline in white matter tract integrity and 959
887 953–961. cognitive performance: a DTI tractography and structural equation modeling study. 960
888 Prakash, R.S., Snook, E.M., Motl, R.W., Kramer, A.F., 2010. Aerobic fitness is associated with Neurobiol. Aging 33, 21–34. 961
T

889 gray matter volume and white matter integrity in multiple sclerosis. Brain Res. 1341, Voss, M.W., Erickson, K.I., Prakash, R.S., Chaddock, L., Malkowski, E., Alves, H., Kim, J.S., 962
890 41–51. Morris, K.S., White, S.M., Wójcicki, T.R., 2010a. Functional connectivity: a source of 963
891 Preacher, K.J., Hayes, A.F., 2004. SPSS and SAS procedures for estimating indirect effects in variance in the association between cardiorespiratory fitness and cognition? 964
C

892 simple mediation models. Behav. Res. Methods Instrum. Comput. 36, 717–731. Neuropsychologia 48, 1394–1406. 965
893 Ramos-Cejudo, J., Gutiérrez-Fernández, M., Otero-Ortega, L., Rodríguez-Frutos, B., Fuentes, Voss, M.W., Prakash, R.S., Erickson, K.I., Basak, C., Chaddock, L., Kim, J.S., Alves, H., Heo, S., 966
894 B., Vallejo-Cremades, M.T., Hernanz, T.N., Cerdán, S., Díez-Tejedor, E., 2015. Brain- Szabo, A.N., White, S.M., 2010b. Plasticity of brain networks in a randomized inter- 967
E

895 derived neurotrophic factor administration mediated oligodendrocyte differentiation vention trial of exercise training in older adults. Front. Aging Neurosci. 2, 1–17. 968
896 and myelin formation in subcortical ischemic stroke. Stroke 46, 221–228. Voss, M.W., Erickson, K.I., Prakash, R.S., Chaddock, L., Kim, J.S., Alves, H., Szabo, A., Phillips, 969
897 Reese, T., Heid, O., Weisskoff, R., Wedeen, V., 2003. Reduction of eddy‐current‐induced S.M., Wójcicki, T.R., Mailey, E.L., 2013a. Neurobiological markers of exercise-related 970
R

898 distortion in diffusion MRI using a twice‐refocused spin echo. Magn. Reson. Med. brain plasticity in older adults. Brain Behav. Immun. 28, 90–99. 971
899 49, 177–182. Voss, M.W., Heo, S., Prakash, R.S., Erickson, K.I., Alves, H., Chaddock, L., Szabo, A.N., Mailey, 972
900 Ruscheweyh, R., Willemer, C., Krüger, K., Duning, T., Warnecke, T., Sommer, J., Völker, K., E.L., Wojcicki, T.R., White, S.M., 2013b. The influence of aerobic fitness on cerebral 973
901 Ho, H., Mooren, F., Knecht, S., 2011. Physical activity and memory functions: an inter- white matter integrity and cognitive function in older adults: results of a one-year 974
R

902 ventional study. Neurobiol. Aging 32, 1304–1319. exercise intervention. Hum. Brain Mapp. 34, 2972–2985. 975
903 Salat, D., Tuch, D., Greve, D., Van Der Kouwe, A., Hevelone, N., Zaleta, A., Rosen, B., Fischl, Weinstein, A.M., Voss, M.W., Prakash, R.S., Chaddock, L., Szabo, A., White, S.M., Wojcicki, 976
904 B., Corkin, S., Rosas, H.D., 2005. Age-related alterations in white matter microstruc- T.R., Mailey, E., McAuley, E., Kramer, A.F., 2012. The association between aerobic 977
O

905 ture measured by diffusion tensor imaging. Neurobiol. Aging 26, 1215–1227. fitness and executive function is mediated by prefrontal cortex volume. Brain 978
906 Schmiedek, F., Li, S.-C., Lindenberger, U., 2009. Interference and facilitation in spatial Behav. Immun. 26, 811–819. 979
907 working memory: age-associated differences in lure effects in the n-back paradigm. Wersching, H., Duning, T., Lohmann, H., Mohammadi, S., Stehling, C., Fobker, M., Conty, 980
908 981
C

Psychol. Aging 24, 203–210. M., Minnerup, J., Ringelstein, E., Berger, K., 2010. Serum C-reactive protein is linked
909 Sheikh, J., Yesavage, J., Brink, T., 1986. Clinical gerontology: a guide to assessment and to cerebral microstructural integrity and cognitive function. Neurology 74, 982
910 intervention. Geriatric Depression Scale (GDS). Recent Evidence and Development 1022–1029. 983
911 of a Shorter Versionpp. 165–173. Westlye, L.T., Walhovd, K.B., Dale, A.M., Bjørnerud, A., Due-Tønnessen, P., Engvig, A., 984
N

912 Smith, S.M., Jenkinson, M., Woolrich, M.W., Beckmann, C.F., Behrens, T.E., Johansen-Berg, Grydeland, H., Tamnes, C.K., Østby, Y., Fjell, A.M., 2009. Life-span changes of the 985
913 H., Bannister, P.R., De Luca, M., Drobnjak, I., Flitney, D.E., 2004. Advances in functional human brain white matter: diffusion tensor imaging (DTI) and volumetry. Cereb. 986
914 and structural MR image analysis and implementation as FSL. NeuroImage 23, Cortex 20, 2055–2068. 987
U

915 S208–S219. Zhu, Z., Johnson, N.F., Kim, C., Gold, B.T., 2013. Reduced frontal cortex efficiency is 988
916 Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E., associated with lower white matter integrity in aging. Cereb. Cortex 25, 138–146. 989
917 Watkins, K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., 2006. Tract-based spatial
918 statistics: voxelwise analysis of multi-subject diffusion data. NeuroImage 31,
919 1487–1505.
990

Please cite this article as: Oberlin, L.E., et al., White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial
working memory in older adults, NeuroImage (2015), http://dx.doi.org/10.1016/j.neuroimage.2015.09.053