Journal of Alzheimer’s Disease xx (20xx) x–xx 1

DOI 10.3233/JAD-161240
IOS Press

1 Combined Dual-Task Gait Training

2 and Aerobic Exercise to Improve
Cognition, Mobility, and Vascular Health

f
3

roo
4 in Community-Dwelling Older Adults
5 at Risk for Future Cognitive Decline1

rP
6 Michael A. Gregorya,b , Narlon C. Boa Sorte Silvab,c , Dawn P. Gilla,b,d,e , Cheri L. McGowanb,d,f ,
7 Teresa Liu-Ambroseg,h , J. Kevin Shoemakerd , Vladimir Hachinskii , Jeff Holmesj

tho
8 and Robert J. Petrellaa,b,c,d,∗
9
a Graduate Program in Health and Rehabilitation Sciences, Faculty of Health Sciences, Western University,

10 London, ON, Canada

11
b Lawson Health Research Institute, London, ON, Canada

12
c School of Kinesiology, Faculty of Health Sciences, Western University, London, ON, Canada
Au
13
d Centre for Studies in Family Medicine, Department of Family Medicine, Schulich School of Medicine and

14 Dentistry, Western University, London, ON, Canada

15
e School of Health Studies, Faculty of Health Sciences, Western University, London, ON, Canada

16
f Department of Kinesiology, University of Windsor, Windsor, ON, Canada

17
g Department of Physical Therapy, Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada
d

18
h Djavad Mowafaghian Centre for Brain Health, Vancouver, BC, Canada

19
i Department of Clinical Neurological Sciences, London Health Sciences Centre, Western University, London,
cte

20 ON, Canada
21
j School of Occupational Therapy, Faculty of Health Sciences, Western University, London, ON, Canada
rre

Accepted 5 February 2017

22 Abstract. This 6-month experimental case series study investigated the effects of a dual-task gait training and aerobic
23 exercise intervention on cognition, mobility, and cardiovascular health in community-dwelling older adults without dementia.
co

24 Participants exercised 40 min/day, 3 days/week for 26 weeks on a Biodex GaitTrainer2 treadmill. Participants were assessed
25 at baseline (V0), interim (V1: 12-weeks), intervention endpoint (V2: 26-weeks), and study endpoint (V3: 52-weeks). The
26 study outcomes included: cognition [executive function (EF), processing speed, verbal fluency, and memory]; mobility: usual
27 & dual-task gait (speed, step length, and stride time variability); and vascular health: ambulatory blood pressure, carotid
Un

28 arterial compliance, and intima-media thickness (cIMT). Fifty-six participants [age: 70(6) years; 61% female] were included
29 in this study. Significant improvements following the exercise program (V2) were observed in cognition: EF (p = 0.002),
30

1 This work was conducted at the Parkwood Research Insti-

tute, in affiliation with the Lawson Health Research Institute and

St. Joseph’s Health Care (London, ON, Canada). Western University, 1465 Richmond St., London, ON N6G 2M1,
∗ Correspondence to: Robert J. Petrella, MD, PhD, Western Canada. Tel.: +1 519 661 2111/Ext. 22119; Fax: +1 519 685 4071;
Centre for Public Health and Family Medicine (2nd Floor), E-mail: petrella@uwo.ca.

ISSN 1387-2877/17/$35.00 © 2017 – IOS Press and the authors. All rights reserved
 2 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

31 processing speed (p < 0.001), verbal fluency [digit symbol coding (p < 0.001), phonemic verbal fluency (p < 0.001)], and
32 memory [immediate recall (p < 0.001) and delayed recall (p < 0.001)]; mobility: usual & dual-task gait speed (p = 0.002
33 and p < 0.001, respectively) and step length (p = 0.001 and p = 0.003, respectively); and vascular health: cIMT (p = 0.002).
34 No changes were seen in the remaining outcomes. In conclusion, 26 weeks of dual-task gait training and aerobic exercise
35 improved performance on a number of cognitive outcomes, while increasing usual & dual-task gait speed and step length in
36 a sample of older adults without dementia.

37 Keywords: Aging, dual-task exercise, executive function, laboratory-based, prevention

f
roo
31 INTRODUCTION In addition to AE and CT, the effect of novel 72

exercise modalities [i.e., dual-task (DT) training] 73

32 As the global population ages, the incidence of on cognition and mobility in older adults has 74

33 dementia is expected to continually rise [1]. An inte- received increasing attention. DT training is a multi- 75

34 gral component to dementia prevention efforts will be dimensional intervention that combines physical and 76

rP
35 the identification of modifiable risk factors associated cognitive tasks in order to directly train the parieto- 77

36 with cognitive worsening [1–4] and the development frontal networks of the brain [20] to divide attention 78

37 of interventions that can improve brain health and and co-ordinate actions more efficiently [21, 22]. In 79

38 functioning in older adults [5]. For instance, car- this perspective, Erickson et al. [22] observed a DT 80

diovascular disease (CVD) risk factors have been training-related ‘shift’ in the location of DT-related

tho
39 81

40 recognized as some of the most readily modifiable brain activity (i.e., reduced activation within the right 82

41 risk factors for dementia [4, 6]. Indeed, the associ- ventral inferior gyrus, right and left superior parietal 83

42 ation between heart and brain health is evident, as lobules, and right dorsal inferior gyrus accompa- 84

43 greater vascular risk factor burden has been found to nied by increased activation within the dorsolateral 85
Au
44 increase the risk of incident dementia over five years prefrontal cortex from pre- to post-training), and sug- 86

45 of follow-up among older adults with mild cogni- gested that this may represent a training-induced 87

46 tive impairment [7]. Moreover, increased CVD risk reorganization of the cortical areas involved in dual- 88

47 factors has been associated with greater task-related tasking processing. DT exercise training improves 89

48 activation and poorer task performance on execu- memory [23, 24], EF [23–26], and global cognition 90

tive function (EF) tasks in community-dwelling older [27], and can reduce the activation within regions
d

49 91

50 adults [8]. of the brain associated with short-term memory dys- 92

Habitual participation in aerobic exercise (AE) functioning [24], and increase DT gait speed [28] in
cte

51 93

52 consistently reduces CVD risk factor burden and cognitively healthy older adults. DT exercise train- 94

53 may be an important strategy to reduce the risk of ing has also been shown to benefit memory and EF, 95

54 cognitive impairment and slow the progression of as well as usual and DT gait speed among elderly 96

55 dementia [4, 9]. Previous meta-analyses suggest that fallers [29] and improve DT performance among 97
rre

56 AE can improve cognitive function within a num- older adults with dementia [30]. Collectively, these 98

57 ber of cognitive domains, including processing speed, observations suggest that DT exercise programs can 99

58 memory, and EF in healthy older adults [10–12] and benefit neural functioning, which may in turn medi- 100

59 can improve verbal fluency in those with indications ate improvements in objective cognitive functioning, 101
co

60 of underlying cognitive impairment [13]. Of particu- dynamic balance, and usual and DT gait performance 102

61 lar interest, EF appears to be exceptionally responsive among older adults. More specifically, early evidence 103

62 to AE training [10] and can also improve following indicates that DT training carried out using variable 104

63 cognitive training [14]. Furthermore, cognitive train- priority training strategy (i.e., allocating attentional 105
Un

64 ing (CT) leads to improvements in EF and memory resources intentionally to specific components of the 106

65 in healthy older adults [14, 15] and in those with cog- DT at times), may lead to greater improvements 107

66 nitive impairment [16]. Although the evidence from in cognitive performance of older adults under DT 108

67 these reviews is promising, recent meta-analyses have conditions in comparison to others strategies, such 109

68 revealed inconsistencies regarding the impact of AE as fixed-priority DT training (i.e., when attention is 110

69 interventions and improvements in aerobic fitness on equally divided between both tasks at all times) [28]. 111

70 cognitive functioning in older adults, and the specific Despite these initial observations, several limita- 112

71 exercise training modality that is best suited to benefit tions within the current literature must be addressed 113

the brain remains to be determined [17–19]. before the cognitive benefits of aerobically based 114
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 3

115 exercise training can be fully understood. Further- (MMSE) score >24] [31], and preserved instrumental 162

116 more, it is crucial to determine the efficacy of activities of daily living (Lawton-Brody Instrumen- 163

117 interventions aimed at simultaneously reducing the tal Activities of Daily Living scale) [32] were 164

118 burden of modifiable dementia risk factors (i.e., CVD invited to participate. Older adults who demon- 165

119 risk factors) and improving cognition and mobility strated significant neurological (i.e., Parkinson’s) or 166

120 in older adults at increased risk for future cogni- orthopaedic (i.e., severe osteoarthritis) conditions, 167

121 tive decline. The current body of literature suggests clinical depression (i.e., >16 on Center for Epi- 168

122 that aerobic-based exercise programs incorporating demiologic Studies-Depression Scale) [33] or at the 169

f
123 some form of cognitive training (e.g., DT) might discretion of the study physician), or BP unsafe 170

roo
124 impart a significantly larger global cognitive benefit for exercise (i.e., 180/100 mmHg or <100/60 mmHg) 171

125 then those focusing on single strategies [5]. Nonethe- [34], and those who reported a recent severe cardio- 172

126 less, the specific type of exercise intervention, along vascular complication (i.e., congestive heart failure, 173

127 with the frequency, duration, and intensity that would stroke), or could not comprehend the questionnaire 174

128 yield cognitive and mobility improvements, as well material were excluded from participation. 175

rP
129 as reduce CVD risk factors, remains equivocal [19].
130 Therefore, the primary objective of this study was Sample size 176

131 to test the hypothesis that 26 weeks of DT gait and

132 AE intervention improves performance on an EF No study to date has observed the impact of 177

tho
133 task. Secondary objectives determined whether the laboratory-based DT gait and AE on EF in older 178

134 26-week exercise intervention: i) improved perfor- adults; however, following reviews of studies using 179

135 mance on cognition tasks across multiple domains, AE [10, 35] and other cycle-based exercise [36] aim- 180

136 including: information processing, verbal fluency, ing to improve cognition [i.e., EF measured via the 181

137 and memory; ii) enhanced usual and DT gait per- Trail Making Test Part B (TMT-B)] in older adults 182
Au
138 formance; iii) reduced 24-h ambulatory systolic and allowed for the selection of an effect size of d = 0.66 183

139 diastolic blood pressure (BP), and decreased vascular for our calculations. The valid and reliable TMT-B 184

140 stiffness [i.e., carotid arterial compliance (CAC) and [37] is specific to EF processes due to its requirements 185

141 intima media thickness (cIMT)]; and iv) stimulated for switching sets and mental tracking throughout the 186

142 positive changes in cognition, mobility, and vascular task [37, 38] and was considered the primary outcome 187

outcomes that were maintained six months following measure. Assuming an alpha of 0.05, 80% power, and
d

143 188

144 the cessation of training. a dropout rate of 10%, 84 participants were required 189
cte

for this study [39]. 190

145 MATERIALS AND METHODS

Baseline variables 191

146 Study design

Participant medical history and demographics 192
rre

147 This study was a 6-month experimental case series were collected at baseline, and included: age, sex, 193

148 coupled with a 6-month no-contact follow-up. Partic- ethnicity, years of formal education, body mass 194

149 ipants were assessed at four time points throughout index, global cognitive functioning, the presence of 195

150 the intervention and follow-up period: baseline (V0), subjective cognitive complaints, and estimated car- 196
co

151 interim (V1: 12-weeks), intervention endpoint (V2: diorespiratory fitness [i.e., predicted maximal oxygen 197

152 26-weeks), and study endpoint (V3: 52-weeks). uptake (VO2 max)]. Global cognitive functioning was 198

assessed using the Montreal Cognitive Assessment 199

153 Participants (MoCA) [40]. Predicted VO2 max was estimated 200
Un

using the Step Test and Exercise Prescription (STEP) 201

154 Participants were recruited from London, ON, tool [41]. 202

155 using town hall announcements, calls to past research

156 participants, and the distribution of advertisements Cognition 203

157 to various locations throughout the community

158 (i.e., Boys & Girls Clubs, Kiwanis Clubs, media EF was assessed using TMT-A and B [37]. For 204

159 outlets). Community-dwelling older adults (60–90 the purposes of this study, TMT-B served as a surro- 205

160 years) without dementia [i.e., no previous demen- gate of EF and the primary cognitive outcome, while 206

161 tia diagnosis and a Mini-Mental State Examination TMT-A served as an index of information processing 207
 4 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

208 speed and a secondary cognitive outcome [37]. Simi- lect a longitudinal two-dimensional image of the 256

209 larly, the valid and reliable [42] Digit-Symbol Coding cephalic portion of the right common carotid artery, 257

210 (DSC) from the Wechsler Adult Intelligence Scale 1-2 cm proximal to the carotid bifurcation. Arterial 258

211 [43] was also used as an index of information pro- diameters were measured leading-edge-to-leading- 259

212 cessing speed. edge at peak systole and end diastole and averaged 260

213 Semantic (animal naming) and phonetic (Con- across three cardiac cycles. Following the acquisi- 261

214 trolled Oral Word Association Test) fluency tasks tion of the arterial diameters, carotid arterial pulse 262

215 were used to evaluate lexical verbal fluency [44]. pressure was inferred through the collection of a 263

f
216 Memory was assessed using the Auditory Verbal single measure of resting supine brachial pulse pres- 264

roo
217 Learning Test (AVLT) [45]. Responses from the sure obtained using automated oscillometry (BPTru, 265

218 immediate and delayed recall trials in the AVLT were Coquitlam, BC, Canada). Anatomical land mark- 266

219 tallied separately and served as the memory out- ing ensured accurate comparisons over time. Carotid 267

220 comes. arterial compliance (CAC) was determined using the 268

following equation: CAC = [π · (Dmax/2)2 – π · 269

rP
221 Mobility (gait) (Dmin/2)2 ]/P, where Dmax; where Dmax was the 270

systolic carotid arterial diameter, Dmin was the dias- 271

222 Spatiotemporal gait characteristics were collected tolic carotid arterial diameter, and P was resting 272

223 using a valid and reliable [46] portable electronic brachial pulse pressure [48]. cIMT was determined by 273

walkway system (GAITRite® System, Software ver-

tho
224 subtracting the carotid arterial lumen diameter from 274

225 sion 4.7.1, CIR Systems, Peekskill, NY, USA). In the outer arterial diameter at end diastole. 275

226 order to avoid capturing acceleration and decelera- In an attempt to control for external factors, 276

227 tion phases of the gait cycle, participant start and end vascular assessments were performed in a quiet, tem- 277

228 points were placed 1.5 meters before and after the perature controlled room (∼23◦ C), and participants 278
Au
229 mat. Participants were required to complete two usual were asked to refrain from the consumption of alco- 279

230 and DT walking trials at a comfortable pace. The hol or participation in moderate-vigorous intensity 280

231 serial sevens (serial 7 s) task was used as the DT con- exercise 24 h, and the consumption of caffeine 12 h, 281

232 dition in the current study (i.e., counting backwards prior to the assessments [49]. 282

233 by 7 s from 100) [29], and there was no instruction

to prioritize either gait or responses to the cognitive
d

234 Intervention: Laboratory-based DT gait and AE 283

235 tasks during the DT trials. Gait performance over training program 284
cte

236 two walks were averaged and used for analysis and
237 gait speed (m/s), step length (cm), and stride time Exercise training utilized a Biodex GaitTrainer2 285

238 variability (CoV, %) were used as outcomes. treadmill (providing visual-spatial feedback related 286

to the user’s step length on a screen fixed atop of the 287

239 Vascular health treadmill) under the supervision of research person- 288
rre

nel. During each session, participants worked through 289

240 24-h ambulatory BP and carotid ultrasonography a 5-min (min) warm-up period, one 15-min stage of 290

241 were used to evaluate vascular health. Following DT gait training, one 15-min stage of moderate inten- 291

242 the gait assessment, participants were fitted with an sity AE (i.e., 75–85% maximal heart rate determined 292
co

243 appropriately sized, valid, and reliable [47] ambu- using the STEP test protocol) [41], and a 5-min cool 293

244 latory BP cuff and monitor (SpacelabsTM 90207 down stage. During the DT gait training stage, partic- 294

245 Ambulatory Blood Pressure Monitor, SpaceLabs ipants walked at a self-selected pace while receiving 295

246 Inc.). Measurements were recorded two times an hour visuospatial step-length feedback and answering cog- 296
Un

247 during the daytime (i.e., 06:00 to 22:00), and once an nitively challenging questions (i.e., verbal fluency 297

248 hour during the nighttime (i.e., 22:00 to 06:00) over and arithmetic). The variable priority DT training 298

249 the subsequent 24-h period, and mean 24-h systolic was used during DT gait training portion of the exer- 299

250 and diastolic BP were considered as outcomes. cise sessions [26]; for the first 7-min, participants 300

251 Carotid arterial diameters were measured follow- prioritized providing correct responses to the verbal 301

252 ing previously published techniques [48]. Briefly, fluency and arithmetic tasks, and after a 1-min break 302

253 after 10 min of supine rest, a 10 MHz linear array (walk without answering questions), participants pri- 303

254 B-mode ultrasonography transducer (Vingmed, GE oritized modifying their step length to achieve or 304

255 Ultrasound A/S, Horton, Norway) was used to col- surpass an individualized step length goal (for the 305
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 5

306 remaining 7-min). Following the DT gait training or medians and interquartile ranges, where appli- 323

307 component, the visuospatial step length feedback was cable. To determine the efficacy of the 26-week 324

308 removed and participants performed 15-min of mod- intervention on EF and whether changes in TMT-B 325

309 erate intensity AE. The incline and speed of the scores were maintained after the no-contact follow- 326

310 treadmill was increased until training heart rate was up, changes in TMT-B scores (time to complete test 327

311 achieved, and the training intensity was monitored in seconds) were compared from baseline (V0) to 12- 328

312 every 5-min throughout the 15-min of AE using a weeks (V1), 26-weeks (V2) and 52-weeks (V3) using 329

313 10-point modified Borg Rating of Perceived Exer- a one-way repeated measure analysis of variance 330

f
314 tion (RPE) scale and the built-in handgrip heart rate (ANOVA) using time as a main effect and post hoc 331

roo
315 monitor on the Biodex treadmill. The intervention tests that employed Bonferroni alpha adjustments. 332

316 was conducted 40 min per day, 3 days/week for 26 Using similar approaches, we further investigated 333

317 weeks. changes in the secondary outcomes: information pro- 334

cessing (DSC and TMT-A), verbal fluency [semantic 335

318 Analysis (animal naming) & phonemic (COWA) fluency], 336

rP
memory (AVLT immediate and delayed recall), usual 337

319 All analyses were performed using IBM® SPSS® and DT gait (speed, step length and stride time vari- 338

320 Statistics for Windows, Version 20.0 (IBM Corp, ability), and 24-h systolic and diastolic BP, CAC 339

321 Armonk, NY). Demographic variables at baseline and cIMT at V2 and V3. Outliers for each outcome 340

tho
322 were summarized as means and standard deviations were identified and removed prior to analyses, and 341

d Au
cte
rre
co
Un

Fig. 1. Participant flow through the dual-task and aerobic exercise intervention and follow-up period.
 6 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

342 Greenhouse-Geiser epsilon adjusted degrees of free- ence of unidentified dementia [MMSE score, mean 368

343 dom were interpreted from the omnibus ANOVA (SD): 29 (1.3)]. 369

344 tests. Friedman tests with alpha adjusted Wilcoxon

345 sign ranked tests were used when violations of nor- Cognition outcomes 370
346 mality were encountered.
Baseline cognitive scores are summarized in 371

347 RESULTS Table 2. The effects of the 26-week DT gait and 372

AE training on the primary and secondary cogni- 373

f
348 Participants tive outcomes are reported in Table 2. The observed 374

roo
change in TMT-B performance from V0 to V1, V2, 375

349 Figure 1 describes participant flow through the and V3 is shown in Fig. 2. A significant difference 376

350 intervention. A total of 109 participants were assessed between TMT-B scores was observed (χ2 (3) = 19.49, 377

351 for eligibility, and 30 were excluded from partic- p < 0.001). Post hoc tests with Bonferroni corrections 378

352 ipation (n = 17 did not meet the inclusion criteria; (significance set at p < 0.008) revealed significant 379

rP
353 n = 13 declined to participate, primarily due to the reductions in the time to complete TMT-B from 380

354 time commitment required for the intervention). This baseline to intervention endpoint (p = 0.002), and a 381

355 left 79 participants who enrolled in the study, 56 of significant difference from baseline was maintained 382

356 whom completed the entire 52-week study. There through the no-contact follow-up period [median 383

tho
357 were no study-related adverse events experienced by (IQR): V3: 55.8 (41.6 to 74.5), p < 0.001]. There were 384

358 any of the participants throughout the intervention no significant differences in TMT-B scores at any 385

359 and follow-up period. Participant characteristics are other time points. 386

360 reported in Table 1. Participants [mean (SD) years: The observed changes in the secondary cogni- 387

361 70.4 (6.2)], were mostly women (61%), primarily tive outcomes from V0 to V2 are summarized in 388
Au
362 Caucasian (95%), highly educated [mean (SD) years: Table 2, and presented in Fig. 3. Significant reduc- 389

363 14.7 (3.2)], and just over half reported that their tions in TMT-A scores were observed following 390

364 memory has deteriorated over the past five years. Fur- the 26-week exercise intervention (p < 0.001), and 391

365 thermore, the participants had relatively preserved these changes were maintained over the 6-month 392

366 objective cognition [MoCA score, mean (SD): 25 follow-up [median (IQR): V3: 25.8 (22.3 to 29.8), 393
d

367 (3.2)] and did not display any indications of the pres- p = 0.005]. At 26-weeks, the participants showed 394

Table 1
cte

Baseline participant demographic and clinical characteristicsa

Characteristic Participants (n = 56)
Age, year 70.4 (6.2)
Female sex, 34 (61%)
Education, year 14.7 (3.2)
rre

Caucasian 53 (95%)
BMI, kg/m2,b 29.6 (4.7)
VO2 max, ml/kg/minc 28.9 (7.8)
Cognitive complaint (ref: 5 years ago)d 31 (55)
MMSE scoree 29 (1.3)
co

MoCA scoree 25 (2.5)

CES-D scoref 6.4 (5.3)
Hypertension 32 (57%)
Hypercholesterolemia 23 (41%)
Un

Type 2 diabetes 7 (12.5%)

Previous cardiovascular event 3 (5%)
Previous stroke 6 (11%)
Osteoarthritis 9 (16%)
SD, standard deviation; BMI, body mass index; MMSE, Mini-Mental Status Examination;
MoCA, Montreal Cognitive Assessment; CES-D, Centre for Epidemiological Studies Depres-
sion Scale. a Data presented as either mean (standard deviation) or no. (%). b Body Mass Index
measured in kg/m2 . c VO2 max was determined using the Step Test and Exercise Prescription
tool. d Participants rated their memory on a scale of 5 (1 = much better, 5 = much worse). e Range
from 0 to 30; lower scores indicate greater cognitive impairment. f Scores above 15 indicate
clinical depression.
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 7

Table 2
Observed changes in study outcomes from baseline (V0) to intervention endpoint (V2)a
Outcomeb V0 V2
Executive Function
TMT-B, s, (n = 51)c 65.6 (53.9 to 87.0) 57.7 (42.6 to 78.4)
Information Processing Speed
TMT-A, s, (n = 50)c 30.5 (26.7 to 36.2) 26.0 (23.0 to 32.3)
DSC, score, (n = 55)d 56.9 (13.8) 60.7 (15.0)
Verbal Fluency

f
Semantic VF, score, (n = 53)e 20.4 (5.1) 21.8 (5.1)
COWA, score, (n = 53)e

roo
13.0 (4.5) 16.5 (4.0)
Memory
AVLT immediate recall, score, (n = 51)f 7.0 (5.3 to 10.8) 11.0 (9.3 to 13.8)
AVLT delayed recall, score, (n = 56)f 8.0 (4.3 to 10.0) 12.0 (7.3 to 14.0)
Usual Gait
Speed, m/s, (n = 56) 1.11 (.19) 1.17 (.18)
Step length, cm, (n = 56) 62.2 (7.1) 64.3 (7.2)

rP
Stride time variability, CoV (%) (n = 45) 1.8 (1.5 to 2.3) 1.8 (1.5 to 2.2)
Dual-task Gait
Speed, m/s (n = 55) 0.81 (0.27) 0.91 (0.25)
Step length, cm, (n = 53) 56.1 (8.3) 59.2 (8.1)
Stride time variability, CoV (%), (n = 44) 3.5 (2.5 to 7) 3.5 (2.0 to 4.4)

tho
Vascular Health
24-h systolic BP, mmHg, (n = 45) 128 (10) 126 (10)
24-h diastolic BP, mmHg, (n = 50) 71 (6) 70 (7)
CAC, mm2 /mmHg × 10–1 , (n = 54) 0.73 (0.54 to 0.96) 0.89 (0.52 to 1.2)
cIMT, cm, (n = 54) 0.63 (0.55 to 0.74) 0.69 (0.63 to 0.80)
TMT, Trail Making Test; DSC, Digit Symbol Coding; VF, verbal fluency; COWA, Controlled Oral Word Association
Au
test; AVLT, auditory verbal learning test; BP, blood pressure; CAC, carotid arterial compliance; cIMT, carotid
intima-media thickness. a Data presented as either mean (standard deviation) or median (interquartile range). b The
removal of outliers results in differing sample sizes for the outcomes. c Lower time to completion indicates greater
performance. d Scores range from 0 to 144; higher scores indicate greater performance. e Scored as the correct number
of unique responses provided in 60 seconds. f Range from 0 to 15; higher scores indicate greater performance.
d

p <.001 performance, the observed improvements DSC 399

160
p = .002 scores, phonemic verbal fluency, and immediate and 400
cte

delayed recall following the exercise intervention 401

140
were maintained after 6-months of follow-up (all 402

120 ≤0.001). 403

Time to Completion (sec)

100
rre

Usual and dual-task gait outcomes 404

80
Changes in usual and DT gait speed, step length, 405
60
and stride time variability from V0 to V2 are summa- 406
co

rized in Table 2. Changes in usual and DT gait and 407

40
stride-time variability from V0 to V1, V2, and V3 are 408

20 presented in Fig. 4. 409

V0 V1 V2 V3 Increased usual gait speed (p = 0.002) and step 410
Un

length (p = 0.001) were observed following the exer- 411

Fig. 2. Trail Making Test (TMT) Part B performance at baseline,
cise intervention; however, after the 6-months of 412
interim (12-weeks), intervention endpoint (26-weeks), and study
endpoint (52-weeks). follow-up the improvements in usual gait speed and 413

step length no longer remained [mean difference 414

(95% CI) gait speed: 0.41 (0.90 to –0.078) m/s, 415

395 significant improvements DSC scores (p = 0.001), p = 0.15; step length: 0.96 (2.5 to –0.54), p = 0.51]. 416

396 phonemic verbal fluency (p < 0.001), and immedi- Increased DT gait speed (p < 0.001) and step length 417

397 ate (p < 0.001) and delayed recall (p < 0.001), but (p = 0.003) were observed following the 26-week 418

398 not semantic verbal fluency. Compared to baseline exercise intervention. After the 6-month follow-up, 419
 8 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

A B p = .001
60 p = .005
80 p = .001
p < .001
50
70
Time to Completion (sec)

Cor rect Responses

40
60
30
50

f
20

roo
40
10

0 30
V0 V1 V2 V3 V0 V1 V2 V3

C D

rP
p < .001 p < .001

16 p < .001 p < .001

16

14 14

12 12

tho
Cor rect Responses
Cor rect Responses

10 10

8 8

6 6
Au
4 4

2 2

0 0
V0 V1 V2 V3 V0 V1 V2 V3
d

E p < .001
30
p < .001
Semantic
cte

Phonemic
25
Cor rect Responses

20
rre

15

10
co

5
V0 V1 V2 V3

Fig. 3. Performance on secondary cognitive outcomes at baseline (V0), interim (V1: 12-weeks), intervention endpoint (V2: 26-weeks), and
Un

study endpoint (V3: 52-weeks). A) Trail Making Test Part A; B) Digit Symbol Coding; C) Auditory Verbal Learning Test immediate recall;
D) Auditory Verbal Learning Test delayed recall; E) Semantic and Phonemic verbal fluency.

420 the improvements in DT gait speed and step length Vascular health outcomes 425

421 no longer remained [mean difference (95% CI) gait

422 speed: 0.63 (0.13 to –0.08) m/s; step length: 1.3 (3.6 Differences in 24-h systolic BP, diastolic BP, CAC, 426

423 to –1.1) cm, all p > 0.05]. There were no observable and cIMT from V0 to V2 are summarized in Table 2. 427

424 reductions in usual and dual-task stride time variabil- Changes in vascular health outcomes from V0 to V1, 428

ity following 26-week exercise intervention. V2, and V3 are presented in Fig. 5. There were no 429
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 9

A 1.6
p < .002
B 80
p = .001
Usual
p < .001 75 p = .003
1.4 Dual-task
70

Step Length (cm)

1.2
Speed (m/sec)

65

1 60

f
55
0.8

roo
50
0.6
45

0.4 40
V0 V1 V2 V3 V0 V1 V2 V3

rP
C D
4
20
3.5 18

tho
3 16
14
CoV (% )

2.5
CoV (% )

12
2 10

1.5 8
Au
6
1
4
0.5 2

0 0
V0 V1 V2 V3 V0 V1 V2 V3
d

Fig. 4. Changes in usual and dual-task (serial 7 subtraction) gait speed, step length, and stride time variability from baseline (V0), interim
cte

(V1; 12-weeks), intervention endpoint (V2; 26-weeks), and study endpoint (V3; 52-weeks). A) Usual and dual-task gait speed; B) Usual
and dual-task step length; C). Usual gait stride time variability; D) Dual-task gait stride time variability. CoV, coefficient of variation.

430 significant changes in 24-h systolic BP, 24-h diastolic Following 26 weeks of treadmill-based DT gait 445

431 BP, or CAC following the 26-week exercise inter- and AE training for older adults without dementia, 446
rre

432 vention (all p > 0.05). Compared to baseline, cIMT improvements in EF were observed and main- 447

433 was higher after the 26-week exercise intervention tained over an additional 26 weeks of follow-up. 448

434 (p = 0.002), but not after the 6-month follow-up. Improvements in other cognitive processes, includ- 449

ing information processing speed, verbal fluency, 450

co

435 DISCUSSION and memory were also observed following the 26- 451

weeks exercise intervention, and these improvements 452

436 The effect of DT gait and AE training on were maintained for at least 26 weeks following the 453

437 cognition completion of the intervention. Performance on the 454

Un

semantic verbal fluency task was the only outcome 455

438 Compared to age- and education-matched nor- that remained unchanged following the interven- 456

439 mative data, the study participants demonstrated on tion, as well as the 26-week no contact follow-up 457

440 average better baseline performance on TMT-A and period. 458

441 -B [50] and semantic verbal fluency task (letters start- Evidence continues to suggest that AE training 459

442 ing with “C”) [51], comparable performance on the alone [54–56], or in combination with cognitive 460

443 DSC [45] and the AVLT [45], and poorer perfor- or DT training [27] can benefit brain health and 461

444 mance on the phonemic verbal fluency task (naming improve cognition in cognitively healthy older adults, 462

animals) [51]. and even among those with objective cognitive 463
 10 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

A 160
SBP

140 DBP

Blood Pressure (mmHg)

120

100

f
80

roo
60

40
V0 V1 V2 V3

rP
B 2.5 C 1 p = .002

0.9
2
CAC (cm 2/mmHg x 10-1)

0.8

tho
cIMT (mm)

1.5 0.7

0.6
1
Au
0.5
0.5
0.4

0 0.3
V0 V1 V2 V3 V0 V1 V2 V3
d

Fig. 5. Changes in 24-h ambulatory systolic and diastolic blood pressure (A), carotid arterial compliance (B), and carotid intima-media
thickness (C) from baseline (V0) to interim (V1; 12-weeks), intervention endpoint (V2; 26-weeks), and study endpoint (V3; 52-weeks).
cte

SBP, systolic blood pressure; DBP, diastolic blood pressure; CAC, carotid arterial compliance; cIMT, carotid intima-media thickness.

464 impairment [35, 57, 58]. Although recent meta- stimulus in combination with an AE intervention, pre- 483

465 analyses have suggested that there is limited vious studies have investigated the cognitive benefits 484
rre

466 high-quality evidence to support the use of AE train- associated with other combined cognitive and physi- 485

467 ing alone as a method to improve cognition in older cal exercise training interventions [60–64]. Although 486

468 adults with [13] or without [17] cognitive impairment, a number of exercise training modalities can benefit 487

469 recent observations suggest that combined cognitive the brain, previous observations together with those 488
co

470 and physical exercise training interventions may pro- from the current study collectively suggest that the 489

471 vide the greatest cognitive benefit [59]. cognitive response to these interventions appear to be 490

472 The results from the current study expands our unique and is likely dependent upon several key fac- 491

473 understanding of the influence of combined physical tors: i) the duration of the intervention, ii) the exercise 492
Un

474 and cognitive exercise training on cognitive function- intensity, and iii) the specific task requirements of the 493

475 ing in older adults. The 26-week exercise intervention cognitive training components of each intervention. 494

476 combined moderate intensity AE with a DT gait train- In contrast to several previous shorter duration stud- 495

477 ing component that required participants to actively ies [60–64], improvements in cognitive functioning 496

478 modify their step length using real-time biofeedback following the exercise intervention were not apparent 497

479 while simultaneously responding to a variety of ver- after 12 weeks of training, and did not emerge until 498

480 bal fluency and arithmetic tasks. Although this is the the completion of the 26-week intervention. In lieu 499

481 only study that the authors are aware of that has inves- of these observations, several methodological differ- 500

482 tigated the cognitive effects of such a unique DT ences may have contributed to the delayed cognitive 501
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 11

502 response to the DT gait and AE training, specifically: not EF. The differences in the executive cognitive 554

503 i) the cognitive and functional status of the partici- response between these studies may be in part due to 555

504 pants in the current study was relatively preserved and discrepancies in the DT requirements and differences 556

505 exercise-related improvements may have required in design of the interventions. For instance, partici- 557

506 more time to manifest; ii) the AE component was rel- pants in the previous study [27] were subject to an 558

507 atively short; iii) the use of a moderate intensity AE a group-based intermittent DT training (i.e., square 559

508 component, which was gradually progressed over the stepping exercise plus cognitive tasks) stimulus dur- 560

509 first two weeks of the intervention until the proper ing 15 min, whereas in the present study, participants 561

f
510 training intensity could be comfortably performed; exercised individually and continuously as part of a 562

roo
511 and iv) the evaluation of cognition using different 15-min DT gait training component. Although DT 563

512 neuropsychological tests where performance may be training can benefit cognition, and specifically EF 564

513 more responsive to exercise training. For instance, [27, 29], questions regarding which type of DT stim- 565

514 Dorfman et al. [29] observed significant reductions ulus and the intensity of that stimulus are best suited 566

515 in TMT-B scores following 12 weeks of treadmill- to improve cognition, still remain. The relationship 567

rP
516 based DT exercise training for older idiopathic fallers. between EF and the control of gait may have allowed 568

517 Although the participants in both studies were of sim- for the current intervention to more directly influence 569

518 ilar age, education, and cognitive status (i.e., MoCA EF than those that employ an unrelated DT condition 570

519 scores), the participants did differ on their previous during training. 571

tho
520 falls history. Cognition, especially EF, is highly asso- The longitudinal observation of the decay of 572

521 ciated with the control of gait, balance, and falls the cognitive benefits that are obtained through 573

522 prevention [65, 66]; thus, when compared to those exercise training has received little attention [5]. 574

523 without a history of falls, older adults with a his- Recently, Rahe et al. [62] observed the maintenance 575

524 tory of falls may have a greater degree of underlying of improved attention up to 1 year of follow-up, 576
Au
525 executive dysfunction, which would be more sensi- while the LIFE trial [67] did not detect any main- 577

526 tive and responsive to interventions directed toward tenance and suggest that the cognitive benefits of 578

527 mitigating falls risk. Differences in baseline TMT- exercise training dissipate after 2 years of follow- 579

528 B scores between the participants in the Dorfman up. Findings from the present study suggest that 580

529 study [29] and the present study [mean (SD): 148.8 the cognitive benefits garnered through the partici- 581

(65.3) versus 69.9 (24.7) s] suggests greater execu- pation in the exercise intervention persist for up to
d

530 582

531 tive deficit among the idiopathic fallers of the former 26-weeks following the cessation of the intervention. 583
cte

532 study, which may have allowed for a more immediate Taken together, it appears that mid-to long-duration 584

533 EF response to training. (i.e., 12- to 26-weeks) exercise training interventions 585

534 The observations presented herein are also aligned can provide cognitive benefits that persist for 26 to 586

535 with previous work that investigated the additional 52-weeks post-training; however, sustained partici- 587

536 cognitive benefit that is provided by including a DT pation in exercise training programs may be required 588
rre

537 training component to a standardized senior’s fitness to prevent the decay of any cognitive benefits that 589

538 program [27]. In example, a previous study reported are achieved. In the current study, we did not track 590

539 by our group [27] employed a 26-week random- participants’ levels of physical activity following the 591

540 ized controlled trial whereby participants performed study endpoint, which limits our understanding with 592
co

541 a standardized senior’s fitness program and mind- regards to the longevity of exercise-driven cognitive 593

542 motor exercise (i.e., square stepping exercise) in improvements. Therefore, further work is required to 594

543 isolation, or with the addition of a cognitive task (i.e., determine the trajectory of the decay in the cognitive 595

544 verbal fluency or arithmetic). Following the inter- benefits garnered through exercise training. 596
Un

545 vention, improved global cognitive functioning was

546 observed among those who performed the standard- The effect of DT gait and AE training on usual 597

547 ized fitness program followed by square-stepping and dual-task gait 598

548 exercise with simultaneous cognitive tasks, compared

549 to those who exercised without additional cogni- Compared to age-matched data, the study partici- 599

550 tive tasks. In contrast to the results of the present pants demonstrated on average comparable usual gait 600

551 study, Gill et al. [27] observed that improvements speed, step length and stride time variability [52], 601

552 in global cognition were driven by increased per- and dual task gait speed, step length, and stride time 602

553 formance on verbal fluency and memory tasks, but variability [48]. 603
 12 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

604 Improvements in usual and DT gait speed and step The effect of DT gait and AE training on 655

605 length were observed following 26 weeks of DT gait vascular health 656

606 and AE training, yet stride time variability remained

607 unchanged. Despite the beneficial effect of training, Compared to age-matched data, the study partic- 657

608 the improvements in usual and DT gait speed and ipants demonstrated on average lower cIMT [53], 658

609 step length were not maintained after 26 weeks of no and similar 24-h systolic BP, 24-h diastolic BP, and 659

610 contact follow-up. Recent meta-analyses have iden- CAC [48]. 660

611 tified increased gait speed as the primary mechanism Following 26 weeks of DT gait and AE train- 661

f
612 by which exercise benefits gait performance [68, 69]. ing, 24-h ambulatory systolic and diastolic BP, and 662

roo
613 Indeed, these suggestions are aligned with the results CAC remained unchanged, while cIMT increased. 663

614 of the current study and those from previous works, After 26 weeks of no-contact follow-up, 24-h sys- 664

615 which observed increased usual and DT gait speed tolic and diastolic BP, CAC, and cIMT were not 665

616 following 12 weeks of treadmill-based DT training significantly different from baseline. CVD risk fac- 666

617 [29] and DT gait speed following 26 weeks of stan- tors, specifically hypertension [75] and the associated 667

rP
618 dard senior’s fitness training combined with single or exacerbations in age-related arterial stiffening [76], 668

619 DT mind-motor exercise training [48]. The influence have been implicated as mechanisms that drive neu- 669

620 of exercise training on usual and DT step length is ropathological changes (i.e., reduced brain volume, 670

621 less definitive, as improvements in step length have white matter hyperintensities, and silent cerebral 671

tho
622 not been consistently found [29, 48]. In contrast to infarct) in the aging brain and the establishment of 672

623 results reported from Gregory and colleagues [48], dementia [77]. However, recent reductions in the inci- 673

624 observations from treadmill-based training interven- dence of cognitive impairment have been attributed 674

625 tions suggest that these programs can increase usual in part to increased efforts to prevent and manage 675

626 and DT step length [48]. Compared to other novel CVD risk factors [78]. Exercise training is a cor- 676
Au
627 cognitive-motor interventions, treadmill-based inter- nerstone lifestyle modification used for CVD risk 677

628 ventions involve a repetitive stepping requirement factor management, and increasing evidence sug- 678

629 that is readily comparable to the demands of usual gests that exercise can benefit cognition [5]. Although 679

630 gait, and thus provide benefits that are more read- exercise-induced adaptations to vascular structure 680

631 ily translatable to daily locomotion. Differences in and function and improved neurovascular coupling 681

the motor requirements of the DT between these have been suggested as primary mechanisms that
d

632 682

633 studies (i.e., treadmill-based versus Square Stepping drive improved cognition post-training [79], the cog- 683
cte

634 Exercise) likely contributed to the discrepancies in nitive benefits that were observed within the current 684

635 the effect of the interventions on usual gait perfor- study emerged without concurrent changes in vascu- 685

636 mance. lar health. 686

637 Stride time variability under usual and DT con- The lack of an observed change in ambulatory BP 687

638 ditions was not influenced by the DT gait and and CAC within the current study may be attributed 688
rre

639 AE intervention. Increased gait variability has been to the level of baseline fitness of the study partic- 689

640 identified as a falls risk factor [70, 71] and is a ipants and the lack of change in predicted VO2 max 690

641 common characteristic of mild cognitive impairment following the intervention. There was no requirement 691

642 [72, 73]. Participants in the present study were, for a history of recent sedentary living within the 692
co

643 on average, cognitively healthy and functionally inclusion criteria, nor was habitual exercise partici- 693

644 independent community-dwelling older adults. Fur- pation quantified upon entry to the study; the blunted 694

645 thermore, these individuals demonstrated relatively vascular response to training could have occurred 695

646 preserved stride time variability at baseline [CoV as a result of participants substituting previously 696
Un

647 (%), median (IQR): 1.8 (1.5 to 2.3) %]. Beauchet performed exercise training with the DT gait and 697

648 et al. [74] determined that only those with the greatest AE intervention. In addition, although aerobically 698

649 variability at baseline (i.e., >4.4%) experience reduc- based exercise training imparts both cardiovascular 699

650 tions in stride time variability following exercise and cognitive benefits, very little is known about 700

651 training. The relatively preserved cognitive and func- whether these benefits occur alongside one another. 701

652 tional status of the participants in the current study Other mechanisms, such as elevations in circulat- 702

653 likely contributed to the lack of observed change in ing growth factors, cortical volume, neurogenesis, 703

654 the gait variability outcomes following the exercise neural efficiency, or cerebral glucose metabolism, 704

intervention. and reductions in oxidative stress, and beta amyloid 705

 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 13

706 burden, that are able to act in a manner independent to since a similar pattern was observed for TMT-A and 756

707 changes in vascular physiology, remain under inves- DSC (i.e., measures information processing speed), 757

708 tigated and may be equally as important to consider improvement in TMT-B performance may not neces- 758

709 [80–84]. sarily reflect changes in EF, but rather be driven by 759

710 Observational studies have identified cIMT as an ameliorated performance in information processing 760

711 index of vascular stiffness, and elevations in cIMT speed. Furthermore, environmental and contextual 761

712 over time have been associated with adverse cardio- cues can also serve as a primer for cognitive perfor- 762

713 vascular events (i.e., myocardial infarction) [85], the mance. For instance, Hupbach et al. [91] found that 763

f
714 development of white matter hyperintensities [86] memories could be automatically reactivated when an 764

roo
715 and stroke [87]. Although exercise training has con- individual returns to an original learning context. The 765

716 sistently been shown to benefit traditional indices of participants in the current study performed the cog- 766

717 vascular health (i.e., BP and arterial compliance), its nitive assessments in a small clinical room that was 767

718 influence on cIMT remains equivocal. Reductions in not used for any other study-related purposes, and 768

719 cIMT have been observed, but this response has only this unique assessment environment may have served 769

rP
720 been found following high-intensity and long dura- to subconsciously prime cognitive performance. The 770

721 tion exercise training [88]. In the current study, due to possibility for contextually cued cognitive perfor- 771

722 baseline fairly high fitness levels and lack of change mance during follow-up assessments and the absence 772

723 in predicted VO2 max post-training, we did not expect of an inactive control group for appropriate compar- 773

tho
724 to see significant changes in cIMT. The observed ele- isons of cognitive performance over time must be 774

725 vations in cIMT post-training are likely the result of considered when interpreting these results. Several 775

726 normal age-related changes to vascular wall structure limitations related to the DT gait assessments must 776

727 that occur in order to maintain intra-arterial pres- also be identified, including i) the task delivery was 777

728 sure and flow homeostasis [89]. Furthermore, these not randomized (i.e., usual gait followed DT condi- 778
Au
729 observed elevations in cIMT are well within what tion), ii) the starting point for the serial subtraction 779

730 is considered the “normal” range for older adults DT was not modified between visits, and iii) per- 780

731 without established CVD [89]. Taken together, these formance on the secondary tasks within the DT gait 781

732 observations suggest that the intensity of the DT gait assessment was not methodologically controlled (i.e., 782

733 and AE intervention was insufficient to prevent the performance on serial 7 s subtraction in isolation, 783

natural progression of age-related elevations in cIMT. without the walking task). Finally, although ideal vas-
d

734 784

cular testing conditions and the associated participant 785

cte

735 Limitations responsibilities were outlined and verbally commu- 786

nicated 24 h prior to the vascular assessments [49], 787

736 The majority of the participants in the current study adherence to these requirements was not evaluated or 788

737 were Caucasian (95%) and highly educated, and enforced. 789

738 nearly two-thirds were women, all of which should be

rre

739 considered when interpreting and generalizing these Conclusions 790

740 findings. The current investigation followed a case

741 study design, and there were no controls or compar- Numerous studies continue to support the use 791

742 ison groups included. The omission of a comparison of cognitive and physical exercise training as an 792
co

743 group does not allow for the determination of whether effective non-pharmacological intervention to mit- 793

744 or not the changes in cognition that were observed igate CVD risk factor burden, improve physical 794

745 during the study occurred as a result of other extra- function, and benefit cognition [92]. During patho- 795

746 neous factors (i.e., increased socialization). There logical cognitive aging, EF and memory are often 796
Un

747 were also limitations associated with the specific out- the first cognitive domains affected [93]; therefore, 797

748 comes used in this study. Cognition was assessed identifying interventions that aim to prevent incipient 798

749 using traditional pen and paper-based neuropsycho- cognitive decline through the simultaneous targeting 799

750 logical outcomes, which may have contributed to the and training of these cognitive domains is of consid- 800

751 occurrence of practice effects. However, as previous erable importance. Treadmill-based DT gait and AE 801

752 observations suggest, the likelihood of encountering training may be an attractive choice, as the cognitive 802

753 practice effects on cognitive testing is significantly requirements of this exercise program (i.e., DT con- 803

754 diminished if assessment sessions are spaced at least trol of gait while providing responses to the verbal 804

755 12 weeks apart [90]. Also, we acknowledge that fluency task) targets and trains both EF and memory 805
 14 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

806 processes. Results from this study indicate that 26 risk factors for Alzheimer’s disease. J Neurol Neurosurg 856

807 weeks of DT gait and AE training can improve func- Psychiatry 86, 1299-1306. 857
[5] Gregory MA, Gill DP, Petrella RJ (2013) Brain health and 858
808 tioning within a number of diverse cognitive domains exercise in older adults. Curr Sports Med Rep 12, 256-271. 859
809 and benefit usual and DT gait performance, but not [6] Montine TJ, Larson EB (2009) Late-life dementias: Does 860

810 influence vascular health, in community-dwelling this unyielding global challenge require a broader view? 861

811 older adults without dementia. These observations JAMA 302, 2593-2594. 862
[7] Li J, Wang YJ, Zhang M, Xu ZQ, Gao CY, Fang CQ, Yan 863
812 support the notion that combined exercise training JC, Zhou HD (2011) Vascular risk factors promote conver- 864
813 interventions impart diverse cognitive and motor ben- sion from mild cognitive impairment to Alzheimer disease. 865

f
814 efits, and that DT gait training may be an effective Neurology 76, 1485-1491. 866

roo
815 method to directly target and train EF and memory. [8] Chuang Y-F, Eldreth D, Erickson KI, Varma V, Harris G, 867
Fried LP, Rebok GW, Tanner EK, Carlson MC (2014) Car- 868
816 Future work is required to determine whether the cog- diovascular risks and brain function: A functional magnetic 869
817 nitive benefits that are associated with DT gait and AE resonance imaging study of executive function in older 870

818 training are greater than what can be achieved follow- adults. Neurobiol Aging 35, 1396-1403. 871

819 ing other exercise training modalities, and whether [9] Barnes DE, Yaffe K, Satariano WA, Tager IB (2003) A 872

rP
longitudinal study of cardiorespiratory fitness and cogni- 873
820 these observations can be replicated in a community- tive function in healthy older adults. J Am Geriatr Soc 51, 874
821 based setting. 459-465. 875
[10] Colcombe SJ, Erickson KI, Raz N, Webb AG, Cohen NJ, 876
McAuley E, Kramer AF (2003) Aerobic fitness reduces 877
822 ACKNOWLEDGMENTS brain tissue loss in aging humans. J Gerontol A Biol Sci 878

tho
Med Sci 58, 176-180. 879

823 The authors would like to thank the study partici- [11] Hindin SB, Zelinski EM (2012) Extended practice and 880
aerobic exercise interventions benefit untrained cognitive 881
824 pants and the staff at the Parkwood Research Institute, outcomes in older adults: A meta-analysis. J Am Geriatr 882
825 in affiliation with Lawson Health Research Institute Soc 60, 136-141. 883

826 in London, Ontario. In addition, we would like to [12] Smith PJ, Blumenthal JA, Hoffman BM, Cooper H, Strau- 884
Au
827 thank the following research staff: Noah Koblinski, man TA, Welsh-Bohmer K, Browndyke JN, Sherwood A 885
(2010) Aerobic exercise and neurocognitive performance: 886
828 Ashleigh De Cruz, Lee Gonzalez, Heather Morton, A meta-analytic review of randomized controlled trials. Psy- 887
829 Stephanie Muise, and Tina Felfeli. chosom Med 72, 239-252. 888

830 This study was supported by a Team Grant from the [13] Gates N, Fiatarone Singh MA, Sachdev PS, Valenzuela 889
M, Singh MAF, Sachdev PS, Valenzuela M (2013) The 890
831 Canadian Institutes of Health research (Grant num-
effect of exercise training on cognitive function in older
d

891
832 ber: 201713), and by the Fellowships in Care of the adults with mild cognitive impairment: A meta-analysis of 892
833 Elderly Research and Rehabilitation, a training award randomized controlled trials. Am J Geriatr Psychiatry 21, 893
cte

834 offered through the Parkwood Research Institute in 1086-1097. 894

[14] Kelly ME, Loughrey D, Lawlor BA, Robertson IH, Walsh 895
835 partnership with St. Joseph’s Health Care Founda-
C, Brennan S (2014) The impact of cognitive training and 896
836 tion. The funding sources had no involvement in this mental stimulation on cognitive and everyday functioning of 897
837 study, data collection, analysis, or interpretation of healthy older adults: A systematic review and meta-analysis. 898

data. Ageing Res Rev 15, 28-43. 899

rre

838
[15] Willis SL, Tennstedt SL, Marsiske M, Ball K, Elias J, 900
839 Authors’ disclosures available online (http:// Koepke KM, Morris JN, Rebok GW, Unverzagt FW, Stod- 901
840 j-alz.com/manuscript-disclosures/16-1240r2). dard AM, Wright E, ACTIVE Study Group for the (2006) 902
Long-term effects of cognitive training on everyday func- 903
tional outcomes in older adults. JAMA 296, 2805-2814. 904
co

841 REFERENCES [16] Klusmann V, Evers A, Schwarzer R, Schlattmann P, Reis- 905

chies FM, Heuser I, Dimeo FC (2010) Complex mental and 906

842 [1] Prince M, Wimo A, Guerchet M, Ali GC, Wu YT, Prina M physical activity in older women and cognitive performance: 907

843 (2015) World Alzheimer Report 2015: The Global Impact A 6-month randomized controlled trial. J Gerontol A Biol 908

of Dementia. Alzheimer’s Disease International. Sci Med Sci 65 A, 680-688. 909

Un

844
845 [2] Daviglus ML, Plassman BL, Pirzada A, Bell CC, Bowen [17] Young J, Angevaren M, Rusted J, Tabet N (2015) Aerobic 910

846 PE, Burke JR, Connolly ES Jr, Dunbar-Jacob JM, Granieri exercise to improve cognitive function in older people with- 911

847 EC, McGarry K, Patel D, Trevisan M, Williams JW Jr (2011) out known cognitive impairment. Cochrane Database Syst 912

848 Risk factors and preventive interventions for Alzheimer dis- Rev, CD005381. 913

849 ease: State of the science. Arch Neurol 68, 1185-1190. [18] Snowden M, Steinman L, Mochan K, Grodstein F, Pro- 914
850 [3] Lehert P, Villaseca P, Hogervorst E, Maki PM, Henderson haska TR, Thurman DJ, Brown DR, Laditka JN, Soares J, 915
851 VW (2015) Individually modifiable risk factors to amelio- Zweiback DJ, Little D, Anderson LA (2011) Effect of exer- 916
852 rate cognitive aging: A systematic review and meta-analysis. cise on cognitive performance in community-dwelling older 917
853 Climacteric 18, 678-689. adults: Review of intervention trials and recommendations 918
854 [4] Xu W, Tan L, Wang HF, Jiang T, Tan MS, Tan L, Zhao QF, for public health practice and research. J Am Geriatr Soc 919
855 Li JQ, Wang J, Yu JT (2015) Meta-analysis of modifiable 59, 704-716. 920
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 15

921 [19] Kelly ME, Loughrey D, Lawlor BA, Robertson IH, Walsh Wolters Kluwer/Lippincott Williams & Wilkins Health, 986
922 C, Brennan S (2014) The impact of exercise on the cognitive Baltimore. 987
923 functioning of healthy older adults: A systematic review and [35] Baker LD, Frank LL, Foster-Schubert K, Green PS, Wilkin- 988
924 meta-analysis. Ageing Res Rev 16, 12-31. son CW, McTiernan A, Plymate SR, Fishel MA, Watson 989
925 [20] Collette F, Olivier L, Van Der Linden M, Laureys S, Delfiore GS, Cholerton BA, Duncan GE, Mehta PD, Craft S (2010) 990
926 G, Luxen A, Salmon E (2005) Involvement of both pre- Effects of aerobic exercise on mild cognitive impairment: 991
927 frontal and inferior parietal cortex in dual-task performance. A controlled trial. Arch Neurol 67, 71-79. 992
928 Cogn Brain Res 24, 237-251. [36] Anderson-Hanley C, Arciero PJ, Brickman AM, Nimon JP, 993
929 [21] Kramer AF, Larish JF, Strayer DL (1995) Training for atten- Okuma N, Westen SC, Merz ME, Pence BD, Woods JA, 994
930 tional control in dual task settings: A comparison of young Kramer AF, Zimmerman EA (2012) Exergaming and older 995

f
931 and old adults. J Exp Psychol Appl 1, 50-76. adult cognition: A cluster randomized clinical trial. Am J 996

roo
932 [22] Erickson KI, Colcombe SJ, Wadhwa R, Bherer L, Peterson Prev Med 42, 109-119. 997
933 MS, Scalf PE, Kim JS, Alvarado M, Kramer AF (2007) [37] Arbuthnott K, Frank J (2000) Trail Making Test, Part B 998
934 Training-induced functional activation changes in dual-task as a measure of executive control: Validation using a set- 999
935 processing: An fMRI study. Cereb Cortex 17, 192-204. switching paradigm. J Clin Exp Neuropsychol 22, 518-528. 1000
936 [23] Eggenberger P, Schumacher V, Angst M, Theill N, de [38] Hagen K, Ehlis A-C, Haeussinger FB, Heinzel S, Dresler 1001
937 Bruin ED (2015) Does multicomponent physical exercise T, Mueller LD, Herrmann MJ, Fallgatter AJ, Metzger 1002

rP
938 with simultaneous cognitive training boost cognitive per- FG (2014) Activation during the Trail Making Test mea- 1003
939 formance in older adults? A 6-month randomized controlled sured with functional near-infrared spectroscopy in healthy 1004
940 trial with a 1-year follow-up. Clin Interv Aging 10, 1335- elderly subjects. Neuroimage 85 Pt 1, 583-591. 1005
941 1349. [39] Fual F, Erdfelder E, Lang A, Buchner A (2007) G*Power: 1006
942 [24] Nishiguchi S, Yamada M, Tanigawa T, Sekiyama K, Kawa- A flexible statistical power analysis program for the social, 1007
943 goe T, Suzuki M, Yoshikawa S, Abe N, Otsuka Y, Nakai R, behavioral, and biomedical sciences. Behav Res Methods 1008

tho
944 Aoyama T, Tsuboyama T (2015) A 12-week physical and 39, 175-191. 1009
945 cognitive exercise program can improve cognitive function [40] Nasreddine ZS, Phillips NA, Bedirian V, Charbonneau S, 1010
946 and neural efficiency in community-dwelling older adults: Whitehead V, Collin I, Cummings JL, Chertkow H, Bédirian 1011
947 A randomized controlled trial. J Am Geriatr Soc 63, 1355- V, Charbonneau S, Whitehead V, Colllin I, Cummings JL, 1012
948 1363. Chertkow H (2005) The Montreal Cognitive Assessment, 1013
949 [25] Forte R, Boreham CAG, Leite JC, De Vito G, Brennan L, MoCA: A brief screening tool for mild cognitive impair- 1014
Au
950 Gibney ER, Pesce C (2013) Enhancing cognitive function- ment. J Am Geriatr Soc 53, 695-699. 1015
951 ing in the elderly: Multicomponent vs resistance training. [41] Stuckey MI, Knight E, Petrella RJ (2012) The Step Test 1016
952 Clin Interv Aging 8, 19-27. and Exercise Prescription Tool in primary care: A critical 1017
953 [26] Silsupadol P, Lugade V, Shumway-Cook A, van Donkelaar review. Crit Rev Phys Rehabil Med 24, 109-123. 1018
954 P, Chou LS, Mayr U, Woollacott MH (2009) Training- [42] Matarazzo JD, Herman DO (1984) Base rate data for the 1019
955 related changes in dual-task walking performance of elderly WAIS-R: Test-retest stability and VIQ-PIQ differences. 1020
956 persons with balance impairment: A double-blind, random- J Clin Neuropsychol 6, 351-366. 1021
d

957 ized controlled trial. Gait Posture 29, 634-639. [43] Wechsler D (2003) Wechsler adult intelligence scale. 1022
958 [27] Gill DP, Gregory MA, Zou G, Liu-Ambrose T, Shigematsu [44] Benton AL, Hamsher, Hamsher KD, Silvan AB (1994) 1023
cte

959 R, Hachinski V, Fitzgerald C, Petrella RJ (2016) The healthy Controlled oral word association test. Multilingual Aphasia 1024
960 mind, healthy mobility trial: A novel exercise program for Examination: Manual of Instructions. AJA Association. 1025
961 older adults. Med Sci Sports Exerc 48, 297-306. [45] van der Elst W, van Boxtel MPJ, van Breukelen GJP, Jolles 1026
962 [28] Silsupadol P, Shumway-Cook A, Lugade V, van Donke- J (2005) Rey’s verbal learning test: Normative data for 1855 1027
963 laar P, Chou LS, Mayr U, Woollacott MH (2009) Effects healthy participants aged 24-81 years and the influence of 1028
964 of single-task versus dual-task training on balance per- age, sex, education, and mode of presentation. J Int Neu- 1029
rre

965 formance in older adults: A double-blind, randomized ropsychol Soc 11, 290-302. 1030
966 controlled trial. Arch Phys Med Rehabil 90, 381-387. [46] Brach JS, Perera S, Studenski S, Newman AB (2008) The 1031
967 [29] Dorfman M, Herman T, Brozgol M, Shema S, Weiss A, reliability and validity of measures of gait variability in 1032
968 Hausdorff J, Mirelman A (2014) Dual-task training on a community-dwelling older adults. Arch Phys Med Rehabil 1033
969 treadmill to improve gait and cognitive function in elderly 89, 2293-2296. 1034
co

970 idiopathic fallers. J Neurol Phys Ther 38, 246-253. [47] Iqbal, Fotherby, Potter (1996) Validation of the SpaceLabs 1035
971 [30] Schwenk M, Zieschang T, Oster P, Hauer K (2010) 90207 automatic non-invasive blood pressure monitor in 1036
972 Dual-task performances can be improved in patients with elderly subjects. Blood Press Monit 1, 367-373. 1037
973 dementia: A randomized controlled trial. Neurology 74, [48] Gregory MA, Gill DP, Zou G, Liu-Ambrose T, Shigematsu 1038
974 1961-1968. R, Fitzgerald C, Hachinski V, Shoemaker K, Petrella RJ 1039
Un

975 [31] Folstein MF, Folstein SE, McHugh PR (1975) “Mini-mental (2016) Group-based exercise combined with dual-task train- 1040
976 state”. A practical method for grading the cognitive state of ing improves gait but not vascular health in active older 1041
977 patients for the clinician. J Psychiatr Res 12, 189-198. adults without dementia. Arch Gerontol Geriatr 63, 18-27. 1042
978 [32] Lawton MP, Brody EM (1969) Assessment of older people: [49] Pickering TG, Hall JE, Appel LJ, Falkner BE, Graves 1043
979 Self-maintaining and instrumental activities of daily living. J, Hill MN, Jones DW, Kurtz T, Sheps SG, Roccella EJ 1044
980 Gerontologist 9, 179-186. (2005) Recommendations for blood pressure measurement 1045
981 [33] Radloff LS (1977) The CES-D Scale: A self-report depres- in humans and experimental animals. Part 1: Blood pressure 1046
982 sion scale for research in the general population. Appl measurement in humans. Hypertension 45, 142-161. 1047
983 Psychol Meas 1, 385-401. [50] Tombaugh TN (2004) Trail Making Test A and B: Nor- 1048
984 [34] Pescatello LS, American College of Sports Medicine (2014) mative data stratified by age and education. Arch Clin 1049
985 ACSM’s guidelines for exercise testing and prescription, Neuropsychol 19, 203-214. 1050
 16 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise

1051 [51] Tombaugh TN, Kozak J, Rees L (1999) Normative data [65] Amboni M, Barone P, Hausdorff JM (2013) Cognitive con- 1116
1052 stratified by age and education for two measures of verbal tributions to gait and falls: Evidence and implications. Mov 1117
1053 fluency: FAS and animal naming. Arch Clin Neuropsychol Disord 28, 1520-1533. 1118
1054 14, 167-177. [66] Herman T, Mirelman A, Giladi N, Schweiger A, Hausdorff 1119
1055 [52] Verlinden VJA, van der Geest JN, Hoogendam YY, Hof- JM (2010) Executive control deficits as a prodrome to falls in 1120
1056 man A, Breteler MMB, Ikram MA (2013) Gait patterns in healthy older adults: A prospective study linking thinking, 1121
1057 a community-dwelling population aged 50 years and older. walking, and falling. J Gerontol A Biol Sci Med Sci 65, 1122
1058 Gait Posture 37, 500-505. 1086-1092. 1123
1059 [53] Lim TK, Lim E, Dwivedi G, Kooner J, Senior R (2008) [67] Sink KM, Espeland MA, Castro CM, Church T, Cohen R, 1124
1060 Normal value of carotid intima-media thickness-a surrogate Dodson JA, Guralnik J, Hendrie HC, Jennings J, Katula J, 1125

f
1061 marker of atherosclersosis: Quantitative assessment by B- Lopez OL, McDermott MM, Pahor M, Reid KF, Rushing 1126

roo
1062 mode carotid ultrasound. J Am Soc Echocardiogr 21, 112- J, Verghese J, Rapp S, Williamson JD (2015) Effect of a 1127
1063 116. 24-month physical activity intervention vs health education 1128
1064 [54] Chapman SB, Aslan S, Spence JS, DeFina LF, Keebler on cognitive outcomes in sedentary older adults. JAMA 314, 1129
1065 MW, Didehbani N, Lu H (2013) Shorter term aerobic exer- 781. 1130
1066 cise improves brain, cognition, and cardiovascular fitness in [68] Howe TE, Rochester L, Neil F, Skelton DA, Ballinger C 1131
1067 aging. Front Aging Neurosci 5, 75. (2011) Exercise for improving balance in older people. 1132

rP
1068 [55] Erickson KI, Kramer AF (2009) Aerobic exercise effects on Cochrane Database Syst Rev, CD004963. 1133
1069 cognitive and neural plasticity in older adults. Br J Sports [69] Plummer P, Zukowski LA, Giuliani C, Hall AM, 1134
1070 Med 43, 22-24. Zurakowski D (2015) Effects of physical exercise interven- 1135
1071 [56] Iuliano E, di Cagno A, Aquino G, Fiorilli G, Mignogna tions on gait-related dual-task interference in older adults: 1136
1072 P, Calcagno G, Di Costanzo A (2015) Effects of different A systematic review and meta-analysis. Gerontology 62, 1137
1073 types of physical activity on the cognitive functions and 94-117. 1138

tho
1074 attention in older people: A randomized controlled study. [70] Hausdorff JM, Rios DA, Edelberg HK (2001) Gait variabil- 1139
1075 Exp Gerontol 70, 105-110. ity and fall risk in community-living older adults: A 1-year 1140
1076 [57] Nagamatsu LS, Chan A, Davis JC, Beattie BL, Graf P, Voss prospective study. Arch Phys Med Rehabil 82, 1050-1056. 1141
1077 MW, Sharma D, Liu-Ambrose T, Nagamatsu LS, Chan A, [71] Springer S, Giladi N, Peretz C, Yogev G, Simon ES, Haus- 1142
1078 Davis JC, Beattie BL, Graf P, Voss MW, Sharma D, Liu- dorff JM (2006) Dual-tasking effects on gait variability: The 1143
1079 Ambrose T (2013) Physical activity improves verbal and role of aging, falls, and executive function. Mov Disord 21, 1144
Au
1080 spatial memory in older adults with probable mild cognitive 950-957. 1145
1081 impairment: A 6-month randomized controlled trial. J Aging [72] Verghese J, Robbins M, Holtzer R, Zimmerman M, Wang C, 1146
1082 Res 2013, 861893. Xue X, Lipton RB (2008) Gait dysfunction in mild cognitive 1147
1083 [58] Ten Brinke LF, Bolandzadeh N, Nagamatsu LS, Hsu impairment syndromes. J Am Geriatr Soc 56, 1244-1251. 1148
1084 CL, Davis JC, Miran-Khan K, Liu-Ambrose T (2014) [73] Montero-Odasso M, Casas A, Hansen KT, Bilski P, Gutma- 1149
1085 Aerobic exercise increases hippocampal volume in older nis I, Wells JL, Borrie MJ (2009) Quantitative gait analysis 1150
1086 women with probable mild cognitive impairment: A 6- under dual-task in older people with mild cognitive impair- 1151
d

1087 month randomised controlled trial. Br J Sports Med i, ment: A reliability study. J Neuroeng Rehabil 6, 35. 1152
1088 248-254. [74] Beauchet O, Launay C, Annweiler C, Fantino B, Allali G, 1153
cte

1089 [59] Law LLF, Barnett F, Yau MK, Gray MA (2014) Effects of De Decker L (2013) Physical training-related changes in 1154
1090 combined cognitive and exercise interventions on cognition gait variability while single and dual tasking in older adults: 1155
1091 in older adults with and without cognitive impairment: A Magnitude of gait variability at baseline matters. Eur J Phys 1156
1092 systematic review. Ageing Res Rev 15, 61-75. Rehabil Med 49, 857-864. 1157
1093 [60] Barnes DE, Santos-Modesitt W, Poelke G, Kramer AF, Cas- [75] Tsao CW, Seshadri S, Beiser AS, Westwood AJ, DeCarli C, 1158
1094 tro C, Middleton LE, Yaffe K (2013) The Mental Activity Au R, Himali JJ, Hamburg NM, Vita JA, Levy D, Larson 1159
rre

1095 and eXercise (MAX) trial: A randomized controlled trial MG, Benjamin EJ, Wolf PA, Vasan RS, Mitchell GF (2013) 1160
1096 to enhance cognitive function in older adults. JAMA Intern Relations of arterial stiffness and endothelial function to 1161
1097 Med 173, 797-804. brain aging in the community. Neurology 81, 984-991. 1162
1098 [61] Fabre C, Chamari K, Mucci P, Massé-Biron J, Préfaut C [76] Seals DR, Desouza CA, Donato AJ, Tanaka H (2008) 1163
1099 (2002) Improvement of cognitive function by mental and/or Habitual exercise and arterial aging. J Appl Physiol 105, 1164
co

1100 individualized aerobic training in healthy elderly subjects. 1323-1332. 1165

1101 Int J Sports Med 23, 415-421. [77] Akinyemi RO, Mukaetova-Ladinska EB, Attems J, Ihara M, 1166
1102 [62] Rahe J, Petrelli A, Kaesberg S, Fink GR, Kessler J, Kalbe E Kalaria RN (2013) Vascular risk factors and neurodegener- 1167
1103 (2015) Effects of cognitive training with additional physical ation in ageing related dementias: Alzheimer’s disease and 1168
1104 activity compared to pure cognitive training in healthy older vascular dementia. Curr Alzheimer Res 10, 642-653. 1169
Un

1105 adults. Clin Interv Aging 10, 297-310. [78] Langa KM (2015) Is the risk of Alzheimer’s disease and 1170
1106 [63] Shah T, Verdile G, Sohrabi H, Campbell A, Putland E, dementia declining? Alzheimers Res. Ther 7, 34. 1171
1107 Cheetham C, Dhaliwal S, Weinborn M, Maruff P, Darby [79] Barnes JN (2015) Exercise, cognitive function, and aging. 1172
1108 D, Martins RN (2014) A combination of physical activity Adv Physiol Educ 39, 55-62. 1173
1109 and computerized brain training improves verbal memory [80] Garciá-Mesa Y, Colie S, Corpas R, Cristòfol R, Comellas F, 1174
1110 and increases cerebral glucose metabolism in the elderly. Nebreda AR, Giménez-Llort L, Sanfeliu C (2015) Oxidative 1175
1111 Transl Psychiatry 4, e487. stress is a central target for physical exercise neuroprotection 1176
1112 [64] Theill N, Schumacher V, Adelsberger R, Martin M, Jäncke against pathological brain aging. J Gerontol A Biol Sci Med 1177
1113 L (2013) Effects of simultaneously performed cogni- Sci 71, 40-49. 1178
1114 tive and physical training in older adults. BMC Neurosci [81] Griffin ÉW, Mullally S, Foley C, Warmington SA, 1179
1115 14, 103. O’Mara SM, Kelly ÁM (2011) Aerobic exercise improves 1180
 M.A. Gregory et al. / Dual-Task Gait Training and Aerobic Exercise 17

1181 hippocampal function and increases BDNF in the serum of of stroke and myocardial infarction: The Rotterdam Study. 1203
1182 young adult males. Physiol Behav 104, 934-941. Circulation 96, 1432-1437. 1204
1183 [82] Lange-Asschenfeldt C, Kojda G (2008) Alzheimer’s dis- [88] Thijssen DHJ, Cable NT, Green DJ (2012) Impact of exer- 1205
1184 ease, cerebrovascular dysfunction and the benefits of exer- cise training on arterial wall thickness in humans. Clin Sci 1206
1185 cise: From vessels to neurons. Exp Gerontol 43, 499-504. (Lond) 122, 311-322. 1207
1186 [83] Lista I, Sorrentino G (2010) Biological mechanisms of [89] Engelen L, Ferreira I, Stehouwer CD, Boutouyrie P, Laurent 1208
1187 physical activity in preventing cognitive decline. Cell Mol S (2012) Reference intervals for common carotid intima- 1209
1188 Neurobiol 30, 493-503. media thickness measured with echotracking: Relation with 1210
1189 [84] Tsai C-L, Wang C-H, Pan C-Y, Chen F-C (2015) The effects risk factors. Eur Heart J 34, 2368-2380. 1211
1190 of long-term resistance exercise on the relationship between [90] Bartels C, Wegrzyn M, Wiedl A, Ackermann V, Ehrenreich 1212

f
1191 neurocognitive performance and GH, IGF-1, and homocys- H (2010) Practice effects in healthy adults: A longitudinal 1213

roo
1192 teine levels in the elderly. Front Behav Neurosci 9, 23. study on frequent repetitive cognitive testing. BMC Neurosci 1214
1193 [85] O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke 11, 118. 1215
1194 GL, Wolfson SK (1999) Carotid-artery intima and media [91] Hupbach A, Hardt O, Gomez R, Nadel L (2008) The dynam- 1216
1195 thickness as a risk factor for myocardial infarction and stroke ics of memory: Context-dependent updating. Learn Mem 1217
1196 in older adults. Cardiovascular Health Study Collaborative 15, 574-579. 1218
1197 Research Group. N Engl J Med 340, 14-22. [92] Bherer L (2015) Cognitive plasticity in older adults: Effects 1219

rP
1198 [86] Bots ML, van Swieten JC, Breteler MM, de Jong PT, van of cognitive training and physical exercise. Ann N Y Acad 1220
1199 Gijn J, Hofman a, Grobbee DE (1993) Cerebral white matter Sci 1337, 1-6. 1221
1200 lesions and atherosclerosis in the Rotterdam Study. Lancet [93] Carlson MC, Xue QL, Zhou J, Fried LP (2009) Executive 1222
1201 341, 1232-1237. decline and dysfunction precedes declines in memory: The 1223
1202 [87] Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE Women’s Health and Aging Study II. J Gerontol A Biol Sci 1224
(1997) Common carotid intima-media thickness and risk Med Sci 64, 110-117. 1225

tho
d Au
cte
rre
co
Un