Skip to main content

Table 3 Studies identified with brain structure and muscle function

From: A systematic review of the evidence that brain structure is related to muscle structure and their relationship to brain and muscle function in humans over the lifecourse

Author

Year

Country and dataset

n

Study design

Mean age (sd)

Male (%)

Brain structure#

Muscle function

Associations*

Brain structure and grip strength

1. Sachdev et al. [32]

2009

Australia, PATH through life project

432

Observational cohort study

M 62.61 (1.42) F 62.62 (1.44)

52.8

Volumes of GM, WM and CSF, ICV and TBV (GM plus WM). Brain atrophy and subcortical atrophy, WMH

Grip strength in writing hand

Study: Total brain WMH volume predicted grip strength in men (beta −0.140, delta R2 0.019, p < 0.05) but not in women (beta −0.140, delta R2 0.018, p > 0.05).

2. Anstey et al. [33]

2007

Australia, PATH through life project

432

Observational cohort study

62.63 (1.43)

51.6

Total, anterior, midbody and posterior corpus callosum (CC) area

Grip strength in writing hand

Study: Grip strength adjusted for sex and ICV was found to correlate with CC midbody area (r = 0.103, p < 0.05), however CC total area and anterior and posterior CC areas did not significantly correlate with grip strength (p > 0.05).

3. Sachdev et al. [34]

2006

Australia, PATH through life project

469

Observational cohort study

M 62.56 (1.44) F 62.53 (1.47)

51.8

Volumes of GM, WM and CSF, ICV and TBV (GM plus WM). Brain atrophy and subcortical atrophy, WMH

Grip strength in writing hand

Study: None, see other articles from the PATH through life project for analysis using this dataset.

4. Sachdev et al. [35]

2005

Australia, PATH through life project

478

Observational cohort study

M 62.56 (1.44) F 62.54 (1.47)

52.3

WMH, ICV

Grip strength in writing hand

Study: Total brain WMH significantly predicted grip strength (beta −0.09, p = 0.002) adjusted for age, sex and depression. Correcting for comorbidity, cognition and brain atrophy did not attenuate the results (beta −0.13, p =0.001).

5. Doi et al. [36]

2012

Japan

110

Cross-sectional study

75.4 (7.1)

50

GM, WM, CSF, brain atrophy (measured using healthy volunteers)

Grip strength

Study: A MLR model found that grip strength is not related to brain atrophy (beta −0.082 (SE 0.005) p = 0.54). Adjusting for age, gender, BMI, education, MMSE, Tokyo Metropolitan Institute of Gerontology Index of Competence, geriatric depression scale and change in walking whilst dual tasking. No other associations given.

6. Hardan et al. [37]

2003

USA, Philadelphia

41 controls

Case–control study

18.6 (8.6)

Not given

Caudate, putamen and total brain volume

Grip strength

Study: Non-significant trends showed a negative correlation between right grip strength and total caudate volume (r = −0.303, p = 0.05) and left grip strength (r = −0.28, p = 0.07) in the control group. Not corrected for age or sex. No relationships given for other measures.

Brain structure and gait speed

7. Piguet et al. [38]

2006

Australia, Sydney Older Person's Study

111

Longitudinal observational cohort study

M 85.29 (2.89) F 85.72 (3.41)

54.5

Cerebellar vermis area, (V1, V2 and V3 and total), Cerebellar volume, cerebral volume and ICV

Timed walk over 5 m, adjusted for lower limb arthritis

Study: None of the brain size measures (cerebellar vermis area, cerebellar volume or cerebral volume) significantly predicted timed walk (p > 0.05) after adjustment for age (but not sex, as was not deemed to be a significant contributor after univariate analyses).

8. Callisaya et al. [39]

2013

Australia, Tasmanian Study of Cognition and Gait (TASCOG)

225

Longitudinal cohort study

71.4 (6.8)

56.4

ICV, GM, WM-lesion free, hippocampal volume, WML

4.6 metre GaitRite computerized walkway (preferred speed)

Study: MLR were performed to investigate the relationship of longitudinal change in brain volumes and gait speed. They found that white matter atrophy (beta 0.25 (CI 0.09-0.40) p = 0.001), greater WML progression (beta −0.89 (CI −1.75- -0.02) p = 0.045), grey matter atrophy (beta 0.25 (CI 0.00-0.19) p = 0.06) and hippocampal atrophy (beta 0.01 (CI 0.00-0.02) p = 0.006) were all associated with a greater decline in gait speed.

9. Srikanth et al. [40]

2010

Australia, TASCOG

385

Longitudinal cohort study

72.2 (7.1)

56

WMLV, TBV

Gait speed using 4.2 m GAITRite system

Study: none, see Callisaya et al. (2013) for analysis using the TASCOG dataset.

10. Srikanth et al. [41]

2009

Australia, TASCOG

294

Longitudinal cohort study

72.3(7.0)

55.4

WMLV, TBV

Gait speed using 4.2 m GAITRite system

Study: none, see Callisaya et al. (2013) for analysis using the TASCOG dataset.

11. Elbaz et al. [42]

2013

France, Three-city study

4010

Cohort study

73.4 (4.6)

38.4

WML volumes

6 metre walk speed (usual and maximum)

Study: Logistic regression stratified by education found that high WML volumes were not associated with slow walking speed among highly educated participants (OR = 0.72), but were associated with a 2-fold-increased risk of slow walking speed among those with low education (OR = 3.19/1.61 = 1.99) (p interaction = 0.026), adjusted for sex, age and total WM volume. Results remained unchanged after adjustment for height, BMI, and MMSE score.

Given: WM volume did not predict walking speed at baseline, adjusted for age, gender and ICV in a MLR (p > 0.05, n = 1510), or decline in walking speed over 7 years, adjusted for age, gender, ICV and baseline walking speed, (p > 0.05, n = 928). A logistic regression found that WM volume was not significantly associated with an increased risk of being in the quartile with the highest walking speed decline (p > 0.05).

12. Dumurgier et al. [43]

2012

France, Three-city study

1623

Cohort study

73.3 (4.1)

39.5

Regional grey matter volumes (sensorimotor cortex; frontal, parietal, temporal, occipital, and limbic lobes; insula; cerebellum; thalamus; basal ganglia nuclei, including the caudate nucleus, putamen and pallidum) and WMLs

Maximum walking speed over 6 metres

Study: A linear regression found that only basal ganglia volume (beta 0.075 (SE 0.025) p = 0.003) was significantly associated with walking speed; driven by caudate nucleus volume (beta 0.114 (SE 0.024) p < 0.001). All other regional GM volumes were not significantly associated with walking speed. A semi-bayes model found again only the basal ganglia volume (beta 0.061 (SE 0.028) p = 0.03) was significantly associated with walking speed; driven by caudate nucleus volume (beta 0.050 (se 0.019) p = 0.007). There was found to be a linear relationship between quartiles of caudate nucleus volume and faster walking speed (p for linear trend (0.001). These relationships were attenuated slightly for total basal ganglia volume by adjusting for MMSE and comorbidity plus smoking but not for caudate nucleus volume. All models adjusted for; age, sex, BMI, education level, ICV, volume of WMLs and silent infarcts. Given: See Elbaz et al. (2013) for Three-City Study data analysis.

13. Dumurgier et al. [44]

2010

France, Three-city study

Baseline 3604, f/u at 4y 1774

Cohort study

Baseline 73.4 (4.6) f/u 71.5 (3.6)

Baseline 38.1%, f/u 38.4%

WMH volume

Maximum walking speed over 6 metres, 1st and 4th follow up, mean 7 years

Study: none

Given: See Elbaz et al. (2013) for Three-City Study data analysis.

14. Soumare et al. [45]

2009

France, Three-city study

1702

Cohort study

72.4 (4.1)

39.4

PVH, deep WMH and total WMH and total WM and ICV

Maximum walking speed over 6 metres, 1st and 4th follow up, mean 7 years

Study: A significantly lower mean walking speed was found in those with a total WMH volume above the 75th percentile compared to those below the 25th (Beta −0.026, p = 0.0003). A similar relationship was found for both deep WMH and PVH. A WMH volume greater than the 90th percentile more than doubled the risk of decline in walking speed compared with subjects with lower volumes of WMH (OR 2.6 (1.5-4.5), p = 0.001). This finding was replicated when looking at PVH but not for deep WMH volume. Given: See Elbaz et al. (2013) for Three-City Study data analysis.

15. Starr et al. [46]

2003

UK, ABC1921 cohort study

97

Longitudinal cohort study

78-79years

59.8

WMH in deep/subcortical, PVH and brain stem, Fazekas score

Self-paced time to walk 6metres

Study: A slower 6metre walk test was associated with increased brain stem lesions (F 7.11, p = 0.009, partial eta2 0.070), but not with WMH (deep) (F 3.33, p = 0.071) or PVH (F 2.47, p = 0.12). Doesn’t state if age and sex are adjusted for in these models. If HADS score and Raven’s score are adjusted for, brainstem lesions are no longer significantly associated with walking time.

16. Manor et al. [47]

2012

USA, Boston,

89 in control group

Case–control study

65.3 (8.2)

48.3

GM, WM, CSF, regional GM volumes; precentral and postcentral gyri, basal ganglia, cerebellum, and dorsolateral prefrontal cortex

75 metre walk test at preferred pace

Study: Within linear regression models, global GM volume and all of the regional GM volumes were not associated with walking speed in the control group (p > 0.005, Bonferroni adjusted). Adjusted for age, sex and body mass.

17. Hajjar et al. [48]

2010

USA, Boston, BP in stroke study (?overlap with Novak et al.)

Non-stroke group 43

Case–control observational study

68 (1)

44

WM, GM (global and regional), CSF normalized for ICV

Gait speed over 12mins at usual pace

Study: Gait speed was not significantly associated with GM volume (p = 0.85), but was significantly associated with WM volume (B = 1.30, p = 0.03) adjusting for age, gender, BMI and antihypertensive use.

18. Novak et al. [49]

2009

USA, Boston (?overlap with Hajjar et al.)

76

Observational study

64.7 (7.2)

47.4

GM, WM, CSF, WMH all as % brain tissue volume. WMH using Wahlund scale

Gait speed over 12mins at normal walking pace

Study: Gait speed was significantly associated with frontal WM normalized for brain tissue volume (R = 0.4, p = .003). Gait speed was significantly associated with frontal GM normalized for brain tissue volume (R = 0.3, p = .01). Adjusted for age and BMI (but not gender). Doesn’t say about other regional brain volumes, ie temporal etc. WMH volumes and PVH and punctuate scores were not associated with gait speed (p > 0.05).

19. Moscufo et al. [50]

2012

USA, Boston, Moscufo study – 2 year f/u

77

Longitudinal cohort study

84 (3.9)

40

WMH volume as % of ICV and regional WMH burden expressed as % of ROI volume. At baseline and 2y f/u.

Gait speed over 2.5 metres, maximum velocity and usual walking speed At baseline and 2y f/u.

Study: Total WMH burden was significantly associated with usual walking speed at baseline but not at follow-up, and maximum walking speed was not associated with total WMH at baseline or follow up. At baseline, regional WMH burden in the splenium of corpus callosum and anterior and superior corona radiata, was significantly associated with both walking measures (p < 0.05) and in addition the body of the corpus callosum was also associated with usual walking speed (p < 0.05). At follow-up, WMH burden in the splenium was significantly associated with both walking measures (p < 0.05) and in the body with maximum walking speed. Change in WMH burden, either total or in any of the 7 regional areas, over 2 years was not associated with a decline in usual walking speed (p > 0.1).

Given: WMH burden is significantly associated with lower gait speed after adjustment for age, sex and BMI (rho = −0.327, p = 0.0008). WM/ICV is not significantly associated with gait speed with or without adjustment (p > 0.05). GM/ICV is significantly associated with gait speed with adjustment for age, gender and BMI (rho = 0.232, p < 0.05). CSF/ICV is significantly associated with gait speed with adjustment for age, sex, BMI (rho = −0.285, p = 0.004).

20. Moscufo et al. [51]

2011

USA, Boston, Moscufo study - baseline

99

Cross-sectional observational study

83(4)

42.4

WM, GM, WMH and CSF volumes all corrected for ICC. Brain atrophy. Regional WMH burden expressed as % of ROI volume.

Gait speed over 2.5metres (done as part of SPPB)

Study: Total WMH burden (i.e. % of ICV) correlates with gait speed (rho = −0.288, p = 0.004). Also all 9x regional burden measurements correlate with gait speed score too except sup. longitudinal fasciculus. No adjustment.

Given: See Moscufo et al. (2012) for analysis using this dataset.

21. Wolfson et al. [52]

2005

USA, Boston, WML and mobility

28 at baseline, 14 at follow up

Prospective longitudinal observational study

SPPB 11or12 mean 81(1.7), SPPB = <8 mean 84(3.4)

64.3

GM, WM, WMSA, CSF, ICCV volumes

Gait velocity over 8metres

Study: Slower baseline gait velocity predicted more WMSA at visit 1 (p < 0.05), but not change in WMSA volume between visit 1 and 2 (p < 0.07). Significant negative relationship of between-visit change in gait velocity to CSF volume (r = 0.733, p < 0.005) and a positive relationship of between-visit change in gait velocity to WM volume (r = 0.558, p < 0.05). Betas not given. Brain volumes normalized for ICCV according to image processing section.

22. Guttmann et al. [53]

2000

USA, Boston, WML and mobility

28 (12 with SPPB score >10 and 16 < 9)

Observational cross-sectional study

SPPB > 10 79(5) SPPB < 9 83(6)

42.9

WM, WMSA, GM, CSF (normalized for ICCV)

Gait velocity over 8metres

Study: Gait velocity was not significantly predicted by age nor WMSA volume (no figures given or p value) adjusted with and without MMSE score.

23. Rosano et al. [54]

2012

USA, Cardiovascular health study

214

Longitudinal observational study

72.3 (3.8)

35.5

Brain volumes (GM, WMH, Prefrontal area, WM, CSF)

Timed 15 ft walk at usual pace

Study: Prefrontal area volume significantly predicted time to walk in a stepwise forward model (beta −0.15, p = 0.02).

24. Barnes et al. [55]

2009

USA, Cardiovascular Health Cognition Study, nested within the CVS Health Study

3375

Prospective, population-based, longitudinal study

75 (no sd)

41

White matter disease and ventricular enlargement

Gait speed over 15 ft

Study: none, see Rosano (2012), Rosano (2006), Rosano (2005) and Longstreth (1996) for analysis using the Cardiovascular Health study dataset

25. Rosano et al. [56]

2006

USA, Cardiovascular health study

321

Longitudinal observational study mean f/u 4 years

78.3 (no sd)

39.3

WMAs, ventricular enlargement

Gait speed at usual pace over 4 metres using GaitMat II

Study: Gait speed was significantly correlated to total WMAs (r = −.18, p < 0.0001) and white matter lesions in the brainstem (r = −.18, p = 0.01). After adjusting for age, slower gait speed was still significantly associated with white matter grade (p = 0.02). Logistic regression found that those in the lowest two quartiles of gait speed (ie < 1.02 m/s) had double the likelihood of having WMH graded 3 or above (p = 0.03), after adjustment for age, race, gender, and prevalent clinical CVD. VE graded >4 was not found to be significantly predicted by gait speed, however VE graded > 5 was, independent of age, gender, race and presence of CVD (OR = 2.91 for 1st vs. 4th quartile, OR 3.82 for 2nd vs 4th quartile)

26. Rosano et al. [57]

2005

USA, Cardiovascular health study

2450

Longitudinal observational study mean f/u 4 years

74.4 (4.7)

43

WMH and ventricular enlargement (graded as minimal, moderate and severe)

Gait speed over 15 ft at usual pace, starting from standing still

Study: Grade of ventricular enlargement was associated with baseline gait speed and mean change in gait speed/year. Gait speed decline was 2.5x that for those with severe VE than minimal VE. (p < 0.001). Grade of WMH was associated with baseline gait speed and mean change in gait speed/year (p = 0.003). In both analyses adjustment had been made for age, sex, race and education and CV risk factors (BMI, systolic BP, antihypertensive meds, internal carotid wall thickness, and ETOH intake) and prevalent CV disease.

27. Silbert et al. [58]

2008

USA, Oregon Brain Aging Study

104

Longitudinal cross-sectional study

85.1 (5.6)

38.5

PV WMH and s/c WMH, total WMH, brain volume, CSF volume, hippocampal volume, ICV

Gait speed over 9 m. Self-selected pace.

Study: Adjusted for age and ICV, higher baseline total WMH vol. was associated with increased rate of change in timed walking in seconds (r2 = 0.08, p = 0.0052). This relationship became non-significant after adjustment for multiple comparisons to threshold p value. PVH volume is associated with increased rate of change in timed walk in seconds (r2 = 0.12, p = .0039). However, baseline subcortical WMH vol. was not related to change in gait performance over time. Higher rate of PVH accumulation is associated with increased rate of change of time to walk 9 m (r2 = 0.15, p = .0453). Adjusted for age, ICV and baseline WMH volume:

Calculated: In an unadjusted GLM, gait speed was predicted by total brain, WMH and hippocampal volume (p < 0.001). The relationship remained significant after adjusting for sex, age, ICV and height, for total brain volume (t = 3.61, p = .004, partial eta squared 4.3%) and WMH (t = −2.80, p = 0.006, partial eta squared 4.4%) but not for hippocampal volume.

28. Marquis et al. [59]

2002

USA, Oregon Brain Aging Study

108

Longitudinal cross-sectional study

83.2 (7.9)

37

Total brain volume, hippocampal volume, ICV

Gait speed over 9 m. Self-selected pace.

Study: Negative correlation between hippocampal volume and time to walk 30 ft (r = −.12). No p value given.

Calculated: See Silbert et al. (2008) for Oregon Brain Aging Study data analysis.

Brain structure and gait speed plus grip strength or isometric knee extension strength (IKES)

29. Rosano et al. [60]

2010

Iceland, AGES-Reykjavik study

795

Longitudinal cohort study

M 75.6 (5.4) F 75.6 (5.7)

41.1

MTR, ICV, brain parenchyma volume, semiquantitative subcortical WMH and PVH and total WMH volume, brain atrophy index

Gait speed over 6 m usual speed and maximal isometric knee extension strength

Study: In men: Time to walk 6metres predicted by WMH volume (beta 0.13, p = 0.02) but not brain atrophy or peak height MTR (adjusted for age and brain size as includes measure of brain atrophy). In women: Usual walking speed predicted by lower MTR height (i.e. indicating abnormal brain tissue) (beta −0.14 (p = 0.01), increased WMH (beta 0.12, p = 0.003) and greater brain atrophy (beta 0.15, p = 0.01) (adjusted for age and brain size). Lower muscle strength associated with peak height MTR (p < 0.005, beta not given).

30. Aribisala et al. [61]

2013

UK, LBC 1936 study

694

Longitudinal cohort study

69.5 (0.7) wave 1 and 72.5 (0.7) wave 2

52.9

TBV, ventricular volume, GM, NAWM and WML at wave 2

6 metre walk (normal walking pace) and grip strength at wave 1 and 2

Study: Grip strength at wave 1 significantly predicts ventricular volume at wave 2 (standardized beta −0.10), however there was no significant association with other brain volumes. 6metre walk at wave 1 predicted TBV (−0.07), ventricular volume (0.09), NAWM (−0.07) and WML (0.11) all p < 0.05. Grip strength at wave 2 was associated with ventricular volume (−0.11) and NAWM (0.08). 6 MW at wave 2 was associated with TBV (−0.07), NAWM (−0.09) and WML (0.11) all p < 0.05. Change in physical function between wave 1 and 2 (i.e. decrease in grip strength or increase in 6 MW) was not significantly associated with any brain volume measure. GM volume did not significantly associate with any of the physical function variables at wave 1 or 2. All analyses were adjusted for age, ICV, age 11 IQ, years of education, social class, comorbidity and smoking status. Corrected for false discovery rate.

31. Rosano et al. [62]

2011

USA, Cardiovascular health study

643

Longitudinal observational study

72.1-72.6 broken down by BP diagnosis

31-42.7 broken down by BP diagnosis

WMH scale 0-9

Gait speed over 15 ft, starting from standstill. Grip strength of dominant hand.

Study: none, see Rosano (2012), Rosano (2006), Rosano (2005) and Longstreth (1996) for analysis using the Cardiovascular Health study dataset.

32. Rosano et al. [63]

2008

USA, Cardiovascular health study

3156

Longitudinal observational study mean f/u 4 years

74 (4.6)

43.2

White matter disease score, brain atrophy score (ventricular enlargement)

Gait speed over 15 ft and grip strength in dominant hand

Study: none, see Rosano (2012), Rosano (2006), Rosano (2005) and Longstreth (1996) for analysis using the Cardiovascular Health study dataset.

33. Longstreth et al. [64]

1996

USA, Cardiovascular health study

3658

Longitudinal observational study

70.7 (no sd)

41.7

MR WMSA graded 0-9

Time to walk 15feet, grip strength in dom and non-dom hand

Study: Time to walk 15 ft correlated with white matter grade (0–9) (r = 0.153, p < 0.001), with adjustment for age, sex and presence of clinically silent stroke on MRI. Same model showed no significant associated between grip strength in dom hand or non-dom hand and white matter grade (p > 0.05).

  1. #All brain structure variables performed using MRI.
  2. *Associations column key: Study = results published within the study; Given = associations calculated by study authors and supplied to us for this review; Calculated = study authors supplied raw data to us and we performed the analysis.