Functional Neural Correlates of Slower Gait Among Older Adults With Mild Cognitive Impairment

Functional Neural Correlates of Slower Gait Among Older Adults With Mild Cognitive Impairment Abstract Background Subtle, but observable, changes in mobility often exist among older adults with mild cognitive impairment (MCI). Notably, these changes are not inconsequential. Therefore, there is a strong interest to better understand the underlying neural correlates of gait slowing among older adults with MCI. In this study, we aimed to characterize patterns of functional connectivity associated with slower gait speed in older adults with MCI. Methods Forty-nine participants aged 60 years and older with MCI were included in the cross-sectional study. All participants underwent assessments of gait speed and resting state functional magnetic resonance imaging. Results In this sample of older adults with MCI, slower usual gait was characterized by altered connectivity between the sensorimotor network (SMN) and the frontoparietal network (FPN) (p < .05)—specifically, slower usual gait was associated with greater connectivity between the supplementary motor area (SMA) and the bilateral ventral visual cortices (p = .01); lower connectivity between the SMA and the bilateral superior lateral occipital cortex (p < .01); and lower connectivity between the SMA and the bilateral frontal eye field (p < .01). Conclusion Altered inter-network functional connectivity between the SMN and FPN may be a neural mechanism for slowing of gait in older adults with MCI. Mild cognitive impairment, Slower gait, Functional connectivity, Older adults Mild cognitive impairment is a clinical entity characterized by cognitive decline greater than expected for an individual’s age and education level that does not interfere with everyday function (1). While individuals with MCI are at increased risk for dementia (2), there is growing recognition that subtle, but observable, changes in mobility (ie, slowing of gait) often exist among this population (3). Notably, these changes are not inconsequential. For example, Doi and colleagues (4) showed that the combination of MCI and slow gait confers a higher risk of disability than each condition independently. Also, high dual-task gait cost (ie, [single-task gait velocity—dual-task gait velocity]/single-task gait velocity)was associated with incident dementia in MCI (5). There is a strong interest to better understand the underlying neural correlates of gait slowing in the prodromal stages of the dementia process, such as in MCI (6,7). Onen and colleagues (6) showed that periventricular leuokaraiosis was associated with slow gait among older adults with MCI. Moreover, Beauchet and colleagues (7) demonstrated that slower gait speed was associated with larger brain ventricle volume among older adults with MCI. However, whether slowing of gait speed among older adults with MCI is also associated with aberrant functional connectivity of the relevant neural networks remains largely unexplored. A better understanding of the functional neural mechanisms underlying slowing of gait speed among older adults with MCI would complement previous structural findings and provide new insights to future strategies to maintain mobility and functional independence among those with MCI. Our current exploratory study is, in part, motivated by the motoric cognitive risk syndrome (MCR) proposed by Verghese and colleagues (8). MCR is operationally defined by the presence of concomitant cognitive complaints and slow gait, without severe functional impairments or dementia (8). Studies showed that MCR is highly prevalent among older adults (8) and incidence rate of dementia was more than doubled among older adults with subjective cognitive complaint, mild cognitive impairment (MCI), and slow gait compared with those without MCR (9). Therefore, we aimed to characterize patterns of functional connectivity associated with slower gait speed in older adults with MCI. We specifically focused on two functional networks, the sensorimotor network (SMN) and the frontoparietal network (FPN). The SMN is actively involved in major aspects of movement, including motor-planning, initiation, execution, and coordination (10). The FPN is involved in top-down attentional control (11) and allocation of available neural resources to important cognitive processes (12), as well as motor planning and motor execution (13). Of relevance, our previous 12-month prospective study showed that lower inter-network connectivity between the SMN and the FPN was associated with subsequent decline in both executive functions and lower extremity physical performance in community-dwelling older adults (14). Our objective for the current study is to determine whether slower gait is characterized by lower connectivity between the SMN and FPN in older adults with MCI. Methods and Measures Study Design and Participants Forty-nine community-dwelling older adults with MCI were included in this cross-sectional study. Mild cognitive impairment was defined using the criteria provided by the National Institute of Aging and Alzheimer’s Association (NIA-AA) (15) as the following: (i) objective cognitive impairment, operationalized in our study as a Montreal Cognitive Assessment (MoCA) score < 26/30 (15); (ii) have subjective memory complaints (SMC); (iii) no significant functional impairment; and (iv) not formally diagnosed with dementia. Participants were recruited from metropolitan Vancouver via advertisement and interested individuals were screened by telephone to confirm general eligibility according to the inclusion and exclusion criteria, followed by an in-person screening session. We included those who: (i) were aged ≥ 60 years; (ii) scored < 26/30 on the MoCA (16). The MoCA is a 30-point test that covers multiple cognitive domains (16). The MoCA has been found to have good internal consistency and test-retest reliability and was able to correctly identify 90% of a large sample of MCI individuals from two different clinics (16); (iii) had SMC, defined as the self-reported feeling of memory worsening with an onset within the last 5 years, as determined by interview (17); (iv) score ≥ 6/8 on the Lawton and Brody (18) Instrumental Activities of Daily Living Scale; (v) preserved general cognition as indicated by a Mini-Mental State Examination (MMSE) (19) score ≥ 24/30; (vi) were right hand dominant as measured by the Edinburgh Handedness Inventory (20); (vii) were living independently in their own homes; (viii) had visual acuity of at least 20/40, with or without corrective lenses; and (ix) provided informed consent. We excluded those who: (i) had a formal diagnosis of neurodegenerative disease, stroke, dementia (of any type), or psychiatric condition; (ii) had clinically significant peripheral neuropathy or severe musculoskeletal or joint disease; (iii) were taking psychotropic medication or medications that may negatively affect cognitive function, such as anticholinergics, including agents with pronounced anticholinergic properties (eg, amitriptyline), major tranquilizers (ie, typical and atypical antipsychotics), and anticonvulsants (eg, gabapentin, valproic acid, etc.); also, not expected to start or are stable on a fixed dose of antidementia medications (eg, donepezil, galantamine, etc.) during the 12-month study period; (iv) had a history indicative of carotid sinus sensitivity; (v) were living in a nursing home, extended care facility, or assisted-care facility; or (vi) were ineligible for MRI scanning. All participants provided written consent and ethics approval was acquired from the Vancouver Coastal Research Health Institute and University of British Columbia’s Clinical Research Ethics Board. Descriptive Variables Age was quantified in years and education level was noted from self-report. Standing height was measured as stretch stature to the 0.1 cm per standard protocol. Weight was measured twice to the 0.1 kg on a calibrated digital scale. Cognitive function was assessed with the MoCA and MMSE as described above. Usual Gait Speed Participants walked at their usual pace along a 4-m path and the elapsed time was recorded using a stopwatch. To avoid acceleration and deceleration effects, participants started walking 1-m before reaching the 4-m path and completed their walk 1 m beyond it. Usual gait speed (m/s) was calculated from the mean of two trials. The test-retest reliability of usual gait speed in our laboratory is 0.95 (ICC) (21). Functional MRI Acquisition All MRI was conducted at the University of British Columbia (UBC) MRI Research Center located at the UBC Hospital on a 3.0 Tesla Intera Achieva MRI Scanner (Phillips Medical Systems Canada, Markham, Ontario) using an eight-channel SENSE neurovascular coil. The session consisted of a resting-state scan with 360 dynamic images of 36 slices (3 mm thick) with the following parameters: repetition time (TR) of 2,000 ms, echo time (TE) of 30 ms, flip angle (FA) of 90 degrees, field of view (FoV) of 240 mm, acquisition matrix 80 × 80. High resolution anatomical MRI T1 images were acquired using the following parameters: 170 slices (1 mm thick), TR of 7.7 ms, TE of 3.6 ms, FA of 8 degrees, FoV of 256 mm, acquisition matrix of 256 × 200. During the resting-state scan, participants were instructed to rest with eyes open and refrain from thinking about anything in particular while remaining stationary for a total duration of 731.9 seconds (12 minutes, 11.9 seconds). All participants underwent MRI session within 2 weeks of the clinical assessment. Functional MRI Data Analysis Preprocessing Image preprocessing was carried out using tools from FSL (FMRIB’s Software Library), MATLAB (Matrix Laboratory), and toolboxes from SPM (Statistical Parametric Mapping). Excess unwanted structures (ie, bones, skull, etc.) in high-resolution T1 images were removed via optimized Brain Extraction Tool (optiBET) (22); rigid body motion correction was completed using MCFLIRT (absolute and relative mean displacement were subsequently extracted and included in the statistical analysis as covariates); spatial smoothing was carried out using Gaussian kernel of Full-Width-Half-Maximum (FWHM) 6.0 mm; temporal filtering was applied with high pass frequency cutoff of 120 seconds. Additional image artifacts were identified through Multivariate Exploratory Linear Optimized Decomposition into Independent Components (MELODIC) and further removed (an average of approximately 20 components were removed per subject). Data points corrupted with large amount of motion were determined via FSL Motion Outliers and the effects of these time points on subsequent analyses were removed using a confound matrix. Prior to data analysis, an additional low pass temporal filtering was also applied to ensure the fMRI signal fluctuated between 0.008<f<0.080 Hz, the optimal bandwidth to examine functional connectivity. Furthermore, the application of a low pass filter eliminated potentially confounding high frequency signals. Functional data were registered to personal high-resolution T1 anatomical images, which were subsequently registered to standardized 152 T1 Montreal Neurological Institute (MNI) space. Noise generated from both physiological and nonphysiological sources was removed through regression of the cerebral-spinal fluid (CSF) signal, white matter signal, and global brain signal. Global signal regression has been reported as both valid and useful step in functional connectivity analyses (23) that may potentially improve specificity (24). The first four volumes of data were discarded to account for delay of the hemodynamic response. Functional connectivity analysis Selection of regions of interest (ROI) was guided by previous work examining connectivity of networks relevant to mobility or gait speed (14). The respective MNI space coordinates for each ROI are presented in Table 1. To better understand the neural substrates of gait speed, we first examined overall inter-network connectivity of the SMN and FPN, then further examined connections between regions of interest in these networks (ie, specific ROI-ROI pair coupling). Table 1. Regions of Interest and Relative MNI Coordinates Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Note: LCB = Left cerebellum; LFEF = Left frontal eye field; LPCG = Left precentral gyrus; LPM = Left premotor; LSLOC = Left lateral occipital cortex; LVV = Left ventral visual; RCB = Right cerebellum; RFEF = Right frontal eye field; RIPS = Inferior parietal sulcus; ROI = Regions of interest; RPCG = Right precentral gyrus; RPM = Right premotor; RSLOC = Right lateral occipital cortex; RSMG = Right supramarginal gyrus; RVV = Right ventral visual; SMA = Supplementary motor area. View Large Table 1. Regions of Interest and Relative MNI Coordinates Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Note: LCB = Left cerebellum; LFEF = Left frontal eye field; LPCG = Left precentral gyrus; LPM = Left premotor; LSLOC = Left lateral occipital cortex; LVV = Left ventral visual; RCB = Right cerebellum; RFEF = Right frontal eye field; RIPS = Inferior parietal sulcus; ROI = Regions of interest; RPCG = Right precentral gyrus; RPM = Right premotor; RSLOC = Right lateral occipital cortex; RSMG = Right supramarginal gyrus; RVV = Right ventral visual; SMA = Supplementary motor area. View Large Given the SMN has left/right/neutral laterality (within the set of ROIs used in this manuscript, neutral referred to only the SMA), an average SMN connectivity was first calculated for the left SMN, right SMN, and neutral SMN. Then, the overall inter-network connectivity between the SMN and FPN was calculated by categorically taking the average of all the pairwise ROI-ROI correlation with similar spatial designation to generate an average network level correlation coefficient reflecting the connectivity between left SMN-FPN, right SMN-FPN, and neutral SMN-FPN (eg, For the left SMN-FPN, average was calculated by averaging correlation coefficient of left precentral gyrus-left ventral visual (LPCG-LVV), left cerebellum-left ventral visual (LCB- LVV) etc.; For the right SMN-FPN, average was calculated by averaging correlation coefficient of right precentral gyrus-right ventral visual (RPCG-RVV), right cerebellum-right ventral visual (RCB-RVV) etc.). Linear regression analyses at the overall inter-network connectivity level (ie, regression model for left SMN-FPN, right SMN-FPN, and neutral SMN-FPN independently) were first conducted prior to examination of the distinct pairwise ROI-ROI connections driving the observed explained variance in gait speed. For each ROI, preprocessed time-series data were extracted with 14 mm spherical regions of interest drawn around their respective MNI coordinates in standard space. Regions of interest time-series data were subsequently cross-correlated to establish functional connectivity maps of their associated neural networks, in which pairwise correlation between time-series extracted from ROI listed above was calculated. Correlation estimates were then Fisher’s z transformed to improve normality before subsequent statistical analyses. Statistical Analyses Statistical analysis was conducted using the IBM SPSS Statistic 19 for Windows (SPSS Inc., Chicago, IL). Alpha was set at p ≤ .05 for all analyses. To achieve our objective, linear regression analysis was performed. Gait speed was entered as the dependent variable; network connectivity of interest was entered as the independent variable. Age and MoCA were included as covariates in all models. Three separate regression analyses were conducted to examine the relationship between gait speed and connectivity at the network level (ie, left SMN-FPN, right SMN-FPN, and Neutral SMN-FPN). Subsequent regression analyses were conducted to examine the association between gait speed and specific pairwise ROI-ROI connectivity. Results Participants Table 2 provides descriptive characteristics of the study sample. Briefly, a total of 49 individuals were included in this cross-sectional study. Table 2 also reports the mean and median gait speed of the entire cohort differentiated by sex. As compared with results from a meta-analysis that congregated findings acquired from 41 studies and reported gait speed of a total of 23,111 adults spanning across different age (25), the mean gait speed of our male participants was similar to average males aged 70–79 years (1.22 m/s vs 1.26 m/s); also, our female participants were similar to average females aged 70–79 years (1.14 m/s vs 1.13 m/s, respectively). Table 2. Participant Characteristics (N = 49) Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Note: *MMSE = Mini-Mental Status Examination; MoCA = Montreal Cognitive Assessment. View Large Table 2. Participant Characteristics (N = 49) Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Note: *MMSE = Mini-Mental Status Examination; MoCA = Montreal Cognitive Assessment. View Large Functional Connectivity After adjusting for the covariates, we found connectivity of neutral SMN-FPN explained a statistically significant amount of variation in gait speed (p < .05; Table 3). There were no significant associations between left or right SMN-FPN and gait speed. Table 3. Network Level Linear Regression Models       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” FPN = Fronto-parietal network; MoCA = Montreal Cognitive Assessment; SMN = Sensori-motor network. View Large Table 3. Network Level Linear Regression Models       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” FPN = Fronto-parietal network; MoCA = Montreal Cognitive Assessment; SMN = Sensori-motor network. View Large Within the neutral SMN-FPN, we performed further tests for the connectivity of the five relevant pairwise ROIs, namely, supplementary motor area-inferior parietal sulcus (SMA-RIPS), SMA-bilateral ventral visual (SMA-BVV), SMA-supramarginal gyrus (SMA-RSMG), SMA-bilateral superior lateral occipital cortex (SMA-BSLOC), and SMA-bilateral frontal eye field (SMA-BFEF). Models constructed with distinct pairwise ROI-ROI connections within the neutral SMN-FPN showed that slower usual gait was associated with greater connectivity between the SMA and BVV cortices (p = .01); lower connectivity between the SMA and the BSLOC (p < .01); and lower connectivity between the SMA and the BFEF (p < .01; Table 4). Table 4. Linear Regression Model with Pairwise ROI Connectivity within Neutral SMN-FPN (N = 49)       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” BFEF = Bilateral frontal eye field; BSLOC = Bilateral occipital cortex; BVV = Bilateral ventral visual; MoCA = Montreal Cognitive Assessment; RIPS = Right inferior parietal sulcus; RSMG = Right supramarginal gyrus; SMA = Supplementary motor area. View Large Table 4. Linear Regression Model with Pairwise ROI Connectivity within Neutral SMN-FPN (N = 49)       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” BFEF = Bilateral frontal eye field; BSLOC = Bilateral occipital cortex; BVV = Bilateral ventral visual; MoCA = Montreal Cognitive Assessment; RIPS = Right inferior parietal sulcus; RSMG = Right supramarginal gyrus; SMA = Supplementary motor area. View Large Discussion In the present cross-sectional study, we demonstrated that slower usual gait speed among older adults with MCI may be characterized by altered connectivity between the SMN and the FPN. In general, slower usual gait was associated with significantly less inter-network connectivity; however, there was one ROI-ROI connection which showed increased inter-network connectivity. Further, the observed variance in gait speed can be attributed to connectivity between the supplementary motor area and regions in the FPN. Thus, we provide preliminary evidence to suggest that preservation of the functional coupling between the SMN and FPN may be critical for the maintenance of usual gait speed among older adults with MCI. Our current exploratory study was, in part, motivated by the concept of MCR. Based on our existing understanding of MCR, it is reasonable to hypothesize that older adults with MCI who have slower gait may be at greater risk for subsequent decline and progression to dementia than those without slower gait. Our current findings support this hypothesis by demonstrating that older adults identified with MCI and slower usual gait speed may have lower overall inter-network functional connectivity between the SMN and the FPN. In a previous 12-month prospective study, we demonstrated that lower connectivity between these two functional networks among community-dwelling older adults with a significant history of falls and without cognitive impairment were significantly associated with greater decline in both executive functions, as measured by the Stroop Test, and general balance and mobility, as measured by the SPPB (14). Moreover, Betzel and colleagues (26) reported age-related systematic decrease in functional connectivity across several large-scale neural networks including the SMN; importantly, the observed functional network decoupling parallels lower brain structural integrity as determined by decline in structural connectivity density in major hubs. Lending additional support to our results, Inman and colleagues (27) also found that compared with healthy older adults, less connectivity was observed between the SMN and FPN during resting-state in stroke survivors—a population that is also at significant risk for falls and dementia (28). Thus, the current and previous findings collectively support our original hypothesis that these two functionally, and anatomically (partially), overlapping networks (14) are of specific interest in understanding the neural basis for the co-occurrence of impaired mobility and cognitive function. Less connectivity between these two networks during resting state may suggest reduced motor preparatory inputs, in anticipation of motor performance, from FPN to the SMN. This, in turn, may impair mobility and increase falls risk. These observations extend our past findings by identifying the specific pair-wise ROIs that contributed to the overall inter-network functional disconnectivity. Key hubs within the SMN and FPN contribute to processing and relaying of visual sensory inputs and conveying the information into appropriate motor outputs (29), including movement planning, preparation and execution (8). In our instance, we found among older adults with MCI, slower gait speed can be explained by significantly lower connectivity between the SMA and the BSLOC, as well as lower connectivity between the SMA and the BFEF. The SMA has strong implications in gait control (30), whereas both the lateral occipital and frontal eye field regions are key components in conducting visual processing (31). This indicates that disrupted communication between key regions of the SMN and FPN may have obstructed the conveyance of visual input to motor output, thus interrupting proper gait control and execution. Additionally, we observed that greater functional coupling between the SMA and the BVV is associated with slower usual gait speed among older adults with MCI. Current understanding of the ventral visual cortex suggests that it is actively involved in the perception of motion—the ability to perceive movement of objects in a given environment (32,33). Older adults often exhibit age-related decline in the capability to detect small magnitudes of motion or distinguish the direction of displacement in space (34)—essential qualities for safe transportation. Hence, it is plausible that among older adults with MCI, gait control/execution and motion perception represent competing cognitive processes such that functionally segregating the SMA and the ventral visual cortex is a compensatory mechanism these individuals adopt to maintain gait speed. Interestingly, we found no significant association between gait speed and connectivity between the SMA and the rostral inferior parietal sulcus of the FPN, in contrast to previous research which showed aberrant inferior parietal connectivity was associated with unstable gait (35) and atrophy of the inferior parietal region was observed among individuals who displayed freezing of gait (36). It may be that our study participants had yet to progress to a similar functionally declined state as those with unstable or freezing of gait to exhibit disturbed connectivity between those particular regions. A key strength of the present study is the focus on investigating inter-network connectivity as opposed to the more commonly researched intra-network connectivity. Regardless, a few limitations should be considered. First, the study participants were not recruited from neurology clinics and were without a formal clinical diagnosis of MCI. Rather, they were categorized as having MCI based on the NIA-AA criteria. Hence, our findings cannot be generalized beyond this population. Second, our sample size may not be powered to examine all the pair-wise ROI connectivity between the SMN and FPN; hence our results may be subject to type II error. Moreover, there is much controversy in regards to global signal regression and potential observation of artificial anticorrelations. However, given the context of the networks under investigation, the effects of induced anticorrelation are less significant. Lastly, the involvement of functional neural correlates of slower gait speed among individuals with MCI may extend beyond that of resting-state functional connectivity of the SMN and FPN; a nonexhaustive list of relevant networks includes the fronto-striatal network (37), fronto-hippocampal network (38), cerebellar network (39), all of which may contribute to gait deficits. Thus, future study with larger sample should consider including resting-state connectivity analysis of more comprehensive selection of networks of interest as well as task-based fMRI paradigm to provide greater clarity. Conclusions Results of this cross-sectional investigation of slower usual gait speed among older adults with MCI highlight the potential importance of the functional coupling between the SMN and FPN. Specifically, lower connectivity between these two functional networks and their specific ROIs were predictive of slower usual gait speed. In light of evidence suggesting neural network functional connectivity may be positively altered via exercise training (40,41), such interventions may promote mobility and functional independence among those with MCI in part by maintaining or strengthening the connectivity between the SMN and FPN. Funding This study was supported by Alzheimer Society Research Program Grant (15-18) to T.L.A. and the Jack Brown and Family Alzheimer Research Foundation Society to T.L.A. C.L.H. is an Alzheimer Society Research Program Doctoral Trainee. T.L.A. is a Canada Research Chair (Tier 2) in Physical Activity, Mobility, and Cognitive Health. Conflict of Interest None reported. References 1. Petersen RC. Mild cognitive impairment. Lancet . 2006; 367: 1979. doi: 10.1016/S0140-6736(06)68881-8 Google Scholar CrossRef Search ADS PubMed  2. Feldman HH, Jacova C. Mild cognitive impairment. Am J Geriatr Psychiatry . 2005; 13: 645– 655. doi: 10.1176/appi.ajgp.13.8.645 Google Scholar CrossRef Search ADS PubMed  3. Verghese J, Robbins M, Holtzer Ret al.   Gait dysfunction in mild cognitive impairment syndromes. J Am Geriatr Soc . 2008; 56: 1244– 1251. doi: 10.1111/j.1532-5415.2008.01758.x Google Scholar CrossRef Search ADS PubMed  4. Doi T, Shimada H, Makizako Het al.   Mild cognitive impairment, slow gait, and risk of disability: a prospective study. J Am Med Dir Assoc . 2015; 16: 1082– 1086. doi: 10.1016/j.jamda.2015.07.007 Google Scholar CrossRef Search ADS PubMed  5. Montero-Odasso MM, Sarquis-Adamson Y, Speechley Met al.   Association of dual-task gait with incident dementia in mild cognitive impairment: results from the gait and brain study. JAMA Neurol . 2017; 74: 857– 865. doi: 10.1001/jamaneurol.2017.0643 Google Scholar CrossRef Search ADS PubMed  6. Onen F, Henry-Feugeas MC, Roy C, Baron G, Ravaud P. Mobility decline of unknown origin in mild cognitive impairment: an MRI-based clinical study of the pathogenesis. Brain Res . 2008; 1222: 79– 86. doi: 10.1016/j.brainres.2008.05.027 Google Scholar CrossRef Search ADS PubMed  7. Beauchet O, Allali G, Annweiler C, Verghese J. Association of motoric cognitive risk syndrome with brain volumes: results from the GAIT study. J Gerontol A Biol Sci Med Sci . 2016; 71: 1081– 1088. doi: 10.1093/gerona/glw012 Google Scholar CrossRef Search ADS PubMed  8. Staines WR, Padilla M, Knight RT. Frontal-parietal event-related potential changes associated with practising a novel visuomotor task. Brain Res Cogn Brain Res . 2002; 13: 195– 202. Google Scholar CrossRef Search ADS PubMed  9. Verghese J, Annweiler C, Ayers Eet al.   Motoric cognitive risk syndrome: multicountry prevalence and dementia risk. Neurology . 2014; 83: 718– 726. doi: 10.1212/WNL.0000000000000717 Google Scholar CrossRef Search ADS PubMed  10. Solodkin A, Hlustik P, Chen EE, Small SL. Fine modulation in network activation during motor execution and motor imagery. Cereb Cortex . 2004; 14: 1246– 1255. doi: 10.1093/cercor/bhh086 Google Scholar CrossRef Search ADS PubMed  11. Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci . 2002; 3: 201– 215. doi: 10.1038/nrn755 Google Scholar CrossRef Search ADS PubMed  12. Sridharan D, Levitin DJ, Menon V. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc Natl Acad Sci USA . 2008; 105: 12569– 12574. doi:10.1073/pnas.0800005105 Google Scholar CrossRef Search ADS PubMed  13. Wymbs NF, Bassett DS, Mucha PJ, Porter MA, Grafton ST. Differential recruitment of the sensorimotor putamen and frontoparietal cortex during motor chunking in humans. Neuron . 2012; 74: 936– 946. doi: 10.1016/j.neuron.2012.03.038 Google Scholar CrossRef Search ADS PubMed  14. Hsu CL, Voss MW, Handy TCet al.   Disruptions in brain networks of older fallers are associated with subsequent cognitive decline: a 12-month prospective exploratory study. PLoS One . 2014; 9: e93673. doi: 10.1371/journal.pone.0093673 Google Scholar CrossRef Search ADS PubMed  15. Albert MS, DeKosky ST, Dickson Det al.   The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement . 2011; 7: 270– 279. doi: 10.1016/j.jalz.2011.03.008 Google Scholar CrossRef Search ADS PubMed  16. Nasreddine ZS, Phillips NA, Bédirian Vet al.   The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc . 2005; 53: 695– 699. doi: 10.1111/j.1532-5415.2005. 53221.x Google Scholar CrossRef Search ADS PubMed  17. Gifford KA, Liu D, Lu Zet al.   The source of cognitive complaints predicts diagnostic conversion differentially among nondemented older adults. Alzheimers Dement . 10:319–327. doi:10.1016/j.jalz.2013.02.007 18. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist . 1969; 9: 179– 186. Google Scholar CrossRef Search ADS PubMed  19. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res . 1975; 12: 189– 198. Google Scholar CrossRef Search ADS PubMed  20. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia . 1971; 9: 97– 113. Google Scholar CrossRef Search ADS PubMed  21. Liu-Ambrose T, Khan KM, Donaldson MG, Eng JJ, Lord SR, McKay HA. Falls-related self-efficacy is independently associated with balance and mobility in older women with low bone mass. J Gerontol A Biol Sci Med Sci . 2006; 61: 832– 838. Google Scholar CrossRef Search ADS PubMed  22. Lutkenhoff ES, Rosenberg M, Chiang Jet al.   Optimized brain extraction for pathological brains (optiBET). PLoS One . 2014; 9: e115551. doi: 10.1371/journal.pone.0115551 Google Scholar CrossRef Search ADS PubMed  23. Fox MD, Zhang D, Snyder AZ, Raichle ME. The global signal and observed anticorrelated resting state brain networks. J Neurophysiol . 2009; 101: 3270– 3283. doi: 10.1152/jn.90777.2008 Google Scholar CrossRef Search ADS PubMed  24. Murphy K, Fox MD. Towards a consensus regarding global signal regression for resting state functional connectivity MRI. NeuroImage . 2016;154:169–173. doi:10.1016/j.neuroimage.2016.11.052 25. Bohannon RW, Williams Andrews A. Normal walking speed: a descriptive meta-analysis. Physiotherapy . 2011; 97: 182– 189. doi: 10.1016/j.physio.2010.12.004 Google Scholar CrossRef Search ADS PubMed  26. Betzel RF, Byrge L, He Y, Goñi J, Zuo XN, Sporns O. Changes in structural and functional connectivity among resting-state networks across the human lifespan. Neuroimage . 2014; 102 ( Pt 2): 345– 357. doi: 10.1016/j.neuroimage.2014.07.067 Google Scholar CrossRef Search ADS PubMed  27. Inman CS, James GA, Hamann S, Rajendra JK, Pagnoni G, Butler AJ. Altered resting-state effective connectivity of fronto-parietal motor control systems on the primary motor network following stroke. Neuroimage . 2012; 59: 227– 237. doi: 10.1016/j.neuroimage.2011.07.083 Google Scholar CrossRef Search ADS PubMed  28. Divani AA, Majidi S, Barrett AM, Noorbaloochi S, Luft AR. Consequences of stroke in community-dwelling elderly: the health and retirement study, 1998 to 2008. Stroke . 2011; 42: 1821– 1825. doi: 10.1161/STROKEAHA.110.607630 Google Scholar CrossRef Search ADS PubMed  29. Pammi VS, Miyapuram KP, Ahmed, Samejima K, Bapi RS, Doya K. Changing the structure of complex visuo-motor sequences selectively activates the fronto-parietal network. Neuroimage . 2012; 59: 1180– 1189. doi: 10.1016/j.neuroimage.2011.08.006 Google Scholar CrossRef Search ADS PubMed  30. Labriffe M, Annweiler C, Amirova LEet al.   Brain activity during mental imagery of gait versus gait-like plantar stimulation: a novel combined functional MRI paradigm to better understand cerebral gait control. Front Hum Neurosci . 2017; 11: 106. doi: 10.3389/fnhum.2017.00106 Google Scholar CrossRef Search ADS PubMed  31. Haxby JV, Grady CL, Horwitz B, Ungerleider LG, Mishkin M, Carson REet al.   Dissociation of object and spatial visual processing pathways in human extrastriate cortex. Proc Natl Acad Sci USA . 1991; 88: 1621– 1625. Google Scholar CrossRef Search ADS PubMed  32. Chapman C, Hoag R, Giaschi D. The effect of disrupting the human magnocellular pathway on global motion perception. Vision Res . 2004; 44: 2551– 2557. doi: 10.1016/j.visres.2004.06.003 Google Scholar CrossRef Search ADS PubMed  33. Hayward J, Truong G, Partanen M, Giaschi D. Effects of speed, age, and amblyopia on the perception of motion-defined form. Vision Res . 2011; 51: 2216– 2223. doi: 10.1016/j.visres.2011.08.023 Google Scholar CrossRef Search ADS PubMed  34. Trick GL, Silverman SE. Visual sensitivity to motion: age-related changes and deficits in senile dementia of the Alzheimer type. Neurology . 1991; 41: 1437– 1440. Google Scholar CrossRef Search ADS PubMed  35. Zhang L, Li TN, Yuan YSet al.   The neural basis of postural instability gait disorder subtype of Parkinson’s disease: a PET and fMRI Study. CNS Neurosci Ther . 2016; 22: 360– 367. doi: 10.1111/cns.12504 Google Scholar CrossRef Search ADS PubMed  36. Kostic VS, Agosta F, Pievani Met al.   Pattern of brain tissue loss associated with freezing of gait in Parkinson disease. Neurology . 2012; 78: 409– 416. doi: 10.1212/WNL.0b013e318245d23c Google Scholar CrossRef Search ADS PubMed  37. Beauchet O, Allali G, Montero-Odasso M, Sejdić E, Fantino B, Annweiler C. Motor phenotype of decline in cognitive performance among community-dwellers without dementia: population-based study and meta-analysis. PLoS One . 2014; 9: e99318. doi: 10.1371/journal.pone.0099318 Google Scholar CrossRef Search ADS PubMed  38. Montero-Odasso M, Muir SW, Speechley M. Dual-task complexity affects gait in people with mild cognitive impairment: the interplay between gait variability, dual tasking, and risk of falls. Arch Phys Med Rehabil . 2012; 93: 293– 299. doi: 10.1016/j.apmr.2011.08.026 Google Scholar CrossRef Search ADS PubMed  39. Manto M, Bower JM, Conforto ABet al.   Consensus paper: roles of the cerebellum in motor control–the diversity of ideas on cerebellar involvement in movement. Cerebellum . 2012; 11: 457– 487. doi: 10.1007/s12311-011-0331-9 Google Scholar CrossRef Search ADS PubMed  40. Voss MW, Prakash RS, Erickson KI, Basak C, Chaddock L, Kim JSet al.   Plasticity of brain networks in a randomized intervention trial of exercise training in older adults. Front Aging Neurosci . 2010; 2. doi:10.3389/fnagi.2010.00032 41. Hsu CL, Best JR, Davis JCet al.   Aerobic exercise promotes executive functions and impacts functional neural activity among older adults with vascular cognitive impairment. Br J Sports Med . 2018; 52: 184– 191. doi: 10.1136/bjsports-2016-096846 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences Oxford University Press

Loading next page...
 
/lp/ou_press/functional-neural-correlates-of-slower-gait-among-older-adults-with-2F5ISQWoKf
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1079-5006
eISSN
1758-535X
D.O.I.
10.1093/gerona/gly027
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Subtle, but observable, changes in mobility often exist among older adults with mild cognitive impairment (MCI). Notably, these changes are not inconsequential. Therefore, there is a strong interest to better understand the underlying neural correlates of gait slowing among older adults with MCI. In this study, we aimed to characterize patterns of functional connectivity associated with slower gait speed in older adults with MCI. Methods Forty-nine participants aged 60 years and older with MCI were included in the cross-sectional study. All participants underwent assessments of gait speed and resting state functional magnetic resonance imaging. Results In this sample of older adults with MCI, slower usual gait was characterized by altered connectivity between the sensorimotor network (SMN) and the frontoparietal network (FPN) (p < .05)—specifically, slower usual gait was associated with greater connectivity between the supplementary motor area (SMA) and the bilateral ventral visual cortices (p = .01); lower connectivity between the SMA and the bilateral superior lateral occipital cortex (p < .01); and lower connectivity between the SMA and the bilateral frontal eye field (p < .01). Conclusion Altered inter-network functional connectivity between the SMN and FPN may be a neural mechanism for slowing of gait in older adults with MCI. Mild cognitive impairment, Slower gait, Functional connectivity, Older adults Mild cognitive impairment is a clinical entity characterized by cognitive decline greater than expected for an individual’s age and education level that does not interfere with everyday function (1). While individuals with MCI are at increased risk for dementia (2), there is growing recognition that subtle, but observable, changes in mobility (ie, slowing of gait) often exist among this population (3). Notably, these changes are not inconsequential. For example, Doi and colleagues (4) showed that the combination of MCI and slow gait confers a higher risk of disability than each condition independently. Also, high dual-task gait cost (ie, [single-task gait velocity—dual-task gait velocity]/single-task gait velocity)was associated with incident dementia in MCI (5). There is a strong interest to better understand the underlying neural correlates of gait slowing in the prodromal stages of the dementia process, such as in MCI (6,7). Onen and colleagues (6) showed that periventricular leuokaraiosis was associated with slow gait among older adults with MCI. Moreover, Beauchet and colleagues (7) demonstrated that slower gait speed was associated with larger brain ventricle volume among older adults with MCI. However, whether slowing of gait speed among older adults with MCI is also associated with aberrant functional connectivity of the relevant neural networks remains largely unexplored. A better understanding of the functional neural mechanisms underlying slowing of gait speed among older adults with MCI would complement previous structural findings and provide new insights to future strategies to maintain mobility and functional independence among those with MCI. Our current exploratory study is, in part, motivated by the motoric cognitive risk syndrome (MCR) proposed by Verghese and colleagues (8). MCR is operationally defined by the presence of concomitant cognitive complaints and slow gait, without severe functional impairments or dementia (8). Studies showed that MCR is highly prevalent among older adults (8) and incidence rate of dementia was more than doubled among older adults with subjective cognitive complaint, mild cognitive impairment (MCI), and slow gait compared with those without MCR (9). Therefore, we aimed to characterize patterns of functional connectivity associated with slower gait speed in older adults with MCI. We specifically focused on two functional networks, the sensorimotor network (SMN) and the frontoparietal network (FPN). The SMN is actively involved in major aspects of movement, including motor-planning, initiation, execution, and coordination (10). The FPN is involved in top-down attentional control (11) and allocation of available neural resources to important cognitive processes (12), as well as motor planning and motor execution (13). Of relevance, our previous 12-month prospective study showed that lower inter-network connectivity between the SMN and the FPN was associated with subsequent decline in both executive functions and lower extremity physical performance in community-dwelling older adults (14). Our objective for the current study is to determine whether slower gait is characterized by lower connectivity between the SMN and FPN in older adults with MCI. Methods and Measures Study Design and Participants Forty-nine community-dwelling older adults with MCI were included in this cross-sectional study. Mild cognitive impairment was defined using the criteria provided by the National Institute of Aging and Alzheimer’s Association (NIA-AA) (15) as the following: (i) objective cognitive impairment, operationalized in our study as a Montreal Cognitive Assessment (MoCA) score < 26/30 (15); (ii) have subjective memory complaints (SMC); (iii) no significant functional impairment; and (iv) not formally diagnosed with dementia. Participants were recruited from metropolitan Vancouver via advertisement and interested individuals were screened by telephone to confirm general eligibility according to the inclusion and exclusion criteria, followed by an in-person screening session. We included those who: (i) were aged ≥ 60 years; (ii) scored < 26/30 on the MoCA (16). The MoCA is a 30-point test that covers multiple cognitive domains (16). The MoCA has been found to have good internal consistency and test-retest reliability and was able to correctly identify 90% of a large sample of MCI individuals from two different clinics (16); (iii) had SMC, defined as the self-reported feeling of memory worsening with an onset within the last 5 years, as determined by interview (17); (iv) score ≥ 6/8 on the Lawton and Brody (18) Instrumental Activities of Daily Living Scale; (v) preserved general cognition as indicated by a Mini-Mental State Examination (MMSE) (19) score ≥ 24/30; (vi) were right hand dominant as measured by the Edinburgh Handedness Inventory (20); (vii) were living independently in their own homes; (viii) had visual acuity of at least 20/40, with or without corrective lenses; and (ix) provided informed consent. We excluded those who: (i) had a formal diagnosis of neurodegenerative disease, stroke, dementia (of any type), or psychiatric condition; (ii) had clinically significant peripheral neuropathy or severe musculoskeletal or joint disease; (iii) were taking psychotropic medication or medications that may negatively affect cognitive function, such as anticholinergics, including agents with pronounced anticholinergic properties (eg, amitriptyline), major tranquilizers (ie, typical and atypical antipsychotics), and anticonvulsants (eg, gabapentin, valproic acid, etc.); also, not expected to start or are stable on a fixed dose of antidementia medications (eg, donepezil, galantamine, etc.) during the 12-month study period; (iv) had a history indicative of carotid sinus sensitivity; (v) were living in a nursing home, extended care facility, or assisted-care facility; or (vi) were ineligible for MRI scanning. All participants provided written consent and ethics approval was acquired from the Vancouver Coastal Research Health Institute and University of British Columbia’s Clinical Research Ethics Board. Descriptive Variables Age was quantified in years and education level was noted from self-report. Standing height was measured as stretch stature to the 0.1 cm per standard protocol. Weight was measured twice to the 0.1 kg on a calibrated digital scale. Cognitive function was assessed with the MoCA and MMSE as described above. Usual Gait Speed Participants walked at their usual pace along a 4-m path and the elapsed time was recorded using a stopwatch. To avoid acceleration and deceleration effects, participants started walking 1-m before reaching the 4-m path and completed their walk 1 m beyond it. Usual gait speed (m/s) was calculated from the mean of two trials. The test-retest reliability of usual gait speed in our laboratory is 0.95 (ICC) (21). Functional MRI Acquisition All MRI was conducted at the University of British Columbia (UBC) MRI Research Center located at the UBC Hospital on a 3.0 Tesla Intera Achieva MRI Scanner (Phillips Medical Systems Canada, Markham, Ontario) using an eight-channel SENSE neurovascular coil. The session consisted of a resting-state scan with 360 dynamic images of 36 slices (3 mm thick) with the following parameters: repetition time (TR) of 2,000 ms, echo time (TE) of 30 ms, flip angle (FA) of 90 degrees, field of view (FoV) of 240 mm, acquisition matrix 80 × 80. High resolution anatomical MRI T1 images were acquired using the following parameters: 170 slices (1 mm thick), TR of 7.7 ms, TE of 3.6 ms, FA of 8 degrees, FoV of 256 mm, acquisition matrix of 256 × 200. During the resting-state scan, participants were instructed to rest with eyes open and refrain from thinking about anything in particular while remaining stationary for a total duration of 731.9 seconds (12 minutes, 11.9 seconds). All participants underwent MRI session within 2 weeks of the clinical assessment. Functional MRI Data Analysis Preprocessing Image preprocessing was carried out using tools from FSL (FMRIB’s Software Library), MATLAB (Matrix Laboratory), and toolboxes from SPM (Statistical Parametric Mapping). Excess unwanted structures (ie, bones, skull, etc.) in high-resolution T1 images were removed via optimized Brain Extraction Tool (optiBET) (22); rigid body motion correction was completed using MCFLIRT (absolute and relative mean displacement were subsequently extracted and included in the statistical analysis as covariates); spatial smoothing was carried out using Gaussian kernel of Full-Width-Half-Maximum (FWHM) 6.0 mm; temporal filtering was applied with high pass frequency cutoff of 120 seconds. Additional image artifacts were identified through Multivariate Exploratory Linear Optimized Decomposition into Independent Components (MELODIC) and further removed (an average of approximately 20 components were removed per subject). Data points corrupted with large amount of motion were determined via FSL Motion Outliers and the effects of these time points on subsequent analyses were removed using a confound matrix. Prior to data analysis, an additional low pass temporal filtering was also applied to ensure the fMRI signal fluctuated between 0.008<f<0.080 Hz, the optimal bandwidth to examine functional connectivity. Furthermore, the application of a low pass filter eliminated potentially confounding high frequency signals. Functional data were registered to personal high-resolution T1 anatomical images, which were subsequently registered to standardized 152 T1 Montreal Neurological Institute (MNI) space. Noise generated from both physiological and nonphysiological sources was removed through regression of the cerebral-spinal fluid (CSF) signal, white matter signal, and global brain signal. Global signal regression has been reported as both valid and useful step in functional connectivity analyses (23) that may potentially improve specificity (24). The first four volumes of data were discarded to account for delay of the hemodynamic response. Functional connectivity analysis Selection of regions of interest (ROI) was guided by previous work examining connectivity of networks relevant to mobility or gait speed (14). The respective MNI space coordinates for each ROI are presented in Table 1. To better understand the neural substrates of gait speed, we first examined overall inter-network connectivity of the SMN and FPN, then further examined connections between regions of interest in these networks (ie, specific ROI-ROI pair coupling). Table 1. Regions of Interest and Relative MNI Coordinates Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Note: LCB = Left cerebellum; LFEF = Left frontal eye field; LPCG = Left precentral gyrus; LPM = Left premotor; LSLOC = Left lateral occipital cortex; LVV = Left ventral visual; RCB = Right cerebellum; RFEF = Right frontal eye field; RIPS = Inferior parietal sulcus; ROI = Regions of interest; RPCG = Right precentral gyrus; RPM = Right premotor; RSLOC = Right lateral occipital cortex; RSMG = Right supramarginal gyrus; RVV = Right ventral visual; SMA = Supplementary motor area. View Large Table 1. Regions of Interest and Relative MNI Coordinates Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Network  ROI  X  Y  Z  SMN  LPCG  −39  −21  55    RPCG  34  −25  53    LCB  −24  −66  −19    RCB  25  −71  −23    LPM  −16  0  57    RPM  20  −17  61    SMA  −5  −1  52  FPN  RIPS  25  −62  53    RVV  36  −62  0    LVV  −44  −60  −6    RSMG  32  −38  38    RSLOC  26  −64  54    LSLOC  −26  −60  52    RFEF  28  −4  58    LFEF  −26  −8  54  Note: LCB = Left cerebellum; LFEF = Left frontal eye field; LPCG = Left precentral gyrus; LPM = Left premotor; LSLOC = Left lateral occipital cortex; LVV = Left ventral visual; RCB = Right cerebellum; RFEF = Right frontal eye field; RIPS = Inferior parietal sulcus; ROI = Regions of interest; RPCG = Right precentral gyrus; RPM = Right premotor; RSLOC = Right lateral occipital cortex; RSMG = Right supramarginal gyrus; RVV = Right ventral visual; SMA = Supplementary motor area. View Large Given the SMN has left/right/neutral laterality (within the set of ROIs used in this manuscript, neutral referred to only the SMA), an average SMN connectivity was first calculated for the left SMN, right SMN, and neutral SMN. Then, the overall inter-network connectivity between the SMN and FPN was calculated by categorically taking the average of all the pairwise ROI-ROI correlation with similar spatial designation to generate an average network level correlation coefficient reflecting the connectivity between left SMN-FPN, right SMN-FPN, and neutral SMN-FPN (eg, For the left SMN-FPN, average was calculated by averaging correlation coefficient of left precentral gyrus-left ventral visual (LPCG-LVV), left cerebellum-left ventral visual (LCB- LVV) etc.; For the right SMN-FPN, average was calculated by averaging correlation coefficient of right precentral gyrus-right ventral visual (RPCG-RVV), right cerebellum-right ventral visual (RCB-RVV) etc.). Linear regression analyses at the overall inter-network connectivity level (ie, regression model for left SMN-FPN, right SMN-FPN, and neutral SMN-FPN independently) were first conducted prior to examination of the distinct pairwise ROI-ROI connections driving the observed explained variance in gait speed. For each ROI, preprocessed time-series data were extracted with 14 mm spherical regions of interest drawn around their respective MNI coordinates in standard space. Regions of interest time-series data were subsequently cross-correlated to establish functional connectivity maps of their associated neural networks, in which pairwise correlation between time-series extracted from ROI listed above was calculated. Correlation estimates were then Fisher’s z transformed to improve normality before subsequent statistical analyses. Statistical Analyses Statistical analysis was conducted using the IBM SPSS Statistic 19 for Windows (SPSS Inc., Chicago, IL). Alpha was set at p ≤ .05 for all analyses. To achieve our objective, linear regression analysis was performed. Gait speed was entered as the dependent variable; network connectivity of interest was entered as the independent variable. Age and MoCA were included as covariates in all models. Three separate regression analyses were conducted to examine the relationship between gait speed and connectivity at the network level (ie, left SMN-FPN, right SMN-FPN, and Neutral SMN-FPN). Subsequent regression analyses were conducted to examine the association between gait speed and specific pairwise ROI-ROI connectivity. Results Participants Table 2 provides descriptive characteristics of the study sample. Briefly, a total of 49 individuals were included in this cross-sectional study. Table 2 also reports the mean and median gait speed of the entire cohort differentiated by sex. As compared with results from a meta-analysis that congregated findings acquired from 41 studies and reported gait speed of a total of 23,111 adults spanning across different age (25), the mean gait speed of our male participants was similar to average males aged 70–79 years (1.22 m/s vs 1.26 m/s); also, our female participants were similar to average females aged 70–79 years (1.14 m/s vs 1.13 m/s, respectively). Table 2. Participant Characteristics (N = 49) Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Note: *MMSE = Mini-Mental Status Examination; MoCA = Montreal Cognitive Assessment. View Large Table 2. Participant Characteristics (N = 49) Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Variables*  Mean (SD)  Age (years)  75.4 (6.3)  Height (cm)  166.0 (11.4)  Weight (kg)  71.6 (14.7)  Sex (M/F)  19/30  MMSE (30 points max)  27.6 (1.4)  MOCA (30 points max)  22.3 (2.6)  Male Mean Gait Speed (m/s)  1.22 (0.20)  Female Mean Gait Speed (m/s)  1.14 (0.23)  Male Median Gait Speed (m/s)  1.15  Female Median Gait Speed (m/s)  1.14  Note: *MMSE = Mini-Mental Status Examination; MoCA = Montreal Cognitive Assessment. View Large Functional Connectivity After adjusting for the covariates, we found connectivity of neutral SMN-FPN explained a statistically significant amount of variation in gait speed (p < .05; Table 3). There were no significant associations between left or right SMN-FPN and gait speed. Table 3. Network Level Linear Regression Models       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” FPN = Fronto-parietal network; MoCA = Montreal Cognitive Assessment; SMN = Sensori-motor network. View Large Table 3. Network Level Linear Regression Models       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .01  Step 1  Age      −0.17  .24  MoCA      0.34  .02  Step 2  Left SMN-FPN connectivity  0.18  0.01  −0.04  .76  Model 2  0.17      .02  Step 1  Age      −0.18  .21  MoCA      0.33  .02  Step 2  Right SMN-FPN connectivity  0.19  0.02  0.12  .40  Model 3  0.17      .01  Step 1  Age      −0.23  .10  MoCA      0.36  .01  Step 2          Neutral SMN-FPN connectivity  0.25  0.08  0.28  .05  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” FPN = Fronto-parietal network; MoCA = Montreal Cognitive Assessment; SMN = Sensori-motor network. View Large Within the neutral SMN-FPN, we performed further tests for the connectivity of the five relevant pairwise ROIs, namely, supplementary motor area-inferior parietal sulcus (SMA-RIPS), SMA-bilateral ventral visual (SMA-BVV), SMA-supramarginal gyrus (SMA-RSMG), SMA-bilateral superior lateral occipital cortex (SMA-BSLOC), and SMA-bilateral frontal eye field (SMA-BFEF). Models constructed with distinct pairwise ROI-ROI connections within the neutral SMN-FPN showed that slower usual gait was associated with greater connectivity between the SMA and BVV cortices (p = .01); lower connectivity between the SMA and the BSLOC (p < .01); and lower connectivity between the SMA and the BFEF (p < .01; Table 4). Table 4. Linear Regression Model with Pairwise ROI Connectivity within Neutral SMN-FPN (N = 49)       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” BFEF = Bilateral frontal eye field; BSLOC = Bilateral occipital cortex; BVV = Bilateral ventral visual; MoCA = Montreal Cognitive Assessment; RIPS = Right inferior parietal sulcus; RSMG = Right supramarginal gyrus; SMA = Supplementary motor area. View Large Table 4. Linear Regression Model with Pairwise ROI Connectivity within Neutral SMN-FPN (N = 49)       Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01        Gait Speed    Independent Variables  R2  R2 Change  Standardized Beta  p-Value  Model 1  0.17      .03  Step 1  Age      −0.17  .22  MoCA      0.35  .02  Step 2  SMA-RIPS  0.18  0.01  0.03  .81  Model 2  0.17      <.01  Step 1  Age      −0.15  .26  MoCA      0.35  .01  Step 2  SMA-BVV  0.28  0.11  −0.33  .01  Model 3  0.17      <.01  Step 1  Age      −0.19  .18  MoCA      0.39  .01  Step 2  SMA-RSMG  0.23  0.06  0.23  .09  Model 4  0.17      <.01  Step 1  Age      −0.24  .08  MoCA      0.27  .05  Step 2  SMA-BSLOC  0.30  0.13  0.36  .01  Model 5  0.17      <.01  Step 1  Age      −0.29  .02  MoCA      0.33  .01  Step 2  SMA-BFEF  0.41  0.24  0.50  <.01  Note: Variables shown under “Step 1” are included as covariates in “Step 2.” BFEF = Bilateral frontal eye field; BSLOC = Bilateral occipital cortex; BVV = Bilateral ventral visual; MoCA = Montreal Cognitive Assessment; RIPS = Right inferior parietal sulcus; RSMG = Right supramarginal gyrus; SMA = Supplementary motor area. View Large Discussion In the present cross-sectional study, we demonstrated that slower usual gait speed among older adults with MCI may be characterized by altered connectivity between the SMN and the FPN. In general, slower usual gait was associated with significantly less inter-network connectivity; however, there was one ROI-ROI connection which showed increased inter-network connectivity. Further, the observed variance in gait speed can be attributed to connectivity between the supplementary motor area and regions in the FPN. Thus, we provide preliminary evidence to suggest that preservation of the functional coupling between the SMN and FPN may be critical for the maintenance of usual gait speed among older adults with MCI. Our current exploratory study was, in part, motivated by the concept of MCR. Based on our existing understanding of MCR, it is reasonable to hypothesize that older adults with MCI who have slower gait may be at greater risk for subsequent decline and progression to dementia than those without slower gait. Our current findings support this hypothesis by demonstrating that older adults identified with MCI and slower usual gait speed may have lower overall inter-network functional connectivity between the SMN and the FPN. In a previous 12-month prospective study, we demonstrated that lower connectivity between these two functional networks among community-dwelling older adults with a significant history of falls and without cognitive impairment were significantly associated with greater decline in both executive functions, as measured by the Stroop Test, and general balance and mobility, as measured by the SPPB (14). Moreover, Betzel and colleagues (26) reported age-related systematic decrease in functional connectivity across several large-scale neural networks including the SMN; importantly, the observed functional network decoupling parallels lower brain structural integrity as determined by decline in structural connectivity density in major hubs. Lending additional support to our results, Inman and colleagues (27) also found that compared with healthy older adults, less connectivity was observed between the SMN and FPN during resting-state in stroke survivors—a population that is also at significant risk for falls and dementia (28). Thus, the current and previous findings collectively support our original hypothesis that these two functionally, and anatomically (partially), overlapping networks (14) are of specific interest in understanding the neural basis for the co-occurrence of impaired mobility and cognitive function. Less connectivity between these two networks during resting state may suggest reduced motor preparatory inputs, in anticipation of motor performance, from FPN to the SMN. This, in turn, may impair mobility and increase falls risk. These observations extend our past findings by identifying the specific pair-wise ROIs that contributed to the overall inter-network functional disconnectivity. Key hubs within the SMN and FPN contribute to processing and relaying of visual sensory inputs and conveying the information into appropriate motor outputs (29), including movement planning, preparation and execution (8). In our instance, we found among older adults with MCI, slower gait speed can be explained by significantly lower connectivity between the SMA and the BSLOC, as well as lower connectivity between the SMA and the BFEF. The SMA has strong implications in gait control (30), whereas both the lateral occipital and frontal eye field regions are key components in conducting visual processing (31). This indicates that disrupted communication between key regions of the SMN and FPN may have obstructed the conveyance of visual input to motor output, thus interrupting proper gait control and execution. Additionally, we observed that greater functional coupling between the SMA and the BVV is associated with slower usual gait speed among older adults with MCI. Current understanding of the ventral visual cortex suggests that it is actively involved in the perception of motion—the ability to perceive movement of objects in a given environment (32,33). Older adults often exhibit age-related decline in the capability to detect small magnitudes of motion or distinguish the direction of displacement in space (34)—essential qualities for safe transportation. Hence, it is plausible that among older adults with MCI, gait control/execution and motion perception represent competing cognitive processes such that functionally segregating the SMA and the ventral visual cortex is a compensatory mechanism these individuals adopt to maintain gait speed. Interestingly, we found no significant association between gait speed and connectivity between the SMA and the rostral inferior parietal sulcus of the FPN, in contrast to previous research which showed aberrant inferior parietal connectivity was associated with unstable gait (35) and atrophy of the inferior parietal region was observed among individuals who displayed freezing of gait (36). It may be that our study participants had yet to progress to a similar functionally declined state as those with unstable or freezing of gait to exhibit disturbed connectivity between those particular regions. A key strength of the present study is the focus on investigating inter-network connectivity as opposed to the more commonly researched intra-network connectivity. Regardless, a few limitations should be considered. First, the study participants were not recruited from neurology clinics and were without a formal clinical diagnosis of MCI. Rather, they were categorized as having MCI based on the NIA-AA criteria. Hence, our findings cannot be generalized beyond this population. Second, our sample size may not be powered to examine all the pair-wise ROI connectivity between the SMN and FPN; hence our results may be subject to type II error. Moreover, there is much controversy in regards to global signal regression and potential observation of artificial anticorrelations. However, given the context of the networks under investigation, the effects of induced anticorrelation are less significant. Lastly, the involvement of functional neural correlates of slower gait speed among individuals with MCI may extend beyond that of resting-state functional connectivity of the SMN and FPN; a nonexhaustive list of relevant networks includes the fronto-striatal network (37), fronto-hippocampal network (38), cerebellar network (39), all of which may contribute to gait deficits. Thus, future study with larger sample should consider including resting-state connectivity analysis of more comprehensive selection of networks of interest as well as task-based fMRI paradigm to provide greater clarity. Conclusions Results of this cross-sectional investigation of slower usual gait speed among older adults with MCI highlight the potential importance of the functional coupling between the SMN and FPN. Specifically, lower connectivity between these two functional networks and their specific ROIs were predictive of slower usual gait speed. In light of evidence suggesting neural network functional connectivity may be positively altered via exercise training (40,41), such interventions may promote mobility and functional independence among those with MCI in part by maintaining or strengthening the connectivity between the SMN and FPN. Funding This study was supported by Alzheimer Society Research Program Grant (15-18) to T.L.A. and the Jack Brown and Family Alzheimer Research Foundation Society to T.L.A. C.L.H. is an Alzheimer Society Research Program Doctoral Trainee. T.L.A. is a Canada Research Chair (Tier 2) in Physical Activity, Mobility, and Cognitive Health. Conflict of Interest None reported. References 1. Petersen RC. Mild cognitive impairment. Lancet . 2006; 367: 1979. doi: 10.1016/S0140-6736(06)68881-8 Google Scholar CrossRef Search ADS PubMed  2. Feldman HH, Jacova C. Mild cognitive impairment. Am J Geriatr Psychiatry . 2005; 13: 645– 655. doi: 10.1176/appi.ajgp.13.8.645 Google Scholar CrossRef Search ADS PubMed  3. Verghese J, Robbins M, Holtzer Ret al.   Gait dysfunction in mild cognitive impairment syndromes. J Am Geriatr Soc . 2008; 56: 1244– 1251. doi: 10.1111/j.1532-5415.2008.01758.x Google Scholar CrossRef Search ADS PubMed  4. Doi T, Shimada H, Makizako Het al.   Mild cognitive impairment, slow gait, and risk of disability: a prospective study. J Am Med Dir Assoc . 2015; 16: 1082– 1086. doi: 10.1016/j.jamda.2015.07.007 Google Scholar CrossRef Search ADS PubMed  5. Montero-Odasso MM, Sarquis-Adamson Y, Speechley Met al.   Association of dual-task gait with incident dementia in mild cognitive impairment: results from the gait and brain study. JAMA Neurol . 2017; 74: 857– 865. doi: 10.1001/jamaneurol.2017.0643 Google Scholar CrossRef Search ADS PubMed  6. Onen F, Henry-Feugeas MC, Roy C, Baron G, Ravaud P. Mobility decline of unknown origin in mild cognitive impairment: an MRI-based clinical study of the pathogenesis. Brain Res . 2008; 1222: 79– 86. doi: 10.1016/j.brainres.2008.05.027 Google Scholar CrossRef Search ADS PubMed  7. Beauchet O, Allali G, Annweiler C, Verghese J. Association of motoric cognitive risk syndrome with brain volumes: results from the GAIT study. J Gerontol A Biol Sci Med Sci . 2016; 71: 1081– 1088. doi: 10.1093/gerona/glw012 Google Scholar CrossRef Search ADS PubMed  8. Staines WR, Padilla M, Knight RT. Frontal-parietal event-related potential changes associated with practising a novel visuomotor task. Brain Res Cogn Brain Res . 2002; 13: 195– 202. Google Scholar CrossRef Search ADS PubMed  9. Verghese J, Annweiler C, Ayers Eet al.   Motoric cognitive risk syndrome: multicountry prevalence and dementia risk. Neurology . 2014; 83: 718– 726. doi: 10.1212/WNL.0000000000000717 Google Scholar CrossRef Search ADS PubMed  10. Solodkin A, Hlustik P, Chen EE, Small SL. Fine modulation in network activation during motor execution and motor imagery. Cereb Cortex . 2004; 14: 1246– 1255. doi: 10.1093/cercor/bhh086 Google Scholar CrossRef Search ADS PubMed  11. Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci . 2002; 3: 201– 215. doi: 10.1038/nrn755 Google Scholar CrossRef Search ADS PubMed  12. Sridharan D, Levitin DJ, Menon V. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc Natl Acad Sci USA . 2008; 105: 12569– 12574. doi:10.1073/pnas.0800005105 Google Scholar CrossRef Search ADS PubMed  13. Wymbs NF, Bassett DS, Mucha PJ, Porter MA, Grafton ST. Differential recruitment of the sensorimotor putamen and frontoparietal cortex during motor chunking in humans. Neuron . 2012; 74: 936– 946. doi: 10.1016/j.neuron.2012.03.038 Google Scholar CrossRef Search ADS PubMed  14. Hsu CL, Voss MW, Handy TCet al.   Disruptions in brain networks of older fallers are associated with subsequent cognitive decline: a 12-month prospective exploratory study. PLoS One . 2014; 9: e93673. doi: 10.1371/journal.pone.0093673 Google Scholar CrossRef Search ADS PubMed  15. Albert MS, DeKosky ST, Dickson Det al.   The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement . 2011; 7: 270– 279. doi: 10.1016/j.jalz.2011.03.008 Google Scholar CrossRef Search ADS PubMed  16. Nasreddine ZS, Phillips NA, Bédirian Vet al.   The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc . 2005; 53: 695– 699. doi: 10.1111/j.1532-5415.2005. 53221.x Google Scholar CrossRef Search ADS PubMed  17. Gifford KA, Liu D, Lu Zet al.   The source of cognitive complaints predicts diagnostic conversion differentially among nondemented older adults. Alzheimers Dement . 10:319–327. doi:10.1016/j.jalz.2013.02.007 18. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist . 1969; 9: 179– 186. Google Scholar CrossRef Search ADS PubMed  19. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res . 1975; 12: 189– 198. Google Scholar CrossRef Search ADS PubMed  20. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia . 1971; 9: 97– 113. Google Scholar CrossRef Search ADS PubMed  21. Liu-Ambrose T, Khan KM, Donaldson MG, Eng JJ, Lord SR, McKay HA. Falls-related self-efficacy is independently associated with balance and mobility in older women with low bone mass. J Gerontol A Biol Sci Med Sci . 2006; 61: 832– 838. Google Scholar CrossRef Search ADS PubMed  22. Lutkenhoff ES, Rosenberg M, Chiang Jet al.   Optimized brain extraction for pathological brains (optiBET). PLoS One . 2014; 9: e115551. doi: 10.1371/journal.pone.0115551 Google Scholar CrossRef Search ADS PubMed  23. Fox MD, Zhang D, Snyder AZ, Raichle ME. The global signal and observed anticorrelated resting state brain networks. J Neurophysiol . 2009; 101: 3270– 3283. doi: 10.1152/jn.90777.2008 Google Scholar CrossRef Search ADS PubMed  24. Murphy K, Fox MD. Towards a consensus regarding global signal regression for resting state functional connectivity MRI. NeuroImage . 2016;154:169–173. doi:10.1016/j.neuroimage.2016.11.052 25. Bohannon RW, Williams Andrews A. Normal walking speed: a descriptive meta-analysis. Physiotherapy . 2011; 97: 182– 189. doi: 10.1016/j.physio.2010.12.004 Google Scholar CrossRef Search ADS PubMed  26. Betzel RF, Byrge L, He Y, Goñi J, Zuo XN, Sporns O. Changes in structural and functional connectivity among resting-state networks across the human lifespan. Neuroimage . 2014; 102 ( Pt 2): 345– 357. doi: 10.1016/j.neuroimage.2014.07.067 Google Scholar CrossRef Search ADS PubMed  27. Inman CS, James GA, Hamann S, Rajendra JK, Pagnoni G, Butler AJ. Altered resting-state effective connectivity of fronto-parietal motor control systems on the primary motor network following stroke. Neuroimage . 2012; 59: 227– 237. doi: 10.1016/j.neuroimage.2011.07.083 Google Scholar CrossRef Search ADS PubMed  28. Divani AA, Majidi S, Barrett AM, Noorbaloochi S, Luft AR. Consequences of stroke in community-dwelling elderly: the health and retirement study, 1998 to 2008. Stroke . 2011; 42: 1821– 1825. doi: 10.1161/STROKEAHA.110.607630 Google Scholar CrossRef Search ADS PubMed  29. Pammi VS, Miyapuram KP, Ahmed, Samejima K, Bapi RS, Doya K. Changing the structure of complex visuo-motor sequences selectively activates the fronto-parietal network. Neuroimage . 2012; 59: 1180– 1189. doi: 10.1016/j.neuroimage.2011.08.006 Google Scholar CrossRef Search ADS PubMed  30. Labriffe M, Annweiler C, Amirova LEet al.   Brain activity during mental imagery of gait versus gait-like plantar stimulation: a novel combined functional MRI paradigm to better understand cerebral gait control. Front Hum Neurosci . 2017; 11: 106. doi: 10.3389/fnhum.2017.00106 Google Scholar CrossRef Search ADS PubMed  31. Haxby JV, Grady CL, Horwitz B, Ungerleider LG, Mishkin M, Carson REet al.   Dissociation of object and spatial visual processing pathways in human extrastriate cortex. Proc Natl Acad Sci USA . 1991; 88: 1621– 1625. Google Scholar CrossRef Search ADS PubMed  32. Chapman C, Hoag R, Giaschi D. The effect of disrupting the human magnocellular pathway on global motion perception. Vision Res . 2004; 44: 2551– 2557. doi: 10.1016/j.visres.2004.06.003 Google Scholar CrossRef Search ADS PubMed  33. Hayward J, Truong G, Partanen M, Giaschi D. Effects of speed, age, and amblyopia on the perception of motion-defined form. Vision Res . 2011; 51: 2216– 2223. doi: 10.1016/j.visres.2011.08.023 Google Scholar CrossRef Search ADS PubMed  34. Trick GL, Silverman SE. Visual sensitivity to motion: age-related changes and deficits in senile dementia of the Alzheimer type. Neurology . 1991; 41: 1437– 1440. Google Scholar CrossRef Search ADS PubMed  35. Zhang L, Li TN, Yuan YSet al.   The neural basis of postural instability gait disorder subtype of Parkinson’s disease: a PET and fMRI Study. CNS Neurosci Ther . 2016; 22: 360– 367. doi: 10.1111/cns.12504 Google Scholar CrossRef Search ADS PubMed  36. Kostic VS, Agosta F, Pievani Met al.   Pattern of brain tissue loss associated with freezing of gait in Parkinson disease. Neurology . 2012; 78: 409– 416. doi: 10.1212/WNL.0b013e318245d23c Google Scholar CrossRef Search ADS PubMed  37. Beauchet O, Allali G, Montero-Odasso M, Sejdić E, Fantino B, Annweiler C. Motor phenotype of decline in cognitive performance among community-dwellers without dementia: population-based study and meta-analysis. PLoS One . 2014; 9: e99318. doi: 10.1371/journal.pone.0099318 Google Scholar CrossRef Search ADS PubMed  38. Montero-Odasso M, Muir SW, Speechley M. Dual-task complexity affects gait in people with mild cognitive impairment: the interplay between gait variability, dual tasking, and risk of falls. Arch Phys Med Rehabil . 2012; 93: 293– 299. doi: 10.1016/j.apmr.2011.08.026 Google Scholar CrossRef Search ADS PubMed  39. Manto M, Bower JM, Conforto ABet al.   Consensus paper: roles of the cerebellum in motor control–the diversity of ideas on cerebellar involvement in movement. Cerebellum . 2012; 11: 457– 487. doi: 10.1007/s12311-011-0331-9 Google Scholar CrossRef Search ADS PubMed  40. Voss MW, Prakash RS, Erickson KI, Basak C, Chaddock L, Kim JSet al.   Plasticity of brain networks in a randomized intervention trial of exercise training in older adults. Front Aging Neurosci . 2010; 2. doi:10.3389/fnagi.2010.00032 41. Hsu CL, Best JR, Davis JCet al.   Aerobic exercise promotes executive functions and impacts functional neural activity among older adults with vascular cognitive impairment. Br J Sports Med . 2018; 52: 184– 191. doi: 10.1136/bjsports-2016-096846 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Journal

The Journals of Gerontology Series A: Biomedical Sciences and Medical SciencesOxford University Press

Published: Feb 17, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off