Introduction
Walking is a physiological milestone of normal human neurodevelopment as well as a crucial part of daily life, and it is no longer regarded solely as a physical factor [
1,
2]. Indeed, gait control is a complex process that involves the integration of motor, perceptual, and cognitive processes [
2]. Specifically, the executive functions, including attentional control, cognitive flexibility, psychomotor processing, inhibition, and goal setting, share with motor abilities and gait control a widespread brain network of prefrontal cortical and subcortical regions [
3]. These include the prefrontal cortex, the medial temporal lobe, and the nigrostriatal system but also the size of ventricles, cerebellum, white matter tracts, and the parietal lobes [
2,
4,
5].
A consistent body of studies reported a relationship between gait abnormalities and early signs of cognitive decline among cognitively healthy participants [
2,
6,
7]. Gait abnormalities include disorders that result in slow, unsteady, staggering, shuffling, and/or asymmetrical walking due to neurological, musculoskeletal, and/or other acquired medical conditions [
8‐
10]. Disorders of gait can be evaluated through clinical visual inspection or through quantitative parameters (e.g., speed, stride length, swing, and stance time) that reflect the observed gait abnormality [
11]. The former is a useful and reliable method in everyday clinical practice, whereas the latter requires technological equipment that can be used to further differentiate individuals according to their cognitive status [
8,
11].
For instance, there is considerable evidence showing that gait abnormalities could predict a cognitive decline over time measured with the Digit Symbol Substitution Test (DSST) [
6,
7,
12‐
16], a measure of psychomotor speed and attention [
17]. Furthermore, gait abnormalities predict the decline in divided attention and cognitive flexibility, as measured with the Trail Making Test part B (TMT-B) [
6,
7]. The presence of gait abnormalities is also longitudinally associated with a decline in global cognition tests [
6,
7]. Indeed, gait disorders have been identified as one of the factors associated with the development of dementia [
8].
Early evidence by Camicioli and colleagues [
18] found that slow gait is evident on clinical examination before or coincident with the development of cognitive impairment in healthy older people. Even though some diagnostic criteria include the presence of gait disturbances in the exclusion criteria of Alzheimer’s disease (AD) [
19], a recent meta-analysis [
20] suggested that gait performance predicts AD dementia (Hazard Ratio—HR = 1.03). The longitudinal study of Kuate-Tegueu and co-authors [
21] showed that gait speed (HR = 1.2) and Trail Making Test part A (HR = 1.4; TMT-A), which requires complex visual scanning and psychomotor speed [
22], were associated with incident AD. This is consistent with a recent study showing that gait abnormalities (slower gait speed, lower cadence, longer double support time, and greater stance time variability) have been associated with AD neuropathology (i.e., beta-amyloid) in cognitively healthy older individuals [
23]. Another study [
24] showed that in older people, cerebral deposition of beta-amyloid is associated with slower gait speed and lower limbs functioning. A recent large multi-database study [
25] showed that higher gait variability can discriminate AD from other neurodegenerative diseases (e.g., Parkinson’s disease, frontotemporal dementia, dementia with Lewy bodies). The authors concluded that high gait variability could be a marker for cortical-related cognitive dysfunctions which alter both cognition and gait control.
The concept of Mild Cognitive Impairment (MCI) has offered a unique window to study the development of AD. MCI is the transitional condition between normal and pathological cognitive aging [
26]. In particular, the amnestic MCI (aMCI) type, namely individuals who experience more memory loss than expected for their age and education and are more likely to develop AD than the non-amnestic type (naMCI), has received increasing attention in the last decade [
26]. In patients with MCI, the prevalence of slow gait or neurological gait abnormalities reaches 46%, almost threefold higher than in healthy older adults without MCI; in addition, neurological gait disorders were more common in patients with aMCI than in those with naMCI [
11]. Interestingly, a growing body of studies revealed that gait disorders may be a risk factor for cognitive deterioration in this population. For instance, Doi and colleagues [
27] found that patients with MCI and slow gait reported greater cognitive deficits on a comprehensive neuropsychological battery, including the Mini-Mental State Examination (MMSE), DSST, TMT-A, and TMT-B, compared to MCI without slow gait, healthy older people with slow gait and without slow gait.
Literature showed different longitudinal studies on MCI or aMCI population, in which the influence of gait abnormalities on cognition was analysed. Buracchio and colleagues [
28] demonstrated that a decline in gait speed occurred about 12 years before MCI, therefore it may be a sensitive marker of cognitive change. Furthermore, individuals with slow gait had 7 times the risk of progressing to dementia and a higher attributable risk than those with cognitive decline alone, who had 3 times the risk of progressing [
29]. Another study showed that slower maximum walking speed and longer time on the Test Timed Up and Go test were predictive of cognitive decline, as assessed according to the Montreal Cognitive Assessment-Japan score decline [
30]. Evidence indicated that aMCI who developed AD had lower gait speed than those who did not develop AD. Both gait speed and gait variability could be markers to early identify aMCI at risk to progress to AD [
31]. Also, the study of Tian and co-authors [
32] confirmed that slower baseline gait speed was associated with a higher hazard of developing aMCI/AD. A study [
33] showed that the presence of at least one copy of apolipoprotein E polymorphism ε4 allele in MCI is longitudinally associated with a decline in both gait performance and global cognition. Intriguingly, one randomized controlled trial [
34] showed that administering donepezil to improve cholinergic neurotransmission in MCI improves gait speed during dual-task, possibly due to an enhancement in frontal functions.
Despite convincing evidence that specific gait parameters can be a risk factor for dementia conversion, no previous studies have investigated which neuropsychological tests would show a greater decline among aMCI patients with and without gait disorders, and what is the prognostic relevance of gait and related neuropsychological functions.
We want to explore if individuals with abnormal gait at the beginning of the study due to neurological (e.g., slow, broad-based, unsteady, stooped, or asymmetrical gait) or musculoskeletal (e.g., injury, pain) deficits, will show a steeper decline on a set of neuropsychological tests, possibly the ones that assess, in addition to global cognition, psychomotor speed, attention, and/or executive functions. In addition, we expect that these findings are the result of the gait profile (i.e., abnormal vs. normal) itself and its possible neural altered mechanisms (e.g., AD pathology) rather than functional, medical (e.g., cerebrovascular accidents, multimorbidity, polypharmacy), and cognitive confounding factors at baseline. Secondarily, we also wanted to explore if the presence of gait disorders is associated with gait related brain measures (e.g., medial temporal regions volume, ventricles size, brain metabolism) Lastly, we want to explore what is the prognostic impact of gait disorders and the significantly affected tests on conversion to AD dementia in aMCI.
Discussion
The present study aimed at investigating the longitudinal trends in cognitive functions in the aMCI population, based on gait profiles at baseline and the prognostic relevance of gait and its related neuropsychological functions.
To our knowledge, there are no longitudinal studies in aMCI that evidence the influence of gait on a set of cognitive tests depending on gait disorders at a specific time point. This study analyzes the effect of the presence of gait disorders on repeated cognitive assessments and structural/functional brain imaging over time and evaluate the prognostic relevance of gait disorders and the significantly affected neuropsychological tests.
We found that gait abnormalities detected by a routinely neurological gait examination are associated with different trends in cognitive tests over time in aMCI. More precisely, when compared to the normal gait group, attention (DSST) and global cognition (MMSE) tests declined faster in the abnormal gait group compared to the normal gait group. Importantly, TMT part A and B uniquely declined over time in the abnormal gait group but not in the normal gait group. In addition, we showed that only ventricles volumes declined faster in the abnormal gait group, however this measure declined also for the normal gait group. Importantly, the presence of gait disorders (HR = 1.7) and the decline in the performance of two (MMSE, HR = 1.09; DSST, HR = 1.03) gait-related cognitive tests were associated with a greater risk of AD dementia conversion in the global aMCI ADNI population.
Our explorative analysis concerning the effect of the presence of gait disorders on a set of cognitive tests showed that some tests decline faster in aMCI with gait abnormalities than in aMCI with a normal gait, whereas other functions decline independently of this grouping variable. Crucially, we showed that psychomotor speed (TMT-A) and divided attention/cognitive flexibility (TMT-B) seem to be uniquely affected by gait abnormalities in aMCI. Less specific tests of cognitive functioning (MMSE, DSST), despite declining faster in the abnormal gait group, are not sensitive to gait disorders. Our findings regarding the link between gait and DSST and TMT-A are supported by previous studies on aging [
6,
7,
12‐
16,
21] and MCI [
27]. Concerning the results of TMT-B, our finding is in line with previous research on aging [
6,
7] and MCI [
27]. This suggests that psychomotor speed, attention, and executive functions are affected by gait and possibly by neuropathological changes in aMCI. A recent study found that additional frontal-executive dysfunction in aMCI increased the risk of dementia conversion compared with single-domain aMCI and that those patients showed diffuse cortical thinning, especially in the frontal areas [
47]. Another research demonstrated that in aMCI the probability of developing dementia in the Alzheimer’s clinical syndrome a year later was significantly predicted by dysexecutive deficits [
48]. Regarding global cognition, our result is in line with previous research that showed that gait abnormalities in healthy older people are longitudinally associated with global cognition performance [
6,
7,
30], this has also been found in MCI [
27]. Here, we extended the present literature by showing that such tests decline in aMCI with gait abnormalities and found that TMT-AB could be a sensible test to gait disorders in aMCI, rather than more general cognitive tests like MMSE or DSST. Indeed, the TMT-AB test is considered a core neuropsychological test to assess cognition and mobility in aging by the Canadian Consortium on Neurodegeneration in Aging [
49].
Conversely, AD-related global cognition (ADAS-13) and auditory-verbal memory (immediate recall) decline over time independently of the presence of abnormal or normal gait. This suggests that in aMCI the decline in global cognition and memory is due to the presence of specific pathological changes potentially associated with AD [
50], rather than with alterations in gait-related brain regions and functions.
Regarding brain alterations related to gait profiles, we showed that ventricles size increases faster in the abnormal compared to the normal gait group. FDG-PET, hippocampal size, and medial temporal lobe size declined regardless of the grouping variable. It could be argued that gait abnormalities are longitudinally associated with faster enlargement of the ventricles because of cortical brain atrophy [
51]. For instance, a study [
52] showed that enlargement of temporal horns and posterior portion of the ventricles is associated with gait instability in healthy older adults. It might be possible that the faster enlargement in the ventricles in the abnormal gait group is due to widespread cortical atrophy and possibly cognitive decline in executive functions/attention. Interestingly, the aMCI group with abnormal gait examination showed larger ventricles and lower FDG-PET metabolism at baseline, suggesting a link between gait disorder and ventricular size and temporoparietal brain metabolism.
In addition, we showed that the presence of gait disorders and MMSE and DSST score decline increased the risk of developing dementia. In accordance with our results, previous studies demonstrated that, in MCI, gait speed may be a sensitive marker of cognitive changes [
28] and that individuals with deficits in gait velocity had a higher risk of progressing to dementia [
29]. Furthermore, in aMCI, gait speed and gait variability may be markers for early detection of the likelihood of progression to AD [
31]. Besides, slower baseline gait speed was associated with a higher hazard of developing aMCI/AD [
32]. Recently, a large multicenter study [
25] showed that higher gait variability could be a marker of AD. In addition to previous studies, we showed that gait-related measures decline, and in particular, the score of the MMSE and DSST tests, are risk markers of future dementia conversion. In contrast to our prediction, the TMT-A and TMT-B were not a prognostic marker for dementia conversion in aMCI despite being negatively affected by gait disorders. Indeed, the study by Kuate-Tegueu and co-authors [
21] demonstrated that a low TMT-A score increased the risk of developing dementia; however, this study did not focus on aMCI but rather on healthy older persons.
Lastly, the presence of gait disorders hampers the autonomy (FAQ) of the individual. Indeed, we found that baseline reduced autonomy (higher FAQ score) and larger ventricles negatively influenced the DSST performance, whereas higher metabolism of the angular, temporal, and posterior cingulate regions (FDG-PET) positively influenced the score in the DSST performance. Conversely, the opposite directions were found for the TMT-A and TMT-B. The number of medical conditions at baseline was not associated with the decline in these tests, possibly because its effect is covered by the other covariates.
The findings of this work are also interesting considering the novel theoretical framework emphasizing the role of embodiment processes in aging. According to the embodiment theories, executive functions/attention and psychomotor speed are grounded in the ability to control and plan motor actions [
53]. The notion that such functions are embodied in the sensorimotor system is also supported by a shared network of brain regions between motor and executive functions [
3]. Indeed, some models of embodiment in aging suggest the importance of bodily information for the maintenance of cognitive abilities [
54‐
56]. Spared motor processing in AD is thought to support cognitive abilities that are not affected by the disease in the early stages, such as motor planning and language comprehension [
57]. Considering this theoretical proposal, we showed that gait in prodromal AD could affect neuropsychological functions related to motor execution and control of gait, raising the issue of the important role of bodily information on cognition. In addition, we found that the presence of gait disorders and executive functions/attention decline are risk factors for developing dementia. This hints that embodiment markers can be useful to detect individuals at greater risk of developing dementia even when the risk factor (i.e., gait and executive decline) is not a core clinical presentation of AD [
19].
This study has certain limitations that must be considered. First, within the aMCI group, there is a strong numerical unbalance between normal gait and abnormal gait group sample size. Due to this disparity, appropriate statistical methods were used accordingly. Second, the neurological gait examination carried out according to the ADNI clinical protocol is categorical; a continuous outcome for walking performance could have improved our results and highlighted subtle changes also in other cognitive domains in aMCI [
6]. Future research in the field of cognitive neuroscience could study embodiment with a specific motor task [
57] in combination with neurophysiological instruments to deepen the understanding of embodiment markers in AD and aMCI. From the clinical point of view, future studies could design preventive cognitive training on executive functions/attention and psychomotor speed in aMCI with an abnormal gait. Indeed, gait and dual-task interventions should be tested to prevent motor and cognitive decline [
58]. Finally, we propose that cognitive decline be monitored using DSST, TMT-A, and TMT-B in patients with aberrant gait aMCI so that test findings can detect probable neurophysiological alterations and signal faster cognitive decline and possible dementia.
Acknowledgements
Data collection and sharing for this project were funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (
www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.
Alzheimer’s Disease Neuroimaging Initative
Lisa C. Silbert1, Betty Lind1, Rachel Crissey1, Jeffrey A. Kaye1, Raina Carter1, Sara Dolen1, Joseph Quinn1, Lon S. Schneider2, Sonia Pawluczyk2, Mauricio Becerra2, Liberty Teodoro2, Karen Dagerman2, Bryan M. Spann2, James Brewer3, Adam Fleisher3, Helen Vanderswag3, Jaimie Ziolkowski4, Judith L. Heidebrink4, Lisa Zbizek-Nulph4, Joanne L. Lord4, Colleen S. Albers5, Ronald Petersen5, Sara S. Mason5, David Knopman5, Kris Johnson5, Javier Villanueva-Meyer6, Valory Pavlik6, Nathaniel Pacini6, Ashley Lamb6, Joseph S. Kass6, Rachelle S. Doody6, Victoria Shibley6, Munir Chowdhury6, Susan Rountree6, Mimi Dang6, Yaakov Stern7, Lawrence S. Honig7, Akiva Mintz7, Beau Ances8, John C. Morris8, David Winkfield8, Maria Carroll8, Georgia Stobbs-Cucchi8, Angela Oliver8, Mary L. Creech8, Mark A. Mintun8, Stacy Schneider8, David Geldmacher9, Marissa Natelson Love9, Randall Griffith9, David Clark9, John Brockington9, Daniel Marson9, Hillel Grossman10, Martin A. Goldstein10, Jonathan Greenberg10, Effie Mitsis10, Raj C. Shah11, Melissa Lamar11, Patricia Samuels11, Ranjan Duara12, Maria T. Greig-Custo12, Rosemarie Rodriguez12, Marilyn Albert13, Chiadi Onyike13, Leonie Farrington13, Scott Rudow13, Rottislav Brichko13, Stephanie Kielb13, Amanda Smith14, Balebail Ashok Raj14, Kristin Fargher14, Martin Sadowski15, Thomas Wisniewski15, Melanie Shulman15, Arline Faustin15, Julia Rao15, Karen M. Castro15, Anaztasia Ulysse15, Shannon Chen15, P. Murali Doraiswamy16, Jeffrey R. Petrella16, Olga James16, Terence Z. Wong16, Salvador Borges-Neto16, Jason H. Karlawish17, David A. Wolk17, Sanjeev Vaishnavi17, Christopher M. Clark17, Steven E. Arnold17, Charles D. Smith18, Gregory A. Jicha18, Riham El Khouli18, Flavius D. Raslau18, Oscar L. Lopez19, MaryAnn Oakley19, Donna M. Simpson19, Anton P. Porsteinsson20, Kim Martin20, Nancy Kowalski20, Melanie Keltz20, Bonnie S. Goldstein20, Kelly M. Makino20, M. Saleem Ismail20, Connie Brand20, Gaby Thai21, Aimee Pierce21, Beatriz Yanez21, Elizabeth Sosa21, Megan Witbracht21, Brendan Kelley22, Trung Nguyen22, Kyle Womack22, Dana Mathews22, Mary Quiceno22, Allan I. Levey23, James J. Lah23, Ihab Hajjar23, Jeffrey M. Burns24, Russell H. Swerdlow24, William M. Brooks24, Daniel H.S. Silverman25, Sarah Kremen25, Liana Apostolova25, Kathleen Tingus25, Po H. Lu25, George Bartzokis25, Ellen Woo25, Edmond Teng25, Neill R Graff-Radford26, Francine Parfitt26, Kim Poki-Walker26, Martin R. Farlow27, Ann Marie Hake27, Brandy R. Matthews27, Jared R. Brosch27, Scott Herring27, Christopher H. van Dyck28, Adam P. Mecca28, Susan P. Good28, Martha G. MacAvoy28, Richard E. Carson28, Pradeep Varma28, Howard Chertkow29, Susan Vaitekunas29, Chris Hosein29, Sandra Black30, Bojana Stefanovic30, Chris (Chinthaka) Heyn30, Ging-Yuek Robin Hsiung31, Ellen Kim31, Benita Mudge31, Vesna Sossi31, Howard Feldman31, Michele Assaly31, Elizabeth Finger32, Stephen Pasternak32, Irina Rachinsky32, Andrew Kertesz32, Dick Drost32, John Rogers32, Ian Grant33, Brittanie Muse33, Emily Rogalski33, Jordan Robson33, M.-Marsel Mesulam33, Diana Kerwin33, Chuang-Kuo Wu33, Nancy Johnson33, Kristine Lipowski33, Sandra Weintraub33, Borna Bonakdarpour33, Nunzio Pomara34, Raymundo Hernando34, Antero Sarrael34, Howard J. Rosen35, Bruce L. Miller35, Micheal W. Weiner35, David Perry35, Raymond Scott Turner36, Kathleen Johnson36, Brigid Reynolds36, Kelly MCCann36, Jessica Poe36, Gad A. Marshall37, Reisa A. Sperling37, Keith A. Johnson37, Jerome Yesavage38, Joy L. Taylor38, Steven Chao38, Jaila Coleman38, Jessica D. White38, Barton Lane38, Allyson Rosen38, Jared Tinklenberg38, Christine M. Belden39, Alireza Atri39, Bryan M. Spann39, Kelly A. Clark39, Edward Zamrini39, Marwan Sabbagh39, Ronald Killiany40, Robert Stern40, Jesse Mez40, Neil Kowall40, Andrew E. Budson40, Thomas O. Obisesan41, Oyonumo E. Ntekim41, Saba Wolday41, Javed I. Khan41, Evaristus Nwulia41, Sheeba Nadarajah41, Alan Lerner42, Paula Ogrocki42, Curtis Tatsuoka42, Parianne Fatica42, Evan Fletcher43, Pauline Maillard43, John Olichney43, Charles DeCarli43, Owen Carmichael43, Vernice Bates44, Horacio Capote44, Michelle Rainka44, Michael Borrie45, T-Y Lee45, Rob Bartha45, Sterling Johnson46, Sanjay Asthana46, Cynthia M. Carlsson46, Allison Perrin47, Anna Burke47, Douglas W. Scharre48, Maria Kataki48, Rawan Tarawneh48, Brendan Kelley48, David Hart49, Earl A. Zimmerman49, Dzintra Celmins49, Delwyn D. Miller50, Laura L. Boles Ponto50, Karen Ekstam Smith50, Hristina Koleva50, Hyungsub Shim50, Ki Won Nam50, Susan K. Schultz50, Jeff D. Williamson51, Suzanne Craft51, Jo Cleveland51, Mia Yang51, Kaycee M. Sink51, Brian R. Ott52, Jonathan Drake52, Geoffrey Tremont52, Lori A. Daiello52, Jonathan D. Drake52, Marwan Sabbagh53, Aaron Ritter53, Charles Bernick53, Donna Munic53, Akiva Mintz53, Abigail O’Connelll54, Jacobo Mintzer54, Arthur Wiliams54, Joseph Masdeu55, Jiong Shi56, Angelica Garcia56, Marwan Sabbagh56, Paul Newhouse57, Steven Potkin58, Stephen Salloway59, Paul Malloy59, Stephen Correia59, Smita Kittur60, Godfrey D. Pearlson61, Karen Blank61, Karen Anderson61, Laura A. Flashman62, Marc Seltzer62, Mary L. Hynes62, Robert B. Santulli62, Norman Relkin63, Gloria Chiang63, Athena Lee63, Michael Lin63, Lisa Ravdin63
1Oregon Health & Science University
2University of Southern California
3University of California – San Diego
4University of Michigan
5Mayo Clinic, Rochester
6Baylor College of Medicine
7 Columbia University Medical Center
8 Washington University, St. Louis
9 University of Alabama—Birmingham
10 Mount Sinai School of Medicine
11 Rush University Medical Center
12 Wien Center
13 Johns Hopkins University
14 University of South Florida: USF Health Byrd Alzheimer’s Institute
15 New York University
16 Duke University Medical Center
17 University of Pennsylvania
18 University of Kentucky
19 University of Pittsburgh
20 University of Rochester Medical Center
21 University of California Irvine IMIND
22 University of Texas Southwestern Medical School
23 Emory University
24 University of Kansas Medical Center
25 University of California, Los Angeles
26 Mayo Clinic, Jacksonville
27 Indiana University
28 Yale University School of Medicine
29 McGill Univ., Montreal-Jewish General Hospital
30 Sunnybrook Health Sciences, Ontario
31 U.B.C. Clinic for AD & Related Disorders
32 St. Joseph’s Health Care
33 Northwestern University
34 Nathan Kline Institute
35 University of California, San Francisco
36 Georgetown University Medical Center
37 Brigham and Women's Hospital
38 Stanford University
39 Banner Sun Health Research Institute
40 Boston University
41 Howard University
42 Case Western Reserve University
43 University of California, Davis – Sacramento
44 Dent Neurologic Institute
45 Parkwood Institute
46 University of Wisconsin
47 Banner Alzheimer's Institute
48 Ohio State University
49 Albany Medical College
50 University of Iowa College of Medicine
51 Wake Forest University Health Sciences
52 Rhode Island Hospital
53 Cleveland Clinic Lou Ruvo Center for Brain Health
54 Roper St. Francis Healthcare
55 Houston Methodist Neurological Institute
56 Barrow Neurological Institute
57 Vanderbilt University Medical Center
58 Long Beach VA Neuropsychiatric Research Program
59 Butler Hospital Memory and Aging Program
60 Neurological Care of CNY
61 Hartford Hospital, Olin Neuropsychiatry Research Center
62 Dartmouth-Hitchcock Medical Center
63 Cornell University