Advertisement

A pilot randomized controlled trial of robotic exoskeleton-assisted exercise rehabilitation in multiple sclerosis

Open AccessPublished:April 04, 2021DOI:https://doi.org/10.1016/j.msard.2021.102936

      Highlights

      • Mobility and cognitive impairment co-occur in persons with multiple sclerosis (PwMS).
      • Robotic-exoskeleton assisted exercise rehabilitation (REAER) may benefit PwMS.
      • This study tested REAER effects on mobility, cognition, and fMRI outcomes in 10 PwMS.
      • REAER improved mobility, cognition, and brain connectivity based on effect sizes.
      • REAER represents a promising rehabilitation approach in PwMS.

      Abstract

      Background

      Co-occurring mobility and cognitive impairments are common, debilitating, and poorly-managed with pharmacological therapies in persons with multiple sclerosis (MS). Exercise rehabilitation (ER), particularly walking ER, has been suggested as one of the best approaches for managing these manifestations of MS. However, there is a focal lack of efficacy of ER on mobility and cognitive outcomes in persons with MS who present with substantial neurological disability. Such severe neurological disability oftentimes precludes the ability for participation in highly-intensive and repetitive ER that is necessary for eliciting adaptations in mobility and cognition. To address such a concern, robotic exoskeleton-assisted ER (REAER) might represent a promising intervention approach for managing co-occurring mobility and cognitive impairments in those with substantial MS disability who might not benefit from traditional ER.

      Methods

      The current pilot single-blind, randomized controlled trial (RCT) compared the effects of 4-weeks of REAER with 4-weeks of conventional gait training (CGT) as a standard-of-care control condition on functional mobility (timed up-and-go; TUG), walking endurance (six-minute walk test; 6MWT), cognitive processing speed (CPS; Symbol Digit Modalities Test; SDMT), and brain connectivity (thalamocortical resting-state functional connectivity (RSFC) based on fMRI) outcomes in 10 persons with substantial MS-related neurological disability.

      Results

      Overall, compared with CGT, 4-weeks of REAER was associated with large improvements in functional mobility (ηp2=.38), CPS (ηp2=.53), and RSFC between the thalamus and ventromedial prefrontal cortex (ηp2=.72), but not walking endurance (ηp2=.01). Further, changes in RSFC were moderately associated with changes in TUG, 6MWT, and SDMT performance, respectively, whereby increased thalamocortical RSFC was associated with improved functional mobility, walking endurance, and CPS (|ρ|>.36).

      Conclusion

      The current pilot RCT provides initial support for REAER as an approach for improving functional mobility and CPS, perhaps based on adaptive and integrative central nervous system plasticity, namely increases in RSFC between the thalamus and ventromedial prefrontal cortex, in a small sample of persons with substantial MS disability. Such a pilot trial provides proof-of-concept data for the design and implementation of an appropriately-powered RCT of REAER in a larger sample of persons with MS who present with co-occurring impairments in both mobility and cognitive functioning.

      Keywords

      1. Introduction

      Multiple sclerosis (MS) is a neurological disease in which the body's immune system attacks its central nervous system (CNS) resulting in axonal damage and myelin loss (
      • Trapp, B.D.
      • Nave, K.-A.
      Multiple sclerosis: an immune or neurodegenerative disorder?.
      ). This CNS disease affects approximately one million people in the United States (
      • Nelson, L.M.
      • Wallin, M.T.
      • Marrie, R.A.
      • et al.
      A new way to estimate neurologic disease prevalence in the United States: illustrated with MS.
      ;
      • Wallin, M.T.
      • Culpepper, W.J.
      • Campbell, J.D.
      • et al.
      The prevalence of MS in the United States: a population-based estimate using health claims data.
      ) and often results in profound mobility and cognitive disability. Mobility disability has been identified as a primary feature of MS (
      • Jock T.
      • Murray, M.
      Multiple Sclerosis: The History of a Disease.
      ) affecting up to 75% of persons with MS (PwMS) (
      • Van Asch, P.
      Impact of mobility impairment in multiple sclerosis 2–patients’ perspectives.
      ). Cognitive impairment is another hallmark consequence of MS that affects 40-65% of patients (
      • Chiaravalloti, N.D.
      • DeLuca, J.
      Cognitive impairment in multiple sclerosis.
      ;
      • Rao, S.M.
      • Leo, G.J.
      • Bernardin, L.
      • Unverzagt, F.
      Cognitive dysfunction in multiple sclerosis.: I. Frequency, patterns, and prediction.
      ;
      • Rocca, M.A.
      • Amato, M.P.
      • De Stefano, N.
      • et al.
      Clinical and imaging assessment of cognitive dysfunction in multiple sclerosis.
      ), is associated with a host of negative health and life outcomes, and like mobility disability, is poorly-managed with pharmacotherapy (
      • DeLuca, J.
      • Chiaravalloti, N.D.
      • Sandroff, B.M.
      Treatment and management of cognitive dysfunction in patients with multiple sclerosis.
      ). Importantly, mobility and cognitive impairments tend to co-occur in PwMS (
      • Benedict, R.H.
      • Holtzer, R.
      • Motl, R.W.
      • et al.
      Upper and lower extremity motor function and cognitive impairment in multiple sclerosis.
      ). The highly burdensome symptoms across two major functional domains underscores the importance of identifying other approaches for addressing co-occurring mobility and cognitive problems that characterize MS, such as exercise rehabilitation (ER).
      Rehabilitation has been identified as perhaps the best method for restoring function in MS (
      • Kraft, G.H.
      Rehabilitation still the only way to improve function in multiple sclerosis.
      ). This is evidenced by a proliferation of research supporting beneficial effects of ER for improving mobility along with a smaller evidence base supporting ER-related cognitive improvements (
      • DeLuca, J.
      • Chiaravalloti, N.D.
      • Sandroff, B.M.
      Treatment and management of cognitive dysfunction in patients with multiple sclerosis.
      ;
      • Motl, R.W.
      • Sandroff, B.M.
      • Kwakkel, G.
      • et al.
      Exercise in patients with multiple sclerosis.
      ). Even a single ER intervention can improve both mobility and cognition in PwMS in general (
      • Briken, S.
      • Gold, S.
      • Patra, S.
      • et al.
      Effects of exercise on fitness and cognition in progressive MS: a randomized, controlled pilot trial.
      ), and those effects might be particularly beneficial if the exercise stimulus involves actual walking (
      • Sandroff, B.M.
      • Bollaert, R.E.
      • Pilutti, L.A.
      • et al.
      Multimodal exercise training in multiple sclerosis: a randomized controlled trial in persons with substantial mobility disability.
      ;
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      ;
      • Sandroff, B.M.
      • Balto, J.M.
      • Klaren, R.E.
      • Sommer, S.K.
      • DeLuca, J.
      • Motl, R.W.
      Systematically developed pilot randomized controlled trial of exercise and cognition in persons with multiple sclerosis.
      ). However, there is a lack of efficacy of ER on mobility and cognitive outcomes in PwMS who present with substantial neurological disability (i.e., PwMS who demonstrate objective deficits in both mobility and cognition) (
      • Motl, R.W.
      • Sandroff, B.M.
      • Kwakkel, G.
      • et al.
      Exercise in patients with multiple sclerosis.
      ). Such severe neurological disability oftentimes precludes participation in highly-intensive and repetitive ER that is necessary for eliciting mobility and cognitive adaptations (
      • Motl, R.W.
      • Sandroff, B.M.
      • Kwakkel, G.
      • et al.
      Exercise in patients with multiple sclerosis.
      ;
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      ). Importantly, persons with substantial MS-related disability likely represent a cohort that has the greatest potential to benefit from ER. Furthermore, adaptive ER approaches such as body-weight supported treadmill training and robot-assisted gait training (i.e., LokoMat) demonstrate overall equivocal evidence for improving mobility and cognition in this population (
      • Lo, A.C.
      • Triche, E.W.
      Improving gait in multiple sclerosis using robot-assisted, body weight supported treadmill training.
      ;
      • Pilutti, L.A.
      • Lelli, D.A.
      • Paulseth, J.E.
      • et al.
      Effects of 12 weeks of supported treadmill training on functional ability and quality of life in progressive multiple sclerosis: a pilot study.
      ;
      • Vaney, C.
      • Gattlen, B.
      • Lugon-Moulin, V.
      • et al.
      Robotic-assisted step training (lokomat) not superior to equal intensity of over-ground rehabilitation in patients with multiple sclerosis.
      ;
      • Gandolfi, M.
      • Geroin, C.
      • Picelli, A.
      • et al.
      Robot-assisted vs. sensory integration training in treating gait and balance dysfunctions in patients with multiple sclerosis: a randomized controlled trial.
      ). For example, one study compared 3-weeks of robot-assisted gait training using the LokoMat (without virtual reality) with 3-weeks of over-ground walking training on quality of life, balance, and mobility outcomes in 49 persons with moderate-to-severe MS disability (
      • Vaney, C.
      • Gattlen, B.
      • Lugon-Moulin, V.
      • et al.
      Robotic-assisted step training (lokomat) not superior to equal intensity of over-ground rehabilitation in patients with multiple sclerosis.
      ). The primary results indicated that robot-assisted gait training did not improve quality of life, balance, or mobility to a greater degree than over-ground walking training in PwMS. Further, a more recent meta-analysis reported that overall, robot-assisted gait training did not exert large, beneficial effects on balance and mobility outcomes compared with walking rehabilitation in PwMS (
      • Xie, X.
      • Sun, H.
      • Zeng, Q.
      • et al.
      Do patients with multiple sclerosis derive more benefit from robot-assisted gait training compared with conventional walking therapy on motor function? A meta-analysis.
      ). Collectively, those adaptive interventions seemingly do not offer additional benefits of actual over-ground walking and lack key interactions between patients and therapists for optimizing the delivery of the intervention. It further is unclear if adaptive ER approaches are superior to conventional gait training (CGT; the current standard-of-care) for improving function in persons with substantial MS disability.
      An alternative approach to ER in persons with substantial MS disability involves the application of robotic exoskeletons (RE). Compared with research on body-weight supported treadmill training and robot-assisted gait training in persons with MS, RE rehabilitation research is in its infancy. Conceptually, RE rehabilitation can help individuals with substantial MS disability actively engage in ER by correcting altered gait patterns during actual over-ground walking (
      • Androwis, G.J.
      • Pilkar, R.
      • Ramanujam, A.
      • Nolan, K.J.
      Electromyography assessment during gait in a robotic exoskeleton for acute stroke.
      ;
      • Ramanujam, A.
      • Cirnigliaro, C.M.
      • Garbarini, E.
      • Asselin, P.
      • Pilkar, R.
      • Forrest, G.F.
      Neuromechanical adaptations during a robotic powered exoskeleton assisted walking session.
      ); this dramatically increases the likelihood of functional adaptations to ER via activity-dependent neuroplasticity (
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      ;
      • Jensen, S.K.
      • Yong, V.W.
      Activity-dependent and experience-driven myelination provide new directions for the management of multiple sclerosis.
      ). RE assistance allows for the application of over-ground, walking-based ER at volumes that are necessary for realizing such functional adaptations in persons with substantial MS-related neurological disability (
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      ). Such a hypothesis is supported by published evidence on the feasibility and initial efficacy of RE use on mobility and balance outcomes in persons with substantial MS disability (
      • Berriozabalgoitia, R.
      • Bidaurrazaga-Letona, I.
      • Otxoa, E.
      • Urquiza, M.
      • Irazusta, J.
      • Rodriguez-Larrad, A.
      Overground robotic program preserves gait in individuals with multiple sclerosis and moderate to severe impairments: a randomized controlled trial.
      ;
      • Drużbicki, M.
      • Guzik, A.
      • Przysada, G.
      • et al.
      Effects of robotic exoskeleton-aided gait training in the strength, body balance, and walking speed in individuals with multiple sclerosis: a single-group preliminary study.
      ;
      • Afzal, T.
      • Tseng, S.-C.
      • Lincoln, J.A.
      • Kern, M.
      • Francisco, G.E.
      • Chang, S.-H.
      Exoskeleton-assisted gait training in persons with multiple sclerosis: a single-group pilot study.
      ;
      • McGibbon, C.
      • Sexton, A.
      • Gryfe, P.
      • et al.
      Effect of using of a lower-extremity exoskeleton on disability of people with multiple sclerosis.
      ). In addition, our previous case-study reported that 4-weeks of RE-assisted ER (REAER) was associated with improved mobility as well as improved cognitive performance relative to CGT in 4 PwMS (
      • Androwis, G.J.
      • Kwasnica, M.A.
      • Niewrzol, P.
      • et al.
      Mobility and cognitive improvements resulted from overground robotic exoskeleton gait-training in persons with MS.
      ). We further hypothesized that the aforementioned functional changes may have been attributable to adaptive neuroplasticity (
      • Pascual-Leone, A.
      • Amedi, A.
      • Fregni, F.
      • Merabet, L.B.
      The plastic human brain cortex.
      ,
      • Straudi, S.
      • Basaglia, N.
      Neuroplasticity-based technologies and interventions for restoring motor functions in multiple sclerosis.
      ), whereby REAER might have changed brain functioning which led to improvements in both mobility and cognitive performance (
      • Sandroff, B.M.
      • Jones, C.D.
      • Baird, J.F.
      • Motl, R.W.
      Systematic review on exercise training as a neuroplasticity-inducing behavior in multiple sclerosis.
      ). As that study (
      • Androwis, G.J.
      • Kwasnica, M.A.
      • Niewrzol, P.
      • et al.
      Mobility and cognitive improvements resulted from overground robotic exoskeleton gait-training in persons with MS.
      ) did not include neuroimaging outcomes, we speculated that such a neural mechanism might involve adaptations in thalamocortical resting-state functional connectivity (RSFC), given previous evidence of associations among increased thalamocortical RSFC, walking-based measures of cardiorespiratory fitness, and cognitive performance with 12-weeks of treadmill walking ER in 10 fully-ambulatory PwMS (
      • Sandroff, B.M.
      • Wylie, G.R.
      • Sutton, B.P.
      • Johnson, C.L.
      • DeLuca, J.
      • Motl, R.W.
      Treadmill walking exercise training and brain function in multiple sclerosis: preliminary evidence setting the stage for a network-based approach to rehabilitation.
      ). Altered RSFC among cortical and subcortical brain regions, including the thalamus, further is associated with mobility and cognitive processing speed deficits in PwMS (
      • Bollaert, R.E.
      • Poe, K.
      • Hubbard, E.A.
      • et al.
      Associations of functional connectivity and walking performance in multiple sclerosis.
      ,
      • Tona, F.
      • Petsas, N.
      • Sbardella, E.
      • et al.
      Multiple sclerosis: altered thalamic resting-state functional connectivity and its effect on cognitive function.
      ). Indeed, RSFC represents an advantageous endpoint for a clinical trial given that this outcome provides a stronger neuroimaging proxy of core neuroplastic changes compared with tasked-based neuroimaging approaches (
      • Kelly, C.
      • Castellanos, F.X.
      Strengthening connections: functional connectivity and brain plasticity.
      ). To address this potential neural mechanism, the current pilot randomized controlled trial (RCT) represents an extension of that previous pilot-study (
      • Androwis, G.J.
      • Kwasnica, M.A.
      • Niewrzol, P.
      • et al.
      Mobility and cognitive improvements resulted from overground robotic exoskeleton gait-training in persons with MS.
      ), whereby we primarily examined the effects of 4-weeks of REAER compared with 4-weeks of CGT on functional mobility (representing an intersection of ambulatory and cognitive functioning), walking endurance, cognitive processing speed (CPS), and thalamocortical RSFC outcomes in 10 persons with substantial MS disability. The primary hypotheses were (a) REAER would be associated with larger improvements in the study outcomes than CGT; and (b) changes in functional mobility, walking endurance, CPS, and thalamocortical RSFC would be inter-correlated.

      2. Methods

      2.1 Experimental design

      The current study represents a two-arm, single-blind RCT on the effects of 4-weeks of REAER compared with 4-weeks of CGT on functional mobility, walking endurance, CPS, and thalamocortical RSFC outcomes in 10 persons with substantial MS disability.

      2.2 Participants

      Participants were recruited directly from Kessler Institute for Rehabilitation, referrals made by patients’ clinician, and a database of previous research participants.
      Participants underwent initial telephone screening to confirm age between 18-75 years, a neurologist-confirmed definitive MS diagnosis, being relapse-free for 30 days, and seizure-free for 90 days.
      Participants subsequently underwent an in-person screening involving administration of the Ambulation Index (AI) (
      • Hauser, S.L.
      • Dawson, D.M.
      • Lehrich, J.R.
      • et al.
      Intensive immunosuppression in progressive multiple sclerosis: a randomized, three-arm study of high-dose intravenous cyclophosphamide, plasma exchange, and ACTH.
      ) for confirmation of substantial MS mobility disability. Persons with AI scores ≥ 2 (i.e., noticeable gait disorder) were included in the study. Participants were included in the study if they had a controlled resting blood pressure (BP) less than 129/89 mmHg (i.e., to minimize the risk of cardiovascular adverse events during training sessions). If a participant had BP higher than 129/89 before a training session on any particular day, BP was measured again after 15 minutes and if it stayed higher than 129/89, training session was rescheduled to a different day. Particpants further were included if they had a passive joint range of motion that (a) fell within normal functional limits for ambulation and (b) was not affected by spasticity at the knee, ankle, or hip joints based on physical therapist (PT)-administered Modified Ashworth Scale scores below 3. Finally, all participants had to fit into the RE based on the manufacturer's guidelines (i.e., height between 1.50 and 1.72 m and weight under 100 kg).
      Participant flow through the study is presented in Fig. 1. Fourteen participants were screened for enrollment, with 12 participants satisfying the aforementioned inclusion/exclusion criteria and randomized to the conditions. Two participants dropped out of the study, resulting in the final analyzed sample of 10 PwMS.

      2.3 Outcome measures

      2.3.1 Functional mobility

      The Timed-up-and-go (TUG) was administered as a measure of functional mobility that is valid and reliable in PwMS (
      • Nilsagard, Y.
      • Lundholm, C.
      • Gunnarsson, L.G.
      • Denison, E.
      Clinical relevance using timed walk tests and ‘timed up and go’testing in persons with multiple sclerosis.
      ,
      • Sebastião, E.
      • Sandroff, B.M.
      • Learmonth, Y.C.
      • Motl, R.W.
      Validity of the timed up and go test as a measure of functional mobility in persons with multiple sclerosis.
      ). Standardized TUG procedures are described elsewhere (
      • Sebastião, E.
      • Sandroff, B.M.
      • Learmonth, Y.C.
      • Motl, R.W.
      Validity of the timed up and go test as a measure of functional mobility in persons with multiple sclerosis.
      ). The primary outcome is the mean of the two trials in seconds. The TUG represents a particularly advantageous study outcome as it strongly encompasses aspects of both mobility and CPS in PwMS (
      • Sebastião, E.
      • Sandroff, B.M.
      • Learmonth, Y.C.
      • Motl, R.W.
      Validity of the timed up and go test as a measure of functional mobility in persons with multiple sclerosis.
      ).

      2.3.2 Walking endurance

      The six-minute walk test (6MWT) was administered as a measure of walking endurance; scores on the 6MWT are valid and reliable in PwMS (
      • Goldman, M.D.
      • Marrie, R.A.
      • Cohen, J.A.
      Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls.
      ). The 6MW took place within a single corridor where participants performed 180° turns, and involved standardized procedures for PwMS (
      • Goldman, M.D.
      • Marrie, R.A.
      • Cohen, J.A.
      Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls.
      ). The primary 6MWT outcome is the total distance traveled (m).

      2.3.3 CPS

      The oral Symbol Digit Modalities Test (SDMT) (
      • Smith, A.
      Symbol Digit Modalities Test: Manual.
      )was administered as a measure of CPS according to standardized procedures (
      • Smith, A.
      Symbol Digit Modalities Test: Manual.
      ). The total number of correct responses in 90 seconds is the primary SDMT outcome. Of note, the SDMT is sensitive to the presence of MS-related cognitive impairment across numerous studies (
      • Benedict, R.H.
      • DeLuca, J.
      • Phillips, G.
      • et al.
      Validity of the symbol digit modalities test as a cognition performance outcome measure for multiple sclerosis.
      ). We further provided alternate forms across testing time points.

      2.3.4 Thalamocortical resting-state functional connectivity

      Participants underwent neuroimaging, which included structural imaging as well as a resting-state scan using an FDA-approved Siemens Skyra 3T clinical imager. Details on MRI sequences and pre-processing are provided in the Supplementary Material. The primary RSFC data were analyzed using a seed-based approach. The 4mm seed was placed in the bilateral thalami (left side coordinates: X=−10, Y=−17, Z=6; right side coordinates: X=2, Y=−22, Z=−2). For each seed, the mean percent signal change was calculated within the seed for each volume of the error term time-series (i.e., the residuals from the deconvolution). The correlation between this time-series of means and every voxel in the brain were then computed, before being converted into z-scores with Fisher's R-to-Z transformation. The resulting z-score maps were entered into the group-level analysis (see Data Analyses below).

      2.4 Experimental condition

      The experimental condition (REAER) involved 4-weeks of supervised and progressive walking training over-ground in the RE (Ekso-GT, Ekso Bionics, Berkley, CA, USA). The REAER intervention itself took place 2 times/week over 4-weeks on a level surface, and led by a licensed PT. The supervised REAER sessions were individualized, but progressed based on training duration and reliance on robotic assistance. The prescribed reduction in robotic assistance over the course of the intervention served as a general guideline as opposed to an absolute requirement. The exercise itself initially consisted of approximately 30-minutes of over-ground walking training using a suggested maximum allowable level of 100% robotic assistance/session (week 1) and eventually progressed to approximately 45-minutes of walking training using a suggested maximum allowable level of 40% robotic assistance/session by the conclusion of the training program (week 4). To optimize participant comfortability and RE functionality, each session began with a 5-minute accommodation period for donning the RE with close guidance from the treating PT and then a 5-minute warm-up. This period was followed by the exercise itself, and then a 5-10 min period for a cool-down and doffing the RE.
      During the exercise session, as participants walked over-ground in the RE, they were guided by the treating PT to follow the kinematic trajectories of the RE for better activating their muscles and moving their limbs in the same profile as the RE. Participants walked in the RE using an initially high level of robotic assistance, which gradually was decreased over the course of each session. The reduction of robotic assistance within a given session was based on the feedback parameters (i.e., percent assistance at each of the powered joints, number of steps taken, distance walked) provided on the controlling monitor of the RE as well as the participant's reported level of fatigue; this was monitored in real-time by the treating PT who manually reduced the percent robotic assistance as appropriate. As the sessions were individualized, participants were encouraged to walk in the RE using less than the maximum allowable level of robotic assistance within each session to increase difficulty (i.e., suggested maximum level of robotic assistance was 100%;90%;80%;70%;70%;60%;50%; and 40% in sessions 1-8, respectively). Within a given session, in the event where a participant reported high levels of fatigue, the level of robotic assistance was temporarily increased and a short break (~3 minutes) was provided. Of note, during the first week of training (sessions 1-2), all participants were provided with bilateral assistive devices (e.g., bilateral canes, or walker) to use while walking in the RE; this feature of the intervention was included to enhance participant comfort and to prepare participants for more difficult REAER as the intervention progressed (i.e., longer duration, less robotic assistance). We further note that the Ekso GT RE is not a self-balancing RE and balance during ambulation is almost always guided by the treating PT along with assistive devices, consistent with the instruction protocol provided by the manufacturer. The notion of reducing the reliance on assistive device during REAER was based on taxing participants’ balance, weight transfer, and lower-extremity loading, presumably increasing the potential for realizing adaptations in mobility. Accordingly, beginning in the second week of the intervention and continuing through the remainder of the intervention, participants were guided and continuously supported by the treating PT to engage in REAER without using the provided assistive device if possible. For each session, percent assistance from the RE at each of the powered joints, number of steps taken, and distance walked/session were recorded. Adherence was operationalized as attendance and compliance was operationalized as the ability to complete each session as prescribed based on session duration and maximum percent robotic assistance.

      2.5 Control condition

      The control comparison condition involved 4-weeks of supervised CGT that accounted for over-ground walking, attention, and social contact; CGT represents the current standard-of-care for rehabilitating gait in MS (
      • Sutliff, M.H.
      • Bennett, S.E.
      • Bobryk, P.
      • et al.
      Rehabilitation in multiple sclerosis: commentary on the recent AAN systematic review.
      ). The CGT sessions were led by a licensed PT, and focused on mobility, gait, balance and lower extremity function. These sessions occurred at the same frequency and duration (i.e., at least 30 minutes walking over-ground per session) as the REAER condition. As was the case for REAER, CGT sessions were individually tailored to participants’ neurological disability by the treating PT whereby participants were encouraged to complete each task to the best of their ability. Sessions included training on elements of stretching, strengthening, ambulation training, balance training, weight support, transfer training, stepping length and width and weight shift during ambulation (see Fig. 2) (
      • Aisen, M.L.
      • Sevilla, D.
      • Fox, N.
      Inpatient rehabilitation for multiple sclerosis.
      ,
      • Haselkorn, J.K.
      • Hughes, C.
      • Rae-Grant, A.
      • et al.
      Summary of comprehensive systematic review: rehabilitation in multiple sclerosis: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology.
      ,
      • Sattelmayer, M.
      • Chevalley, O.
      • Steuri, R.
      • Hilfiker, R.
      Over-ground walking or robot-assisted gait training in people with multiple sclerosis: does the effect depend on baseline walking speed and disease related disabilities? A systematic review and meta-regression.
      ,
      • Sutliff, M.H.
      • Bennett, S.E.
      • Bobryk, P.
      • et al.
      Rehabilitation in multiple sclerosis: commentary on the recent AAN systematic review.
      ). Throughout the 4-weeks, CGT progressed in terms of duration (matching that of the REAER condition) and difficulty of training elements (e.g., narrowing base of support, performing activities with eyes closed, etc.). Within each session, the licensed PT monitored participants’ form, provided support when needed, and ensured that the assigned exercises were completed as prescribed. As was the case for REAER, participants reported their levels of fatigue in real-time, and the PT made accommodations to the CGT activities accordingly (i.e., either decreasing difficulty or providing a short break if necessary). For CGT, adherence was operationalized as attendance and compliance was operationalized as the ability to complete each session as prescribed based on session duration and number of exercises performed.
      Fig. 2
      Fig. 2(A) Participant with MS engaging in REAER (Ekso-GT, Ekso Bionics, Inc.); (B) Participant with MS engaging in CGT.

      2.6 Procedure

      The study was approved by the KF Institutional Review Board and all participants provided written informed consent. Upon satisfaction of the inclusion/exclusion criteria, all participants underwent baseline testing. Baseline testing took place across two sessions at KF, separated by 48 hours, to minimize fatigue. The first baseline session involved TUG, 6MWT, and SDMT administration, and the second session involved the MRI scan. All baseline testing was administered by treatment-blinded assessors. Participants were subsequently randomly assigned to the experimental or control conditions using a computerized random numbers generator by a researcher who was uninvolved in the administration of outcome assessments. After the 4-week study period, participants completed follow-up outcome assessments at KF, using alternate forms where applicable, which were administered by the same treatment-blinded assessors as baseline.

      2.7 Data analysis

      Study data were analyzed using SPSS version 26 (IBM Inc., Armonk, NY); RSFC data further were analyzed using R (version 3.5.3). Given the pilot nature of this RCT, a per-protocol (i.e., completer's) analytic approach was adopted and accordingly included 10 PwMS who completed all study procedures. We first analyzed possible baseline differences in demographic and clinical characteristics, as well as in the study endpoints using independent-samples t-tests and chi-square difference tests. If a given outcome differed between the groups at baseline, it was included in subsequent models as a covariate. The primary analytic models involved separate, mixed-factor analyses of variance (ANOVA) on TUG, 6MWT, and SDMT. Time (i.e., baseline vs. follow-up) was included as a within-subjects factor and group (i.e., REAER vs. CGT) was included as a between-subjects factor. Effect sizes from the ANOVAs were expressed as partial eta-squared (ηp2), with values of .01, .06, and .14 interpreted as small, moderate, and large, respectively (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ). We further computed effect sizes for change scores (i.e., follow-up minus baseline) between the conditions, expressed as Cohen's d (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ) with values interpreted as small, moderate, and large based on criteria of 0.2, 0.5, and 0.8, respectively. For the thalamocortical RSFC outcomes, a Linear Mixed Effects model was used. The factors were ‘time’ (as above) and ‘group’ (as above); ‘side’ (left vs. right) was also included, to account for the bilateral seeds; and ‘subject’ was included as a random factor. The results of this whole-brain analysis were corrected for multiple comparisons by using an individual voxel probability threshold of p<.01 and a cluster threshold of 65 voxels (voxel dimension=2.3 × 2.3 × 4.0 mm). Monte Carlo simulations, using 3dClustSim (version AFNI_17.2.16, compile date: 9/19/2017) showed this combination to result in a corrected alpha level of p<.05. As we were interested in REAER as a possible neuroplasticity-inducing behavior, we performed bivariate Spearman's rho rank-order correlations (ρ) among changes in TUG, 6MWT, SDMT, and thalamocortical RSFC (
      • Rousselet, G.A.
      • Pernet, C.R.
      Improving standards in brain-behavior correlation analyses.
      ), with effect sizes of .1, .3., and .5 interpreted as small, moderate, and large, respectively (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ). Of note, given the small sample size of the present early-stage trial, the aforementioned analytic models were not likely to reach statistical significance; therefore, we underscore the importance of interpreting effect size estimates for all primary and secondary analyses. This is particularly important for early-stage experimental research for reducing Type II error and providing evidence against prematurely dismissing an intervention that might be beneficial (
      • Mohr, D.C.
      • Spring, B.
      • Freedland, K.E.
      • et al.
      The selection and design of control conditions for randomized controlled trials of psychological interventions.
      ).

      3. Results

      3.1 Descriptive characteristics

      Demographic and clinical characteristics of the sample are presented in Table 1. The sample consisted of highly-educated persons with relapsing-remitting MS who demonstrated functional mobility, walking endurance, and CPS that is consistent with other samples of persons with substantial MS disability (
      • Sandroff, B.M.
      • Bollaert, R.E.
      • Pilutti, L.A.
      • et al.
      Multimodal exercise training in multiple sclerosis: a randomized controlled trial in persons with substantial mobility disability.
      ). There were no significant group differences in most baseline characteristics. However, the CGT group demonstrated significantly worse baseline walking endurance relative to the REAER group (p=.03). Given the baseline difference in 6MWT performance, as a conservative approach, we performed ANCOVA on follow-up 6MWT performance, controlling for baseline scores.
      Table 1Baseline demographic and clinical characteristics of 10 persons with substantial MS disability.
      VariableOverall Sample (N=10)REAER (N=6)CGT (N=4)p-value
      Age (years)49.9 (8.4)46.5 (5.2)55.0 (9.6).14
      Sex (% female)6/10 (60%)3/6 (50%)3/4 (75%).49
      Education (years)16.9 (1.4)16.7 (1.5)17.3 (1.2).58
      MS Type: RRMS (n, %) Progressive (n, %)9/10 (90.0%) 1/10 (10.0%)5/6 (83.3%) 1/6 (16.7%)4/4 (100.0%) 0/4 (0.0%).39
      AI (median, range)4.0 (2 - 5)3.5 (2 - 5)5.0 (4 - 5).06
      Note: All data presented as mean (SD), unless otherwise noted; p-value based on independent samples t-tests or chi-square difference tests; * denotes statistical significance at p < .05; REAER = robotic exoskeleton-assisted exercise rehabilitation; CGT = conventional gait training; RRMS = relapsing-remitting multiple sclerosis; AI = Ambulation Index.

      3.2 Adherence and compliance

      All ten participants who were retained in the study attended all 8 REAER or CGT sessions over the 4-week period. Regarding compliance, all participants completed each session as prescribed across both conditions. As a manipulation check, participants in the REAER group took approximately 59% more steps during session 8 compared with session 1.

      3.3 Primary outcomes

      Data on TUG performance, 6MWT performance, SDMT performance, and thalamocortical RSFC per group per time point are presented in Table 2 and Fig. 3.
      Table 2Functional Mobility, Walking Endurance, Cognitive Processing Speed, and Thalamocortical RSFC Outcomes in 10 persons with Multiple Sclerosis.
      REAER (n=6)CGT (n=4)
      VariableBaselineFollow-upBaselineFollow-up
      TUG (s)19.3 (10.9)14.3 (7.1)22.4 (4.9)22.8 (3.2)
      6MWT (m)291.0 (103.7)322.7 (129.4)147.5 (26.7)145.7 (27.7)
      SDMT (raw score)41.8 (11.8)51.0 (7.0)54.0 (20.5)49.5 (18.5)
      Thalamocortical RSFC (z-score)−0.028 (0.014)0.021 (0.049)−0.001 (0.075)−0.125 (0.113)
      Note: All data presented as mean (SD); REAER = robotic exoskeleton assisted exercise rehabilitation; CGT = conventional gait therapy; TUG = timed up-and-go; 6MWT = six-minute walk test; SDMT = Symbol Digit Modalities Test; RSFC = resting-state functional connectivity.
      Fig. 3
      Fig. 3Changes in (A) TUG performance; (B) 6MWT performance; (C) SDMT performance; and (D) thalamocortical RSFC in 10 persons with MS who underwent 4-weeks of REAER or CGT.
      Note: Values represent estimated marginal means per time point along with standard error bars and 95% confidence intervals.
      TUG. There was a non-significant time × group interaction on TUG performance(F(1,8)=4.92,p=.06,ηp2=.38; Fig. 3a). Although non-significant, the interaction was associated with a large effect (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ), whereby the REAER group demonstrated substantial improvements in TUG performance relative to minimal change in TUG performance for the CGT group(d=−1.20). There further were non-significant main effects of time and group on TUG performance.
      6MWT. There was no significant difference in follow-up 6MWT performance after controlling for baseline scores. Although the REAER group demonstrated improved 6MWT distance compared with minimal change for the CGT group(d=0.80; Fig. 3b), this effect was likely due to differences in baseline 6MWT performance.
      SDMT. There was a statistically significant time × group interaction on SDMT scores(F(1,8)=9.15,p=.02,ηp2=.53; Fig. 3c); this interaction was large in magnitude based on effect size estimates (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ). Persons who underwent REAER demonstrated substantial improvements in SDMT performance (i.e., 9.2 points) relative to substantial worsening for those who underwent CGT (i.e., 4.5 points)(d=1.42). There further were no other significant effects.
      Thalamocortical RSFC. There was a statistically significant time × group interaction on RSFC between the thalamus and ventromedial prefrontal cortex(F(1,8)=20.15,p<.01,ηp2=.72; Fig. 3d); this interaction was large in magnitude based on effect size estimates (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ). Persons who underwent REAER demonstrated increases in thalamocortical RSFC compared with substantial decreases in RSFC for those who underwent CGT(d=1.64) (see Fig. 4). There were no other significant effects.
      Fig. 4
      Fig. 4Increases in resting-state functional connectivity between the thalamus and ventromedial prefrontal cortex for persons with MS who underwent REAER relative to those who underwent CGT thalamocortical RSFC.

      3.4 Correlations

      Spearman's rho (ρ) rank-order correlations among changes in TUG performance, 6MWT performance, SDMT performance, and thalamocortical RSFC are presented in Table 3. Although non-statistically significant, changes in RSFC were moderately associated with changes in TUG, 6MWT, and SDMT performance, respectively, based on effect size estimates (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ), whereby increased thalamocortical RSFC was associated with improved functional mobility, walking endurance, and CPS. These were not statistically significant, likely due to the small sample size. Interestingly, improvement in TUG performance was significantly associated with improvement in SDMT performance whereby reduced time to complete the TUG was associated with increases in SDMT scores from baseline to follow-up; the magnitude of this association was large based on effect size estimates (
      • Cohen, J.
      Statistical Power Analysis for the Behavioral Sciences.
      ).
      Table 3Spearman's rho (ρ) rank-order correlations among changes in TUG, 6MWT, SDMT, and thalamocortical RSFC in 10 persons with multiple sclerosis.
      VariableΔ TUGΔ 6MWTΔ SDMTΔ Thalamocortical RSFC
      Δ TUG
      Δ 6MWT−.21
      Δ SDMT−.75*.18
      Δ Thalamocortical RSFC−.39.36.38
      Note: * denotes p < .05 based on two-tailed test; Δ = change (follow-up minus baseline); TUG = Timed up-and-go; 6MWT = Six-minute walk test; SDMT = Symbol Digit Modalities Test; RSFC = Resting-state functional connectivity

      4. Discussion

      To our knowledge, the current pilot study represents the first RCT of REAER effects on the intersection of mobility, cognition, and brain connectivity compared with an active, standard-of-care control condition among a small sample of persons with substantial MS disability. The primary novel results indicated that relative to CGT, 4-weeks of REAER was associated with large improvements in functional mobility, CPS, and RSFC between the thalamus and ventromedial prefrontal cortex in 10 PwMS based on effect size estimates. This is consistent with other preliminary work reporting on REAER-related benefits on aspects of mobility in persons with substantial MS disability (
      • Afzal, T.
      • Tseng, S.-C.
      • Lincoln, J.A.
      • Kern, M.
      • Francisco, G.E.
      • Chang, S.-H.
      Exoskeleton-assisted gait training in persons with multiple sclerosis: a single-group pilot study.
      ,
      • Berriozabalgoitia, R.
      • Bidaurrazaga-Letona, I.
      • Otxoa, E.
      • Urquiza, M.
      • Irazusta, J.
      • Rodriguez-Larrad, A.
      Overground robotic program preserves gait in individuals with multiple sclerosis and moderate to severe impairments: a randomized controlled trial.
      ,
      • Drużbicki, M.
      • Guzik, A.
      • Przysada, G.
      • et al.
      Effects of robotic exoskeleton-aided gait training in the strength, body balance, and walking speed in individuals with multiple sclerosis: a single-group preliminary study.
      ,
      • Kozlowski, A.J.
      • Fabian, M.
      • Lad, D.
      • Delgado, A.D.
      Feasibility and safety of a powered exoskeleton for assisted walking for persons with multiple sclerosis: a single-group preliminary study.
      ,
      • McGibbon, C.
      • Sexton, A.
      • Gryfe, P.
      • et al.
      Effect of using of a lower-extremity exoskeleton on disability of people with multiple sclerosis.
      ). Consistent with our a priori hypotheses, those improvements in thalamocortical RSFC were moderately associated with improvements in functional mobility and CPS based on effect size estimates. Such a pattern of results provides preliminary evidence supporting REAER as a possible neuroplasticity-inducing behavior in persons with substantial MS mobility disability, whereby REAER-related improvements brain connectivity might explain downstream adaptations in motor and cognitive function (
      • Sandroff, B.M.
      • Jones, C.D.
      • Baird, J.F.
      • Motl, R.W.
      Systematic review on exercise training as a neuroplasticity-inducing behavior in multiple sclerosis.
      ). This is particularly exciting considering that REAER might represent an innovative approach for rehabilitating the highly prevalent and burdensome consequences of co-occurring mobility and cognitive disability in MS that may even be superior to the current standard-of-care.
      We have previously hypothesized that exercise might improve mobility and cognition in PwMS based on integrative CNS plasticity (i.e., PRIMERS hypothesis (
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      )). The current RCT of REAER effects on neuroimaging and functional outcomes represents an initial test of that hypothesis, as REAER is an exercise behavior that clearly involves cognitive-motor interactions. From a neurophysiological perspective, over-ground, RE-assisted walking is a particularly complex behavior that requires the rapid processing and integration of multisensory input. The RE provides motor assistance to participants by driving their angular joints of the lower extremity through a repetitive, predefined trajectory to complete the gait pattern. By extension, successfully walking using RE assistance requires processing of proprioceptive, locomotor, postural, visuospatial, and cardiorespiratory sensory information (i.e., stimulation of the cardiorespiratory system via movement (
      • Evans, N.
      • Hartigan, C.
      • Kandilakis, C.
      • Pharo, E.
      • Clesson, I.
      Acute cardiorespiratory and metabolic responses during exoskeleton-assisted walking overground among persons with chronic spinal cord injury.
      )), among other sources, based on research reporting that acute RE use taxes those physiological systems (
      • Contreras-Vidal, J.L.
      • Kilicarslan, A.
      • Huang, H.H.
      • Grossman, R.G.
      Human-centered design of wearable neuroprostheses and exoskeletons.
      ,
      • Eijlers, A.J.
      • Dekker, I.
      • Steenwijk, M.D.
      • et al.
      Cortical atrophy accelerates as cognitive decline worsens in multiple sclerosis.
      ,
      • Gagnon, D.H.
      • Escalona, M.J.
      • Vermette, M.
      • et al.
      Locomotor training using an overground robotic exoskeleton in long-term manual wheelchair users with a chronic spinal cord injury living in the community: lessons learned from a feasibility study in terms of recruitment, attendance, learnability, performance and safety.
      ). That information must be rapidly integrated for preparing an appropriate and active motor response (i.e., consciously taking an additional step according to the RE's controlled, predefined trajectory). As the RE serves to correct a participant's pathological gait, a single bout of RE-assisted walking over-ground also requires a substantial degree of motor learning and cognitive processing, as walking using RE-assistance biomechanically deviates from a person's habitual walking pattern. In the current study, over the course of the REAER intervention, the maximum allowable degree of robotic assistance was progressively reduced across sessions. Such a progressive reduction rendered the actual walking process more difficult, whereby ambulation in the RE was completed largely by the participant's own effort (and not the RE) by the end of the intervention. Such a progression over time presumably requires adaptations in communication at the brain-systems level, in order for the participant to keep up with the progressively increasing demands of REAER. Consistent with the PRIMERS hypothesis, those adaptations at the brain-systems level should transfer to functions that rely on such neural circuitry (
      • Sandroff, B.M.
      • Motl, R.W.
      • Reed, W.R.
      • Barbey, A.K.
      • Benedict, R.H.
      • DeLuca, J.
      Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
      ).
      Collectively, the observed pattern of results provides initial support for the PRIMERS hypothesis, given large intervention effects, based on effect size estimates, favoring REAER over the active control condition (i.e., CGT) that qualitatively requires fewer neurophysiological demands to complete. Relatedly, the present results provide additional support for the PRIMERS hypothesis given moderate-sized correlations between changes in thalamocortical RSFC and functional outcomes. As thalamic neuroimaging outcomes have been linked with both mobility and cognition in PwMS (e.g., (
      • Tona, F.
      • Petsas, N.
      • Sbardella, E.
      • et al.
      Multiple sclerosis: altered thalamic resting-state functional connectivity and its effect on cognitive function.
      ;
      • Eijlers, A.J.
      • Dekker, I.
      • Steenwijk, M.D.
      • et al.
      Cortical atrophy accelerates as cognitive decline worsens in multiple sclerosis.
      ;
      • Motl, R.W.
      • Zivadinov, R.
      • Bergsland, N.
      • Benedict, R.H.
      Thalamus volume and ambulation in multiple sclerosis: a cross-sectional study.
      )), it is plausible that perhaps the observed REAER-related changes in thalamocortical RSFC drove changes in TUG and SDMT performance, respectively. This is consistent with research in persons with stroke that reports on large, REAER-related improvements in cognitive performance (
      • Manuli, A.
      • Maggio, M.G.
      • Latella, D.
      • et al.
      Can robotic gait rehabilitation plus Virtual Reality affect cognitive and behavioural outcomes in patients with chronic stroke? A randomized controlled trial involving three different protocols.
      ) as well as strengthened cortical connectivity (i.e., EEG-based frontoparietal effective connectivity) using the Ekso-GT RE (
      • Calabrò, R.S.
      • Naro, A.
      • Russo, M.
      • et al.
      Shaping neuroplasticity by using powered exoskeletons in patients with stroke: a randomized clinical trial.
      ). Such convergent evidence in neurological populations provides initial support for REAER as a potential neuroplasticity-inducing behavior. Although non-statistically significant (likely due to the small sample size), the large intervention effect on TUG performance is particularly noteworthy, considering that the TUG is a measure that captures the intersection of mobility and cognition in MS (
      • Sebastião, E.
      • Sandroff, B.M.
      • Learmonth, Y.C.
      • Motl, R.W.
      Validity of the timed up and go test as a measure of functional mobility in persons with multiple sclerosis.
      ). To that end, perhaps REAER may represent a novel exercise intervention for simultaneously improving interrelated mobility and cognitive functions in PwMS (
      • Motl, R.W.
      • Sandroff, B.M.
      • DeLuca, J.
      Exercise training and cognitive rehabilitation: a symbiotic approach for rehabilitating walking and cognitive functions in multiple sclerosis?.
      ). Indeed, this is supported by a large, statistically significant correlation (ρ=−.75) between change in functional mobility and change in CPS.
      Another noteworthy result of the present study includes the large, statistically significant time-by-group interaction on CPS, whereby those who underwent REAER demonstrated a 9.2-point improvement in SDMT scores, compared with a 4.5-point decrease for those who underwent CGT. Taken together, such a pattern of results suggests that REAER was associated with a clinically-meaningful improvement in CPS (
      • Morrow, S.A.
      • Drake, A.
      • Zivadinov, R.
      • Munschauer, F.
      • Weinstock-Guttman, B.
      • Benedict, R.H.
      Predicting loss of employment over three years in multiple sclerosis: clinically meaningful cognitive decline.
      ) in just 4-weeks. This adds to the growing body of evidence supporting ER for improving cognition in PwMS (
      • Sandroff, B.M.
      • Motl, R.W.
      • Scudder, M.R.
      • DeLuca, J.
      Systematic, evidence-based review of exercise, physical activity, and physical fitness effects on cognition in persons with multiple sclerosis.
      ), as the observed improvements in SDMT score might partly be attributable to central mechanisms (i.e., enhanced thalamocortical RSFC). However, recommendations for REAER as a possible treatment for MS-related cognitive impairment are premature, as we did not exclusively recruit persons with objective CPS impairments a priori (
      • DeLuca, J.
      • Chiaravalloti, N.D.
      • Sandroff, B.M.
      Treatment and management of cognitive dysfunction in patients with multiple sclerosis.
      ), given the pilot nature of the present RCT. Of note, the sample did demonstrate some degree of CPS impairment on average based on an age-, sex-, and education-adjusted z-score of −1.0 SD units (i.e., 16th percentile). Improvement on the SDMT was not likely attributable to learning effects (i.e., practice effects) over a short testing interval considering the application of alternate forms across testing time points coupled with a clinically-meaningful worsening of CPS for those who underwent the control condition. We further note that the present pilot RCT was not designed as a non-inferiority trial; but rather, the CGT condition was selected as an active control condition that accounted for the effects of social contact, attention, and aspects of gait training. Although CGT happens to represent the standard-of-care for mobility rehabilitation in MS, we did not expect the CPS outcomes to improve (
      • Barbarulo A.M.
      • Giacomo L.
      • Elisabetta S.
      • Luigi T.
      • Dario G.
      • Mariateresa E.
      • Teresa C.
      • et al.
      Integrated cognitive and neuromotor rehabilitation in multiple sclerosis: a pragmatic study.
      ) compared with the REAER condition. Nevertheless, the current results support the design and implementation of future larger RCTs examining the effects of REAER for managing CPS impairment in PwMS.
      There are several strengths of the current study, such as the single-blind RCT design, inclusion of neuroimaging outcomes and an active control condition representing the standard-of-care for MS mobility rehabilitation, along with high adherence/compliance rates. There are also several limitations. The small sample was likely underpowered for detecting robust changes in mobility and cognition; this further limits the generalizability of the results amongst the MS population at-large. As such, we applied a per-protocol analytic approach (as opposed to intent-to-treat principles) and report on effect sizes in addition to p-values (
      • Cutter, G.
      Effect size–one size doesn’t fit all.
      ), given that this preliminary trial can promote innovation and provide the effect sizes necessary for informing a larger, appropriately-powered trial of the efficacy of REAER in PwMS who demonstrate both mobility and cognitive impairment. Further, although the demographic and clinical characteristics of the current sample are consistent with other samples of persons with substantial MS-related disability, the small sample introduces a possibility of a selection bias, whereby the results may not be fully generalizable to the MS population at large. Finally, although the sample demonstrated some degree of baseline CPS impairment on average, not all participants had SDMT scores that were ≥ 1.0 SDs below the age-, sex-, and education-adjusted normative mean. Nonetheless, the present pilot RCT provides proof-of-concept data supporting REAER for improving functional mobility and CPS in PwMS, perhaps based on adaptive neuroplasticity. Subsequent exercise trials using REAER should pre-screen participants for impairments in both domains of functioning (
      • Motl, R.W.
      • Sandroff B.M.
      Current perspectives on exercise training in the management of multiple sclerosis.
      ).

      5. Conclusions

      The current pilot RCT provides preliminary support for REAER as an approach for improving functional mobility and CPS, largely based on effect size estimates, in a small sample of persons with substantial MS disability. Such improvements may be based on adaptive and integrative CNS plasticity, namely increases in RSFC between the thalamus and ventromedial prefrontal cortex. Collectively, this pilot trial provides proof-of-concept data for the design and implementation of an appropriately-powered RCT of REAER in a larger sample of PwMS who present with co-occurring impairments in both mobility and cognitive functioning.

      Role of funding source

      The authors acknowledge the support from funding agencies in particular the National Multiple Sclerosis Society, USA (Collaborative Network of New Jersey), Award Number: CA1069-A-7; and the Avidan Foundation- New Jersey, USA.

      Credit author statement

      GJA, BMS, GY, and JD: contributed to the conceptualization and design of the study; GJA, PN and FF: contributed to data curation; GJA, PN and GRW: contributed to processing collected data; GJA and BMS: performed the formal and statistical analysis; GJA and JD contributed to funding acquisition and GJA and BMS contributed to writing original draft. All authors contributed to the interpretation of the data, manuscript writing and revision.

      Declaration of Competing Interest

      None.

      Acknowledgment

      The authors acknowledge the support from funding agencies in particular the National Multiple Sclerosis Society, USA (Collaborative Network of New Jersey), Award Number: CA1069-A-7; and the Avidan Foundation- New Jersey, USA.

      Appendix. Supplementary materials

      References

        • Trapp, B.D.
        • Nave, K.-A.
        Multiple sclerosis: an immune or neurodegenerative disorder?.
        Annu. Rev. Neurosci. 2008; 31: 247-269
        • Nelson, L.M.
        • Wallin, M.T.
        • Marrie, R.A.
        • et al.
        A new way to estimate neurologic disease prevalence in the United States: illustrated with MS.
        Neurology. 2019; 92: 469-480
        • Wallin, M.T.
        • Culpepper, W.J.
        • Campbell, J.D.
        • et al.
        The prevalence of MS in the United States: a population-based estimate using health claims data.
        Neurology. 2019; 92: e1029-e1040
        • Jock T.
        • Murray, M.
        Multiple Sclerosis: The History of a Disease.
        Demos medical publishing, 2004
        • Van Asch, P.
        Impact of mobility impairment in multiple sclerosis 2–patients’ perspectives.
        Eur. Neurol. Rev. 2011; 6: 115-120
        • Chiaravalloti, N.D.
        • DeLuca, J.
        Cognitive impairment in multiple sclerosis.
        Lancet Neurol. 2008; 7: 1139-1151
        • Rao, S.M.
        • Leo, G.J.
        • Bernardin, L.
        • Unverzagt, F.
        Cognitive dysfunction in multiple sclerosis.: I. Frequency, patterns, and prediction.
        Neurology. 1991; 41: 685-691
        • Rocca, M.A.
        • Amato, M.P.
        • De Stefano, N.
        • et al.
        Clinical and imaging assessment of cognitive dysfunction in multiple sclerosis.
        Lancet Neurol. 2015; 14: 302-317
        • DeLuca, J.
        • Chiaravalloti, N.D.
        • Sandroff, B.M.
        Treatment and management of cognitive dysfunction in patients with multiple sclerosis.
        Nat. Rev. Neurol. 2020; : 1-14
        • Benedict, R.H.
        • Holtzer, R.
        • Motl, R.W.
        • et al.
        Upper and lower extremity motor function and cognitive impairment in multiple sclerosis.
        J. Int. Neuropsychol. Soc. 2011; 17: 643-653
        • Kraft, G.H.
        Rehabilitation still the only way to improve function in multiple sclerosis.
        Lancet North Am. Ed. 1999; 354: 2016-2017
        • Motl, R.W.
        • Sandroff, B.M.
        • Kwakkel, G.
        • et al.
        Exercise in patients with multiple sclerosis.
        Lancet Neurol. 2017; 16: 848-856
        • Briken, S.
        • Gold, S.
        • Patra, S.
        • et al.
        Effects of exercise on fitness and cognition in progressive MS: a randomized, controlled pilot trial.
        Mult. Scler. J. 2014; 20: 382-390
        • Sandroff, B.M.
        • Bollaert, R.E.
        • Pilutti, L.A.
        • et al.
        Multimodal exercise training in multiple sclerosis: a randomized controlled trial in persons with substantial mobility disability.
        Contemp. Clin. Trials. 2017; 61: 39-47
        • Sandroff, B.M.
        • Motl, R.W.
        • Reed, W.R.
        • Barbey, A.K.
        • Benedict, R.H.
        • DeLuca, J.
        Integrative CNS plasticity with exercise in MS: the PRIMERS (PRocessing, Integration of Multisensory Exercise-Related Stimuli) conceptual framework.
        Neurorehabil. Neural Repair. 2018; 32: 847-862
        • Sandroff, B.M.
        • Balto, J.M.
        • Klaren, R.E.
        • Sommer, S.K.
        • DeLuca, J.
        • Motl, R.W.
        Systematically developed pilot randomized controlled trial of exercise and cognition in persons with multiple sclerosis.
        Neurocase. 2016; 22: 443-450
        • Lo, A.C.
        • Triche, E.W.
        Improving gait in multiple sclerosis using robot-assisted, body weight supported treadmill training.
        Neurorehabil. Neural Repair. 2008; 22: 661-671
        • Pilutti, L.A.
        • Lelli, D.A.
        • Paulseth, J.E.
        • et al.
        Effects of 12 weeks of supported treadmill training on functional ability and quality of life in progressive multiple sclerosis: a pilot study.
        Arch. Phys. Med. Rehabil. 2011; 92: 31-36
        • Vaney, C.
        • Gattlen, B.
        • Lugon-Moulin, V.
        • et al.
        Robotic-assisted step training (lokomat) not superior to equal intensity of over-ground rehabilitation in patients with multiple sclerosis.
        Neurorehabil. Neural Repair. 2012; 26: 212-221
        • Gandolfi, M.
        • Geroin, C.
        • Picelli, A.
        • et al.
        Robot-assisted vs. sensory integration training in treating gait and balance dysfunctions in patients with multiple sclerosis: a randomized controlled trial.
        Front. Hum. Neurosci. 2014; 8: 318
        • Xie, X.
        • Sun, H.
        • Zeng, Q.
        • et al.
        Do patients with multiple sclerosis derive more benefit from robot-assisted gait training compared with conventional walking therapy on motor function? A meta-analysis.
        Front. Neurol. 2017; 8: 260
        • Androwis, G.J.
        • Pilkar, R.
        • Ramanujam, A.
        • Nolan, K.J.
        Electromyography assessment during gait in a robotic exoskeleton for acute stroke.
        Front. Neurol. 2018; 9: 630
        • Ramanujam, A.
        • Cirnigliaro, C.M.
        • Garbarini, E.
        • Asselin, P.
        • Pilkar, R.
        • Forrest, G.F.
        Neuromechanical adaptations during a robotic powered exoskeleton assisted walking session.
        J. Spinal Cord Med. 2018; 41: 518-528
        • Jensen, S.K.
        • Yong, V.W.
        Activity-dependent and experience-driven myelination provide new directions for the management of multiple sclerosis.
        Trends Neurosci. 2016; 39: 356-365
        • Berriozabalgoitia, R.
        • Bidaurrazaga-Letona, I.
        • Otxoa, E.
        • Urquiza, M.
        • Irazusta, J.
        • Rodriguez-Larrad, A.
        Overground robotic program preserves gait in individuals with multiple sclerosis and moderate to severe impairments: a randomized controlled trial.
        Arch. Phys. Med. Rehabil. 2020;
        • Drużbicki, M.
        • Guzik, A.
        • Przysada, G.
        • et al.
        Effects of robotic exoskeleton-aided gait training in the strength, body balance, and walking speed in individuals with multiple sclerosis: a single-group preliminary study.
        Arch. Phys. Med. Rehabil. 2020;
        • Afzal, T.
        • Tseng, S.-C.
        • Lincoln, J.A.
        • Kern, M.
        • Francisco, G.E.
        • Chang, S.-H.
        Exoskeleton-assisted gait training in persons with multiple sclerosis: a single-group pilot study.
        Arch. Phys. Med. Rehabil. 2020; 101: 599-606
        • McGibbon, C.
        • Sexton, A.
        • Gryfe, P.
        • et al.
        Effect of using of a lower-extremity exoskeleton on disability of people with multiple sclerosis.
        Disabil. Rehabil. 2021; : 1-8
        • Kozlowski, A.J.
        • Fabian, M.
        • Lad, D.
        • Delgado, A.D.
        Feasibility and safety of a powered exoskeleton for assisted walking for persons with multiple sclerosis: a single-group preliminary study.
        Arch. Phys. Med. Rehabil. 2017; 98: 1300-1307
        • Androwis, G.J.
        • Kwasnica, M.A.
        • Niewrzol, P.
        • et al.
        Mobility and cognitive improvements resulted from overground robotic exoskeleton gait-training in persons with MS.
        in: Paper presented at: 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2019
        • Straudi, S.
        • Basaglia, N.
        Neuroplasticity-based technologies and interventions for restoring motor functions in multiple sclerosis.
        Multiple Sclerosis: Bench to Bedside. Springer, 2017: 171-185
        • Pascual-Leone, A.
        • Amedi, A.
        • Fregni, F.
        • Merabet, L.B.
        The plastic human brain cortex.
        Annu. Rev. Neurosci. 2005; 28: 377-401
        • Sandroff, B.M.
        • Jones, C.D.
        • Baird, J.F.
        • Motl, R.W.
        Systematic review on exercise training as a neuroplasticity-inducing behavior in multiple sclerosis.
        Neurorehabil. Neural Repair. 2020; 1545968320921836
        • Sandroff, B.M.
        • Wylie, G.R.
        • Sutton, B.P.
        • Johnson, C.L.
        • DeLuca, J.
        • Motl, R.W.
        Treadmill walking exercise training and brain function in multiple sclerosis: preliminary evidence setting the stage for a network-based approach to rehabilitation.
        Mult. Scler. J.–Exp., Transl. Clin. 2018; 42055217318760641
        • Tona, F.
        • Petsas, N.
        • Sbardella, E.
        • et al.
        Multiple sclerosis: altered thalamic resting-state functional connectivity and its effect on cognitive function.
        Radiology. 2014; 271: 814-821
        • Bollaert, R.E.
        • Poe, K.
        • Hubbard, E.A.
        • et al.
        Associations of functional connectivity and walking performance in multiple sclerosis.
        Neuropsychologia. 2018; 117: 8-12
        • Kelly, C.
        • Castellanos, F.X.
        Strengthening connections: functional connectivity and brain plasticity.
        Neuropsychol. Rev. 2014; 24: 63-76
        • Hauser, S.L.
        • Dawson, D.M.
        • Lehrich, J.R.
        • et al.
        Intensive immunosuppression in progressive multiple sclerosis: a randomized, three-arm study of high-dose intravenous cyclophosphamide, plasma exchange, and ACTH.
        N. Engl. J. Med. 1983; 308: 173-180
        • Nilsagard, Y.
        • Lundholm, C.
        • Gunnarsson, L.G.
        • Denison, E.
        Clinical relevance using timed walk tests and ‘timed up and go’testing in persons with multiple sclerosis.
        Physiother. Res. Int. 2007; 12: 105-114
        • Sebastião, E.
        • Sandroff, B.M.
        • Learmonth, Y.C.
        • Motl, R.W.
        Validity of the timed up and go test as a measure of functional mobility in persons with multiple sclerosis.
        Arch. Phys. Med. Rehabil. 2016; 97: 1072-1077
        • Goldman, M.D.
        • Marrie, R.A.
        • Cohen, J.A.
        Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls.
        Mult. Scler. J. 2008; 14: 383-390
        • Smith, A.
        Symbol Digit Modalities Test: Manual.
        Western Psychological Services, Los Angeles1982
        • Barbarulo A.M.
        • Giacomo L.
        • Elisabetta S.
        • Luigi T.
        • Dario G.
        • Mariateresa E.
        • Teresa C.
        • et al.
        Integrated cognitive and neuromotor rehabilitation in multiple sclerosis: a pragmatic study.
        Front. Behav. Neurosci. 2018; 196
        • Benedict, R.H.
        • DeLuca, J.
        • Phillips, G.
        • et al.
        Validity of the symbol digit modalities test as a cognition performance outcome measure for multiple sclerosis.
        Mult. Scler. J. 2017; 23: 721-733
        • Sutliff, M.H.
        • Bennett, S.E.
        • Bobryk, P.
        • et al.
        Rehabilitation in multiple sclerosis: commentary on the recent AAN systematic review.
        Neurology. 2016; 6: 475-479
        • Sattelmayer, M.
        • Chevalley, O.
        • Steuri, R.
        • Hilfiker, R.
        Over-ground walking or robot-assisted gait training in people with multiple sclerosis: does the effect depend on baseline walking speed and disease related disabilities? A systematic review and meta-regression.
        BMC Neurol. 2019; 19: 93
        • Haselkorn, J.K.
        • Hughes, C.
        • Rae-Grant, A.
        • et al.
        Summary of comprehensive systematic review: rehabilitation in multiple sclerosis: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology.
        Neurology. 2015; 85: 1896-1903
        • Aisen, M.L.
        • Sevilla, D.
        • Fox, N.
        Inpatient rehabilitation for multiple sclerosis.
        J. Neurol. Rehabil. 1996; 10: 43-46
        • Cohen, J.
        Statistical Power Analysis for the Behavioral Sciences.
        Second Edition. 12 Lawrence Erlbaum Associates Inc, Hillsdale, New Jersey1988: 13
        • Rousselet, G.A.
        • Pernet, C.R.
        Improving standards in brain-behavior correlation analyses.
        Front. Hum. Neurosci. 2012; 6: 119
        • Mohr, D.C.
        • Spring, B.
        • Freedland, K.E.
        • et al.
        The selection and design of control conditions for randomized controlled trials of psychological interventions.
        Psychother. Psychosom. 2009; 78: 275-284
        • Evans, N.
        • Hartigan, C.
        • Kandilakis, C.
        • Pharo, E.
        • Clesson, I.
        Acute cardiorespiratory and metabolic responses during exoskeleton-assisted walking overground among persons with chronic spinal cord injury.
        Top. Spinal Cord Injury Rehabil. 2015; 21: 122-132
        • Gagnon, D.H.
        • Escalona, M.J.
        • Vermette, M.
        • et al.
        Locomotor training using an overground robotic exoskeleton in long-term manual wheelchair users with a chronic spinal cord injury living in the community: lessons learned from a feasibility study in terms of recruitment, attendance, learnability, performance and safety.
        J. Neuroeng. Rehabil. 2018; 15: 12
        • Contreras-Vidal, J.L.
        • Kilicarslan, A.
        • Huang, H.H.
        • Grossman, R.G.
        Human-centered design of wearable neuroprostheses and exoskeletons.
        AI Magazine. 2015; 36: 12-22
        • Eijlers, A.J.
        • Dekker, I.
        • Steenwijk, M.D.
        • et al.
        Cortical atrophy accelerates as cognitive decline worsens in multiple sclerosis.
        Neurology. 2019; 93: e1348-e1359
        • Motl, R.W.
        • Zivadinov, R.
        • Bergsland, N.
        • Benedict, R.H.
        Thalamus volume and ambulation in multiple sclerosis: a cross-sectional study.
        Neurodegener. Dis. Manag. 2016; 6: 23-29
        • Manuli, A.
        • Maggio, M.G.
        • Latella, D.
        • et al.
        Can robotic gait rehabilitation plus Virtual Reality affect cognitive and behavioural outcomes in patients with chronic stroke? A randomized controlled trial involving three different protocols.
        J. Stroke Cerebrovasc. Dis. 2020; 29104994
        • Calabrò, R.S.
        • Naro, A.
        • Russo, M.
        • et al.
        Shaping neuroplasticity by using powered exoskeletons in patients with stroke: a randomized clinical trial.
        J. Neuroeng. Rehabil. 2018; 15: 1-16
        • Motl, R.W.
        • Sandroff, B.M.
        • DeLuca, J.
        Exercise training and cognitive rehabilitation: a symbiotic approach for rehabilitating walking and cognitive functions in multiple sclerosis?.
        Neurorehabil. Neural Repair. 2016; 30: 499-511
        • Morrow, S.A.
        • Drake, A.
        • Zivadinov, R.
        • Munschauer, F.
        • Weinstock-Guttman, B.
        • Benedict, R.H.
        Predicting loss of employment over three years in multiple sclerosis: clinically meaningful cognitive decline.
        Clin. Neuropsychol. 2010; 24: 1131-1145
        • Sandroff, B.M.
        • Motl, R.W.
        • Scudder, M.R.
        • DeLuca, J.
        Systematic, evidence-based review of exercise, physical activity, and physical fitness effects on cognition in persons with multiple sclerosis.
        Neuropsychol. Rev. 2016; 26: 271-294
        • Cutter, G.
        Effect size–one size doesn’t fit all.
        Mult. Scler. Relat. Disord. 2020; 102587
        • Motl, R.W.
        • Sandroff B.M.
        Current perspectives on exercise training in the management of multiple sclerosis.
        Expert. Rev. Neurother. 2020; 20: 855-865