Gait Adaptation Study

Multiple sclerosis (MS) is a neurological disorder that affects nearly 1 million adults in the United States (Wallin, Culpepper et al. 2019). Mobility and cognitive dysfunction are highly prevalent and debilitating consequences of MS, whereby upwards of 90% of patients present with mobility disability and upwards of 65% of patients demonstrate cognitive impairment (Chiaravalloti and DeLuca 2008, van Asch 2011, Chiaravalloti, Genova et al. 2015). Mobility and cognitive disability tend to co-occur in pwMS (i.e., cognitive-motor coupling), perhaps based on damage to neural substrates that are important for both domains of functioning (Benedict, Holtzer et al. 2011) (e.g., disruptions in the central nervous system (CNS) connectivity (Janssen, Boster et al. 2013, Wojtowicz, Mazerolle et al. 2014, Sbardella, Tona et al. 2015, Neva, Lakhani et al. 2016, Meijer, Eijlers et al. 2017, Saleh, Sandroff et al. 2018, van Geest, Douw et al. 2018, Chaves, Wallack et al. 2019, Hoxha, Glassen et al. 2019). Alarmingly, MS-related mobility and cognitive disability are both poorly managed with pharmacotherapy (DeLuca, Chiaravalloti et al. 2020). This underscores the importance of rehabilitation as an approach to managing MS-related mobility and cognitive dysfunction.

One particularly promising approach for rehabilitating MS-related mobility and cognitive dysfunction involves gait adaptability training (GAT, i.e., navigating real-world environments using precision stepping and obstacle avoidance) using mixed/augmented reality. Gait adaptability is inherently involved in most activities of daily living, where interaction with the physical environment requires proactive and reactive gait adaptations (Weerdesteyn, Hollands et al. 2018). GAT repeatedly and simultaneously trains mobility and cognitive processes and progressively increases in difficulty over time, leading to adaptations in both functions. This is supported by preliminary evidence of GAT-related improvements in mobility and cognition in neurological populations (van Ooijen, Heeren et al. 2015, Lau, Regis et al. 2022), including one recent clinical trial of virtual reality, dual-task treadmill training that involved some elements of gait adaptability that reported improvements in both mobility and cognition in pwMS (Hsieh, Mirelman et al. 2020, Galperin, Mirelman et al. 2022). However, there have been no trials of GAT, per se, on mobility and cognition in MS to date.

Of further importance, GAT research, in general, has not focused on restoring mobility and cognitive functioning in those who present with impairments in both mobility and cognition. Such research introduces the risk of ceiling effects and does not focus on the population presenting with the problems being studied, limiting inferences of treatment or restoration of function. Moreover, GAT research has not examined intervention effects on shared neural substrates of co-occurring, MS-related mobility and cognitive disability for establishing such an approach as a neuroplasticity-inducing behavior. Indeed, we have hypothesized that the neural processing and integration of multisensory inputs required for the regulation of physiological systems during acute walking becomes more efficient over time with ongoing, repeated, and progressively more challenging bouts of walking (i.e., GAT), resulting in adaptations in mobility and cognition, based on proximal adaptations in CNS connectivity (Sandroff, Motl et al. 2018). By extension, enhancing the multisensorial experience of walking with GAT should maximize adaptations in CNS connectivity, mobility, and cognition in pwMS (Sandroff, Motl et al. 2018). This proposal positions GAT as an innovative rehabilitative approach for enhancing the multisensorial demands of walking for eliciting meaningful mobility, cognitive, and neural adaptations in pwMS-related mobility and cognitive impairment, thus overcoming two major limitations associated with the rigor of the prior research.

Our preliminary EEG data suggest that avoiding obstacles during walking (i.e., gait adaptability) is associated with abnormal CNS (within-brain) effective connectivity (EC) in networks that are critical for mobility and cognition and in brain-to-muscle connectivity (BMC) in pwMS (Saleh, Glassen et al. 2020). This suggests that perhaps progressive, repetitive GAT might reverse such abnormalities and result in neurophysiological and functional benefits over time in pwMS who present with dysfunction. Thus, the proposed randomized controlled trial (RCT) in aim 1 \& 2 will examine the effects of GAT relative to an active control condition (basic physical training; BPT) on mobility and cognition, along with innovative endpoints of CNS connectivity and BMC in 30 pmMS participants. Indeed, including functional and neural endpoints is essential for providing initial empirical support for GAT as a biologically grounded, neuroplasticity-inducing, rehabilitative approach for managing the prevalent, debilitating, and co-occurring mobility and cognitive disability in MS.

In aim 1.a: Before recruitment and enrollment of MS participants, the mixed reality (GAT) training paradigm will be tested on group of 15 healthy subjects. The purpose is to determine the feasibility of the use of the mixed reality (MR) game with simultaneous EEG testing/recording. This goal of this pilot phase of aim 1 on healthy subjects is to confirm system's usability and comfort level by recording the Mixed Reality Concerns (MRC) questionnaire to determine trust in the MR system, the Virtual Embodiment Questionnaire (VEQ) to determine their comfort level with seeing an avatar represent themselves by copying their movement pattern, and the System Usability Survey (SUS) for overall usability of the software.

In aims 1.b \&2: We propose a single-blind, 2-arm RCT that compares the effects of a 12-session GAT intervention with an active control condition on functional and neuroimaging outcomes in 30 participants from three groups: 1) MS (n=30) with mobility and cognitive impairment. Upon satisfying inclusion/exclusion criteria, participants will undergo baseline mobility, cognitive, and neuroimaging outcome assessments. Following baseline testing, participants will be randomly assigned to either the intervention (GAT) or active control (BPT) condition. GAT will involve completing 12 training sessions using mixed reality (MR) over 4 weeks. BPT will involve completing 12 sessions of upper extremity movement and walking without purposeful gait adaptations over 4 weeks (i.e., walking alone). After the 4-week study period, participants will complete the same assessments as at baseline (i.e., follow-up) with blinded assessors.

GAT Intervention Condition: Participants will be trained for 12 sessions (4 weeks, 3 sessions per week) in a supervised paradigm that involves navigating an MR environment. This duration of GAT is consistent with research in neurological populations reporting mobility and cognitive improvements (Heeren, Van Ooijen et al. 2013, van Ooijen, Heeren et al. 2015, Timmermans, Roerdink et al. 2016, Hulzinga, Seuthe et al. 2022). The MR environment, built in Unity engine using the Mixed Reality Toolkit, will include externally cued targets that will appear as virtual objects. Participants are expected to perform precise real word activities like stacking books on shelves, picking up objects, etc (see Figure 1 in section 10). Microsoft HoloLens 2 MR goggles will be used to project these targets onto the real-world environment. For the current proposal, a library of 5 training applications during walking will be standardized across participants. Figure 1 shows an example of two applications. Within a given training session, the applications will be delivered in a predefined order across participants. The GAT prescription is primarily based on walking, but it involves integrative dual task activities like reaching, grasping, precision stepping. The progression in GAT training sessions will be individualized based on patients' performance and will involve increasing task difficulty. Specifically, the difficulty of the GAT tasks will progressively increase based on actual performance. Participants will wear motion sensors, and sensor information will be streamed and processed in real time to calculate performance accuracy and to adapt the game properties to provide a progression of training conditions, and to allow participants to practice the task more naturally and with high success. Once participants achieve a criterion of 80% accuracy on a given task, difficulty will incrementally increase upon the next administration of the task. Task difficulty will increase where GAT condition will require higher accuracy in precision stepping, providing higher number and faster rate of obstacle representations, and more stimulations in navigating the virtual environment.

BPT Condition: Training will take place with the same training dosage (duration) as the GAT condition; however, there will be no environmental or cognitive challenges applied. To control for upper extremity activities in the GAT training conditions (stacking objects, grasping, etc), BPT will include 10 minutes of practicing single upper extremity tasks while seated (reaching, grasping), followed by performing 30 minutes of single walking training in the gait lab without any cognitive demands.

In aim 3: We propose an exploratory study to measure the neural and behavioral characteristics of 40 children participants (4 groups with 10 children (13-17 years) with MS, MOG, NMOSD and gender- age- matched children with no neurological conditions). Upon satisfying inclusion/exclusion criteria, participants in this aim will attend a one visit session and complete the baseline measurements in aims 1\&2 detailed above (mobility, cognitive, and neuroimaging outcome assessments).

We plan to recruit 70 subjects [30 subjects (aim 1\&2); 40 subjects (aim 3)] in the pilot phase of the study. We decided on the sample size based on the feasibility to acquire enough pilot data to calculate effect size.

Aim 1a consists of one visit for ~2 hours. Aim 1b \& 2 consist of: three assessment visits ~3 hours each; and 12 training visits 1 hour each. A total of 21 hours over 12 weeks to complete study.

Aim 3 consists of one visit 3 hours long

Aim 1a Age 18 to 80 yrs Ability to attend to computer screen and follow instructions in English which will be assessed by MiniMental test requiring a participant to score >24 (out of 30 max score) Right-handed as determined by research diagnostic criteria of the Edinburgh Handedness Inventory (Oldfield 1971) Self report of normal or corrected-to-normal vision

Exclusion Criteria Psychiatric disorders, only those requiring treatment because the medications may affect brain activation Neurological disorder that may affect cognition, balance, and/or physical wellness such as brain injury, or stroke.

Orthopedic injuries or neuromuscular disorder that may affect balance or mobility Engagement in other cognitive or physical training program while enrolled in this study Ongoing relapse or steroid treatment during the 30 days preceding enrollment

Aim 1b \& 2:

Inclusion Criteria Age 18 to 80 yrs Clinically definite MS diagnosis(Thompson et al, 2018) Ability to attend to computer screen and follow instructions in English which will be assessed by MiniMental test requiring a participant to score >24 (out of 30 max score) Right-handed as determined by research diagnostic criteria of the Edinburgh Handedness Inventory (Oldfield 1971) Self report of normal or corrected-to-normal vision

Exclusion Criteria Psychiatric disorders, only those requiring treatment because the medications may affect brain activation Neurological disorder that may affect cognition, balance, and/or physical wellness such as brain injury, or stroke.

Orthopedic injuries or neuromuscular disorder that may affect balance or mobility Engagement in other cognitive or physical training program while enrolled in this study Ongoing relapse (new or returning neurological symptoms) or steroid treatment during the 30 days preceding enrollment

Aim 3 (* applicable to controls):

General criteria will be reviewed by RC and PI. The medical report and/or list of medications will be reviewed by Co-PI, a Neurologist.

Inclusion Criteria:

Pediatric Onset Multiple Sclerosis (POMS) diagnosed according to the most recent diagnostic criteria Or MOG or NMOSD MS Age 13 to 17 yrs* Ability to attend to computer screen and follow instructions in English which will be assessed by MiniMental test for children (Salvador, Moura et al. 2019) with score >31 out of max score of 37 Self report normal or corrected-to-normal vision*

Exclusion Criteria Psychiatric disorders, only those requiring treatment because the medications may affect brain activation * Neurological disorders (other than POMS, MOG, and NMOSD) that may affect cognition, balance or mobility, for example brain injury, stroke, epilepsy, autism, etc.

Orthopedic injuries neuromuscular disorder that may affect balance or mobility Engagement in other cognitive or physical training program while enrolled in this study

Aim 1b and Aim 2b involve testing GAT vs BPT. GAT Intervention Condition: Participants will be trained for 12 sessions (4 weeks, 3 sessions per week) in a supervised paradigm in the GAIT lab at SHP, Rutgers RBHS. This duration of GAT is consistent with research in neurological populations reporting mobility and cognitive improvements [21], [109-111]. The VR training is programmed to allow training pwMS to walk while stepping on tracks, avoiding obstacles, picking up grocery items in a virtual environment, etc. (see Figure 4). The RW conditions are designed based on the FALLS [112]training condition and include slalom walking, steps, and obstacle avoidance (see Figure 5). A similar training setup has been implemented in stroke research and improving OA and Stroop task success rates[17]. We note that the cognitive challenges incorporated with GAT approximate those of the highly successful APT cognitive rehabilitation paradigm[113]. As GAT focuses on increasing the accuracy and speed of responding to environmental challenges that occur in the real world, the training stimulus will not progress in terms of walking speed across sessions. Regarding the progression of environmental challenges, GAT will require progressively higher accuracy in precision stepping, providing a higher number and faster rate of obstacle representations, and more stimulations in navigating the virtual environment. If a participant misses a given training session, a make-up session will be scheduled. If a participant misses multiple sessions, they will have up to 6-weeks to complete the 4-week training paradigm. Only cases where participants attend ≥10 of the 12 possible sessions will be included in the completers analysis.

BPT is designed based on conventional walking training where there is no innovative component in training conditions.

Aim 3 is a single session arm of the study where the independent variables are the subject groups that we will compare in terms of physical and cognitive function, and neurophysiology assessments (pediatric MS, NMOSD, MOG, and HC).

Aim 1a. - System user experience Mixed Reality Concerns (MRC) questionnaire to determine trust in the MR system, the Virtual Embodiment Questionnaire (VEQ) to determine subjects' comfort level with seeing an avatar represent themselves by copying their movement pattern, and the System Usability Survey (SUS) for overall usability of the software.

• Neurophysiology measures
• Resting-state brain connectivity
• EEG activity during task

Aim 1b \& 2 outcomes: A battery of outcome assessment will include evaluation of disability, physical function, cognitive function, and neurophysiology measures, see list below.

Disability

1. Expanded Disability Status Score- EDSS (Kurtzke 1983)
2. MS functional composite (MSFC)

Physical Function

1. 25ftWT, primary outcome
2. Dynamic Gait Index (DGI), primary outcome
3. 6 min walking test (endurance test)
4. Cardiovascular measure (resting heart rate) Cognitive Function BICAMS- Brief International Assessment for MS Neurophysiology measures
5. Resting-state brain connectivity 2. Computed Muscle Control- CMC during walking (EEG-EMG coherence)

Aim 3 outcomes: A battery of outcome assessment will include evaluation of physical function, cognitive function and neurophysiology measures, see table 1.

Physical Function

• 25ftWT, primary outcome
• Dynamic Gait Index (DGI), primary outcome
• 6 min walking test (endurance test)
• Cardiovascular measure (resting heart rate) Cognitive Function
• Brief neuropsychological battery for children which include selective reminding test, SDMT, trail making test, and Vocabulary test from the Wechsler Intelligence Scale for children [brief neuropsychological battery for children: a screening tool for cognitive impairment in childhood and juvenile multiple sclerosis]

Neurophysiology measures

• Resting-state brain connectivity
• Computed Muscle Control- CMC during walking (EEG-EMG coherence)