Fatigue Limits Motor and Cognitive Improvements after High-intensity Exercise Prior to Balance Training over Telehealth in People with Spinocerebellar Ataxia Authors: Chelsea E. Macpherson, Fatima Awad, Vruta Rana, Sheng-Han Kuo, Lori Quinn Published: 2025 ABSTRACT ObjectiveThis pilot randomized controlled trial explored feasibility, and preliminary effects from remotely delivered high- versus low-intensity exercise prior to balance training for people with Spinocerebellar Ataxia (SCA).MethodsTwenty participants with SCA (types 1, 2, 3, or 6) were randomized to high- or low-intensity exercise (30 min), followed by balance training (30 min), delivered via telehealth twice-weekly for 8-weeks. Exercises were progressive and individualized based on ataxia severity, mobility, and home equipment. Outcomes included disease-specific measures (Scale for Assessment and Rating of Ataxia at home, Cerebellar Cognitive Affective Syndrome Scale), and fatigue (Fatigue Severity Scale) assessed at baseline, mid- and post-intervention.ResultsEighteen participants completed the intervention with high adherence. Both groups improved on disease-specific measures, with greater gains in cognition and reduced fatigue in the low-intensity group.ConclusionsRemote delivery of exercise and balance training is feasible in SCA. Fatigue may limit tolerability of higher exercise intensities. National Institutes of Health Clinical Trials Registration Number: NCT05826171 Spinocerebellar ataxias (SCA) are a group of rare neurogenetic disorders characterized by symptoms of ataxia that affect balance, gait, motor learning, and motor control (Diallo et al., 2020). The cerebellum is the most affected brain region, and progressive cerebellar degeneration leads to motor learning deficits that impact functional independence and capacity for neuroplastic adaptation (Mitoma et al., 2020). Strategies for clinical management focus largely on symptoms; however, they often fail to address impairments in motor learning. These motor learning impairments likely influence sustained benefits following rehabilitation interventions (Miyai et al., 2012; Morton, 2006). Evidence supporting rehabilitation interventions for people with SCA (PwSCA) continues to evolve, with studies evaluating different exercise types (e.g., aerobic exercise, balance training, vestibular exercise, and task specific training), but also multi-modal interventions, group therapies, and alternative therapies (e.g., biofeedback, exergaming, etc.) (Bogaert et al., 2023; Heusel-Gillig & Hall, 2023; Ilg et al., 2014; Jabri et al., 2022; Marquer et al., 2014; Matsugi, 2017; Milne et al., 2017; Winser et al., 2018). Most studies have evaluated effects from individual exercises such as balance training (Keller & Bastian, 2014; Rodríguez-Díaz et al., 2018; Velázquez-Pérez et al., 2019) and more recently aerobic exercise (Barbuto et al., 2020; Di Liegro et al., 2019; Müller et al., 2020) given their potential to improve ataxia symptoms, enhance neuroplasticity, and delay neurodegeneration (Mellesmoen et al., 2018; Velázquez-Pérez et al., 2019). Although condition specific exercise guidelines are starting to emerge (e.g., Parkinson Disease (Osborne et al., 2021), Multiple Sclerosis (Kalb et al., 2020)), many neurodegenerative conditions including cerebellar ataxias continue to follow the American College of Sports Medicine Guidelines for special populations (Riebe et al., 2018). For aerobic exercise specifically, the American College of Sports Medicine recommends engaging in moderate intensity (50–70% of heart rate maximum) exercise 30-minutes a day 3–5 days per week, for a total of 150-minutes weekly, or vigorous intensity (70–85% of heart rate maximum) exercise at least 2 days per week, for a total of 75-minutes weekly. Combinations of both moderate and high intensity exercise can also be performed to meet set guidelines (e.g., 150-minutes of moderate to vigorous physical activity (Sabag et al., 2022)). One such method is to use high-intensity interval training. High-intensity interval training includes repeated bouts of high intensity exercise (upwards of 80% of heart rate maximum) followed by varying lengths of active or passive recovery periods (Sabag et al., 2022). Results of high intensity interval training for people with cardiac conditions (e.g., coronary artery disease, post myocardial infarct, heart failure) as well as neurodegenerative conditions (e.g., Parkinson Disease, Multiple Sclerosis) have shown comparability in outcomes to continuous forms of moderate to vigorous intensity exercise the American College of Sports Medicine recommends engaging in moderate intensity (50–70% of heart rate maximum) exercise 30-minutes a day 3–5 days per week, for a total of 150-minutes weekly, or vigorous intensity (70–85% of heart rate maximum) exercise at least 2 days per week, for a total of 75-minutes weekly. Combinations of both moderate and high intensity exercise can also be performed to meet set guidelines (e.g., 150-minutes of moderate to vigorous physical activity (Sabag et al., 2022)). One such method is to use high-intensity interval training. High-intensity interval training includes repeated bouts of high intensity exercise (upwards of 80% of heart rate maximum) followed by varying lengths of active or passive recovery periods (Sabag et al., 2022). Results of high intensity interval training for people with cardiac conditions (e.g., coronary artery disease, post myocardial infarct, heart failure) as well as neurodegenerative conditions (e.g., Parkinson Disease, Multiple Sclerosis) have shown comparability in outcomes to continuous forms of moderate to vigorous intensity exercise For balance training, interventions should incorporate components of proactive, reactive, and steady state balance for a well-rounded and multidimensional training (Gill-Body et al., 2021; Sparrow et al., 2016). Guidelines for balance training in PwSCA have not yet been established, however, gathering from other neurodegenerative conditions such as Parkinson Disease, multidimensional balance training should occur for 30–120 minutes per session, at a frequency of two to three times per week, for at least 5 to 10 weeks to see changes on condition specific motor scales, dynamic tests of balance (e.g., the Functional Gait Assessment), and measures of balance confidence (e.g., the ABC Scale) (Osborne et al., 2021). Additionally, in order to achieve the greatest functional and cognitive benefit, balance training should be guided by increasing task difficulty that reaches a person’s individual challenge point using theoretical frameworks from both Gentile’s taxonomy of tasks and the challenge point hypothesis (Bae et al., 2025; Conradsson et al., 2017; Gentile, 2000; Guadagnoli & Lee, 2004). Studies in PwSCA and other neurological conditions, highlight the importance of moderate to vigorous intensity exercise (Ahlskog, 2011; Barbuto et al., 2020) and the importance of incorporating challenging tasks of progressive difficulty for enhanced motor learning and neuroprotective benefit (Guadagnoli & Lee, 2004; Hirsch & Farley, 2009; Hornby et al., 2020; Osborne et al., 2021). However, balance training and aerobic exercise may independently activate different neural recovery mechanisms, and combining each intervention may offer additional benefit for PwSCA. Motor priming refers to a non-conscious process in which exposure to one stimulus influences the response to a subsequent, separate stimulus, thus facilitating learning or performance (Stoykov & Madhavan, 2015). Recent literature suggests that when aerobic exercise is positioned prior to task-specific exercise (e.g., balance training) it may provide effects of “priming” that boost motor performance (Macpherson et al., 2023; Moriarty et al., 2019; Steib et al., 2018; Stoykov & Madhavan, 2015). Mechanistically this is thought to occur as aerobic exercise improves neuroplasticity, by way of increasing blood flow, oxygenation, receptor activity, and release of neurotropic factors (El-Sayes et al., 2019), while balance training increases cortical thickness and excitability (Rogge et al., 2018; Taube et al., 2007; Ueta et al., 2022). The cerebellum is heavily involved in motor learning, and degeneration of cerebellar structures disrupts neural connectivity, motor learning ability, and functional mobility (De Zeeuw & Ten Brinke, 2015). Motor priming through exercise has potential to boost motor learning and advance functional gains. To date, no rehabilitation interventions have targeted motor learning via long-term processes of motor priming with exercise in PwSCA. This pilot randomized controlled trial was informed by prior pilot study work (Macpherson et al., 2023). The current study aimed to evaluate the feasibility and preliminary efficacy of an 8-week remote exercise intervention that compared high intensity exercise prior to balance training (HIGH-BT) to low intensity exercise prior to balance training (LOW-BT) in people with genetically confirmed SCA types 1, 2, 3, or 6. Methods This pilot randomized control trial ran from April 2023 to March 2024. The study protocol was registered on the National Institute of Health Clinical Trials Registry (NCT0582617, Date of Trial Registration: 2023-03-22), and followed guidelines established by the Consolidated Standards of Reporting Trials (Boutron et al., 2017). Recruitment Participants with SCA (n=20) were recruited from Columbia University Irving Medical Center Ataxia & Tremor Clinic, ClinicalTrials.gov, the National Ataxia Foundation (study recruitment posting, support group talks), and by word of mouth. Those who expressed interest received information by study personnel and were screened over the phone by the study principal investigator (CM). Eligibility Criteria Participants were included if they were (1) age >18 years; (2) had genetic confirmation of SCA1, 2, 3, 6 or 7; (3) had a Scale for Ataxia Rating and Assessment (Schmitz-Hubsch et al., 2006) score between 8–25/40, capturing mild-moderate disease; (4) were able to walk with or without assistive devices; (5) successfully completed the Physical Activity Readiness Questionnaire (S. Thomas et al., 1992) to confirm no medical contraindications to exercise; and (6) had care partner availability during intervention. Participants were excluded if they had (1) severe non-ataxic motor symptoms measured by Inventory of Non-Ataxia Signs (Jacobi et al., 2013); (2) any visual complications associated with ataxia (e.g., retinal or optic nerve involvement); (3) spontaneous nystagmus; (4) other concurrent disease of the cerebellum (e.g., stroke, multiple sclerosis); (5) cardiopulmonary diseases that would restrict exercise; or (6) currently engaged in >3 week aerobic exercise and/or balance training. All eligible participants were enrolled in the study by the study principal investigator (CM), and signed informed consent approved by the Teachers College, Columbia University Institutional Review Board (TCIRB#23-088). Randomization A block stratified randomization module was developed in REDCap (Vanderbilt University, Nashville, TN, USA) that randomized participants in blocks of four by age (> or < 65 years), and sex (male, female). Randomization was conducted using REDCap’s block randomization tool after baseline assessment. After randomization, participants were informed of their group allocation (either HIGH-BT, or LOW-BT) via secure e-mail communication (compliant with the Health Insurance Portability and Accountability Act (HIPAA)) between the study principal investigator (CM) and the participant. Participants were instructed not to disclose their group assignment in online ataxia forums or support groups to minimize contamination. Study Materials Prior to baseline assessments, participants were sent a study packet containing a gait-belt to reduce fall risk, a chest-worn smart-phone holder for digital balance and functional outcome assessments via Mon4t Encephalog Balance Assessment System (Mon4t LLC, Binyamina, Israel), and a wrist-worn consumer-grade wearable device (FitBit Charge HR 5, San Francisco, CA, USA). Safety Considerations All participants and their care-partners underwent a 30-minute safety training session via telehealth with the study interventionist (CM) prior to study initiation (Week 0). 1 Training included establishment of safe environments, participant guarding (Clarke et al., 2013; Kalra et al., 2004; Vloothuis et al., 2019), and medical safety (e.g., warning signs of an adverse response to exercise) (Macpherson, 2024). Additionally, participants developed an emergency action plan 2 with the interventionist, which consisted of location of exercise sessions, emergency contact name, medical providers, and local hospital of choice. These details were shared with participants’ care-partners and/or emergency contacts. For all telehealth sessions, a HIPAA-secure automatic invite containing location, time, and type of session was sent to the participant, care-partner, or emergency contact. In the event of a medical emergency, the emergency action plan was activated. All adverse events were documented and reported to the study principal investigator, research team, and Teachers College, Columbia University IRB. Outcome Assessments Primary metrics of feasibility included: (1) recruitment rate, defined as the percentage of participants referred from medical providers who then signed an informed consent and enrolled in the study; (2) retention rate, defined as the percentage of enrolled participants who completed the intervention and post-intervention assessments; (3) safety, defined as the number of adverse and serious adverse events which were reported during the trial; (4) adherence to the intervention, defined as the attendance to all sessions of the intervention; (5) cardiovascular tolerability to the intervention which was defined as meeting target heart rate reserve or rating of perceived exertion 70% of the time, and (6) acceptability, captured via a custom post-intervention questionnaire that was administered to participants via REDCap survey (see Appendix A) (Macpherson, 2024). The questionnaire consisted of 11 closed-ended questions scored on a 5-point Likert scale, that discussed four domains: overall satisfaction, interaction with the physical therapist, telehealth, and self-efficacy for exercise. There were seven open ended questions that evaluated four domains: motivation for participation, program assessment, change in function, integration of technology or telehealth. Primary clinical outcomes were assessed by a trained clinical evaluator, who remained constant throughout the trial and was blinded to group allocation (VR). Outcomes included: (1) the Scale for Assessment and Rating of Ataxia at Home (SARAHome) (Grobe-Einsler et al., 2021) and (2) the Cerebellar Cognitive Affective Syndrome scale (CCAS Scale) (Hoche et al., 2018) both of which were video-recorded for quality assurance. The SARAHome Scale is a remote version of the original Scale for Assessment and Rating of Ataxia scale. It consists of 5 of the 8 original Scale for Assessment and Rating of Ataxia items (gait, stance, speech, nose-finger test, fast alternating hand movements). Total scores range from 0 to 28, with 0 reflecting no presence of ataxia and 28 reflecting severe ataxia. The SARAHome was found to be highly correlated with the original Scale for Assessment and Rating of Ataxia scale (r = 0.98, p < 0.0001) (Grobe-Einsler et al., 2021). The Cerebellar Cognitive Affective Syndrome scale is a 10-item scale that provides pass/fail indications per item (representing the total score out of 10 points), with a maximum raw score of 120 points. Pass/fail criteria are used to determine if a participant had possible (one failed test item), probable (two failed test items), or clinically definite (≥ three failed test items) CCAS. A cut-off score of 80/120 has been documented as criteria to obtain intervention for cognitive impairment (Hoche et al., 2018). Secondary outcomes of functional mobility and balance included the Timed Up and Go (TUG) test (Macpherson, C.E. et al., 2025; Podsiadlo & Richardson, 1991), 30 second Sit to Stand (30 sec STS) test (Khalil et al., 2010; Polidori et al., 2024), and measures of static postural sway (neutral stance, feet together eyes open, feet together eyes closed, tandem stance, single leg stance right, single leg stance left) (Macpherson, C.E. et al., 2025). These outcomes were digitally captured using the Mon4t Encephalog Balance Assessment System (Mon4t LLC, Binyamina, Israel) (Karlinsky et al., 2022; Tchelet et al., 2019). Additionally, several patient reported outcomes were included to understand the impact of this intervention on balance confidence through the Activities Specific Balance Confidence (ABC) Scale (Powell & Myers, 1995), fatigue via the 49-item Fatigue Severity (FSS-49) Scale (Krupp et al., 1988), quality of life and perceptions of disease related function through the Neurological Quality of Life (NeuroQofL) Scale (Cella et al., 2011), Patient Reported Measure for Ataxia (PROM-Ataxia) (Schmahmann et al., 2021), and the Patient Global Impression of Change Score (PGIC) (Rampakakis et al., 2015). To understand effects of this intervention on motor adaptation, we explored the use of a remotely delivered visuomotor adaptation task (Tsay et al., 2022). Primary outcomes of motor and cognitive function were administered over a HIPAA-secure telehealth platform (Zoom for Healthcare Inc, San Jose, CA, USA) at baseline 1 (Week 0), baseline 2 (Week 8), mid-intervention (week 12), and post-intervention (Week 16). All participants were required to have a care-partner present at baseline (Week 0), however for remaining assessments and intervention sessions the study team advised participants about the need for a care-partner based on fall risk from initial balance test scores (e.g., TUG <14 sec, stance in positions for Static Posturography >10 sec, CCAS scale raw score <80/120, and clinical judgement). All patient reported outcomes were administered via secure REDCap survey, at all four assessment points, except for the PGIC score and the post-intervention acceptability questionnaire, which were measured at post-intervention (Week 16). Intervention Once randomized, all participants underwent an 8-week control period to establish a baseline and evaluate performance effects on any of the outcome measures. During this time, participants were asked to continue with their typical activities of daily living, and not to engage in any novel forms of exercise. After the control period, participants proceeded with 8-weeks of 1-on-1 remote intervention sessions delivered on Zoom Healthcare by a licensed physical therapist (CM) twice weekly, for 1 hour (16 sessions). The first half (25-minutes) of each session consisted of high intensity interval training for the HIGH-BT group, or low intensity exercise for the LOW-BT group, with a 5-minute transition period (Roig et al., 2016; R. Thomas et al., 2016), prior to the second half which consisted of 30-minutes of intensive balance training for both groups. Exercises were tailored to the participant based on ataxia severity and functional presentation (Barbuto et al., 2020), and the physical therapist (CM) provided individualized cueing to ensure safe, proper body mechanics with exercise. All sessions were recorded for intervention fidelity and adhered to guidelines for practice (Macpherson, 2024; Macpherson et al., 2023). Details of the exercise protocol for both HIGH-BT and LOW-BT groups are shown in Table 1, and case exemplars for the HIGH-BT and LOW-BT group are provided in Appendix B that detail exercise progressive exercise programs for a PwSCA who participated in this study. Heart rate intensity for both groups was individualized, and determined by either the Karvonen Formula (Camarda et al., 2008) or the Brawner Formula (Mendes, 2021) for participants who had medications that interacted with heart rate function. Heart rate was monitored digitally during each session from Fitbit Charge HR 5 and perceived effort from subjective ratings of perceived exertion ratings on the Borg CR10 Scale. Each were obtained from participants at 3–5 minute intervals, as well as during post-intervention recovery at minutes 1, 2 and 5 (Macpherson, 2024). Both groups underwent the same balance training program. The balance training protocol was based in part on previously used training protocols for PwSCA and other neurodegenerative conditions (Callesen et al., 2018; Gunn et al., 2015; Sparrow et al., 2016). Fitting with the Challenge Point Hypothesis, the balance training exercises focused on the skill, the learner, and the difficulty of the tasks to be learned (Guadagnoli & Lee, 2004). To optimize motor learning, balance training exercises were individualized (Barbuto et al., 2020; Statton et al., 2018), variable, progressive, and included active participation in problem solving. Guided by Gentile’s Taxonomy of Tasks (Huxham et al., 2001), vision, proprioception, and vestibular senses were integrated within a range of tasks challenging proactive, reactive and steady state balance (Gill-Body et al., 2021; Keller & Bastian, 2014; Pérez-Avila et al., 2004; Velázquez-Pérez et al., 2019). Statistical Analysis Sample size was determined a priori from pilot data of the modified Scale for Assessment and Rating of Ataxia scale (Macpherson et al., 2023) where power was set to 0.8, α = 0.05 on IBM SPSS Statistics Version 29 (IBM SPSS Inc, Chicago, IL, USA), and a sample of 20 participants with SCA was estimated for this pilot randomized controlled trial study. Data analysis was performed in IBM SPSS Statistics Version 29 (IBM SPSS Inc, Chicago, IL, USA), while data visualizations were performed in R (R Core Group, Vienna, Austria) (Lin Pedersen, T, n.d.; R Core Team, (2021); Revelle, W. 2024; Singmann, H. et al., 2024; Wickham, H. et al., 2019, 2023). Metrics of feasibility and demographic data for all participants were summarized using descriptive statistics. Primary clinical outcome measures of SARAHome, CCAS scale raw score, and PROM-Ataxia were log transformed to achieve normality. Patient reported secondary outcome measures were normally distributed, while digital recordings of functional mobility and balance were log transformed to achieve normality. Except for the PGIC score, all quantitative outcome measures were analyzed with both descriptive statistics and a repeated-measures analysis of variance using the effect of group (HIGH-BT, or LOW-BT) and time (Baseline Week 0, Baseline Week 8, Mid-Intervention Week 12). All primary and secondary outcome measures met Mauchly’s test for the assumption of sphericity (p > .05), except for the SARAHome, where Mauchly’s test for the assumption of sphericity was violated (χ2(5) = .33, p = .006), and therefore degrees of freedom were corrected using the Greenhouse-Geisser correction (ɛ = .65). Post-Intervention (Week 16) data were evaluated with an alpha set at p < .05. Post-hoc tests used Bonferroni test for the CCAS scale raw score, and the FSS-49. The PGIC score was analyzed with descriptive statistics for Week 16. Qualitative data from the post-intervention survey was divided into analysis of Likert questions or yes/no responses, and open-ended questions. Responses to Likert questions and yes/no responses were analyzed with descriptive statistics. Responses to open ended questions were inductively coded into themes (Meissner et al., n.d.). Themes were determined by iteratively breaking down the data into smaller groups then creating codes that linked themes with interview questions. Themes were organized in a hierarchical coding frame and assigned numerical values for analysis. Any discrepancies were addressed by team discussion. Results Descriptive and Demographic Data Participant characteristics are in Table 2. Most participants had SCA type 6 (n=9), followed by SCA type 3 (n=8), SCA type 2 (n=2), and SCA type 1 (n=1). Participants had a mean (standard deviation) disease duration of 4.87(3.29) years and mean(SD) SARAHome score of 8.71(2.68) points, suggesting motor function was in early disease stages. Participants had a mean(SD) CCAS scale raw score of 95.68(10.21), and CCAS scale total score of 8.11(1.59), suggesting probable cerebellar cognitive affective syndrome. Feasibility Metrics Recruitment and Retention Participants were recruited from November 16, 2022, through September 30, 2023. Data collection was completed by March 4, 2024. Retention rate was 90% (n=18). Attrition rate per group was 10%, as the HIGH-BT group had one participant medically withdrawn from the study (n=9 for analysis), and the LOW-BT group had one participant self-withdraw (n=9 for analysis) (see Figure 1). Acceptability Acceptability criteria defined a priori was met by both groups (see Appendix 1). The post-intervention questionnaire revealed most participants preferred the option of telehealth over in-person sessions (83%, n=15). Analysis of Likert questions revealed intervention satisfaction was rated as mean(SD) 4.6(0.3) while therapist interaction was rated 5.0(0.0) (Likert scale 1–5; 1 = strongly disagree, 3 = neutral, and 5 = strongly agree). Participants also noted that the intervention resulted in improvements in walking or balance post-intervention (mean(SD) 4.5(0.6)), provided information regarding the management of motor symptoms (4.6(0.5)), and instilled the importance of disease-specific exercises (4.7(0.6)). Qualitative analysis revealed that telehealth sessions were convenient, reduced time, energy, and cost involved in travelling to in-person sessions. Some participants recommended having an option for hybrid care as an idea for future trials. Primary reasons for participating in the study were to 1) maintain functional independence, 2) reduce disease severity, and 3) be proactive with research efforts. Participants expressed greatest intervention benefits from therapist interaction/support, development of a new exercise program, and the intervention building confidence for mobility. Alternatively, for aspects of the intervention that were deemed not enjoyable, participants mentioned 1) challenging motor and cognitive assessments; 2) the use of FitBit as being difficult to readily and consistently monitor heart rate throughout the intervention (e.g., difficulty navigating device prompts, delays in obtaining heart rate, etc.); and 3) the need for a longer intervention period (beyond 8 weeks) to see greater functional improvements. Adverse Events There was one adverse event of low back pain that occurred outside of the intervention (work-related) to a participant in the HIGH-BT group. The participant was cleared by a medical doctor to continue, however, the low back pain affected intervention tolerability. The intervention was modified to reduce high to moderate intensity of exercise and integrate core training as well as coaching on body mechanics for the remainder of the study (session 6/16 onward). The participant’s back pain had resolved by the end of the trial. There were no serious adverse events during the trial. This met definitions of safety defined a priori for this pilot randomized controlled trial. Cardiovascular Tolerability Throughout the 8-week intervention, the LOW-BT group consistently demonstrated cardiovascular tolerability, with maintenance of heart rate reserve <40%, and rating of perceived exertion between 1–3 on the Borg CR10 scale across all sessions. In contrast, the HIGH-BT group showed reduced physiological markers of cardiovascular tolerability (e.g., heart rate reserve), particularly at high over moderate intensities despite reporting rating of perceived exertion within acceptable ranges at both levels (see Table 3). This discrepancy highlights a potential disconnect between physiological and self-reported measures of exercise intensity, especially as exercise intensity increases. Balance Tolerability The HIGH-BT group consistently showed decreased tolerability to engage in balance training after aerobic exercise in comparison to the LOW-BT group. While the LOW-BT group completed mean (SD) of 10.0(2.2) balance exercises per session, the HIGH-BT group only completed 8.5(2.4) largely due to self-reported fatigue as evidenced by request of frequent rest breaks, difficulty concentrating on tasks (e.g., further explanation required, task break down). Further, LOW-BT group progressed challenge of balance exercises on 83.4(14.4) percent of sessions, while the HIGH-BT group only progressed balance exercises on 69.55(14.5) percent of sessions. Preliminary Efficacy of Primary Outcomes SARAHome The SARAHome scores showed no main effect for group or time, and no interaction effects (p>.05) on repeated measures analysis of variance (see Table 4). Descriptive plots, and interaction plots for the SARAHome over time are displayed in Figure 2. CCAS Scale Raw Score There was an interaction between group and time over the 16-week intervention F(3, 48) = 5.16, p = .04. There was no effect of group p > .05, however there was a significant effect of time F(3, 48) = 19.92, p < .001. Post-hoc comparisons of time using a Bonferroni correction revealed significant differences between baseline Week 0 and mid-intervention Week 12 (mean difference 8.17 points, 95% CI [4.08, 12.62], p < .001), as well as post-intervention week 16 (mean difference 9.44 points, 95% CI [5.52, 13.37], p < .001). Additionally, there were significant differences between baseline Week 8 and post-intervention Week 16 (mean difference 7.28 points, 95% CI [2.40, 12.15], p < .01) (See Table 4, Figure 3). PROM-Ataxia There was no interaction between group and time over the 16-week intervention (p > .05). There were no main effects of group or time (p > .05) (see Table 4). Descriptive plots, and interaction plots for the PROM-Ataxia over time are displayed in Figure 4. Preliminary Efficacy of Secondary Outcomes Patient Reported Outcomes The Fatigue Severity Scale showed no main effect for group or time (p > .05), however there was an interaction effect between group and time, (F (2, 32) = 4.64, p = .018, partial η2 = .19). Post hoc pairwise comparisons used a Bonferroni correction and indicated Fatigue Severity Scale scores showed no significant differences across time points in either group (p > .05). There were no main effects for time or group (p > .05), and there were no interaction effects for the ABC scale, Fatigue Visual Analogue Scale, and NeuroQofL scale (p > .05), (see Table 5, Figures 5 A–C). Digital Measures of Functional Mobility and Balance For the TUG test, there was a main effect of time F(1, 3) = 4.36, p = .02, however no effect of group, and no interaction effects (p > .05). There were no main effects for time or group (p > .05), and there were no interaction effects for the 30 second sit to stand test (p > .05), or measures of digital posturography in any stance position (Neutral, Feet together eyes open, Feet together eyes closed, Tandem) (p > .05). Results for digital measures of functional mobility and balance are in Table 6, and Appendix C. Overall Impression of Change The median Patient Global Impression of Change scores between the HIGH-BT and LOW-BT groups were identical with a median of 6 (IQR = [5.0, 6.0]), where a score of 6 signifies that post-intervention, participants perceived their condition to be “better” with a “definite improvement that has made a real and worthwhile difference.” There were no differences between groups (U = 35.5, p = .96, r = .22). Effect Estimates Across Outcome Measures Figure 6 displays the mean difference between groups across outcome measures used in this pilot-Randomized Controlled Trial. Confidence intervals showed inclusion of zero across measures which suggests that the observed differences may be due to chance. Discussion This pilot randomized controlled trial assessed feasibility and explored motor and cognitive outcomes after an 8-week remotely delivered intervention of high-intensity exercise prior to balance training versus low-intensity exercise prior to balance training in PwSCA. The intervention was safe and feasible, with low participant attrition, high intervention adherence and acceptance in both groups. Although our results show similar post-intervention improvements on primary endpoints and at least moderate improvements on the Patient Global Impression of Change score for both groups, the LOW-BT group showed higher tolerability for intensive balance training, greater changes in global cognitive function on the CCAS scale, and improvements in fatigue on the Fatigue Severity Scale in comparison to the HIGH-BT group at post-intervention (p < .01). The use of digital outcomes (e.g., digital posturography) in this study was thought to improve sensitivity of detecting subtle changes in score on standard clinical measures (e.g., SARAHome). However, no observable differences were found across groups. The use of digital assessments in PwSCA may not be sensitive to change during an 8-week exercise trial, and instead, may be better served to detect changes longitudinally. Overall, these findings reveal an emerging story of fatigue, exercise intensity and tolerability with unique challenges to capture subtle changes in function after this remotely delivered exercise trial. Telehealth provides access to medical services within a person’s home environment, therefore reducing common barriers seen within a clinical environment (e.g., mobility impairments, time, transportation, financial burdens, care-partner availability, etc.) (Barbuto et al., 2020; Ben-Pazi et al., 2018; Chirra et al., 2019; De Marchi et al., 2021). However, telehealth has its own barriers for use, and so, considerations should be made for access to technology, technological literacy, impact of cognition, functional independence and fall risk, environmental constraints, validity and reliability of remotely delivered outcomes, as well as the availability of a care-partner for assessment and intervention delivery (Ben-Pazi et al., 2018; Chirra et al., 2019; Macpherson, 2024). In this study, we employed care-partner training, an emergency protocol, cut-offs for cognitive function, and fall risk pertaining to the involvement of a care-partner during assessments and/or intervention sessions. For safe assessment and intervention delivery, supervision was provided by a care-partner (if a participant was deemed fall risk), and a licensed physical therapist delivered the intervention. From our efforts to promote safety, there were no adverse events for assessment delivery, and one adverse event that occurred outside the intervention session (e.g., recurrence of low back pain for a participant while at work) that resolved by incorporating body mechanics and core stability to remaining sessions. Despite high levels of intervention acceptance from both groups, participants in the HIGH-BT group showed less intervention tolerability than their intervention counterparts (LOW-BT). For aerobic exercise, as intensity rose, the HIGH-BT group showed greater difficulty maintaining target heart rate and rating of perceived exertion. For balance training, although both groups were matched for time and task-intensity, group differences revealed the HIGH-BT group showed an appreciable decrease in their capacity to engage in balance training after aerobic exercise. Specifically, participants in the HIGH-BT group requested more frequent breaks, had difficulty concentrating on complex balance tasks (e.g., required task breakdown, additional explanations) and completed less balance exercises per session than those in the LOW-BT group. Exercise tolerability may be limited at high intensities for PwSCA, and these higher intensities may impair further engagement in intensive forms of multi-modal exercise. Although higher intensity aerobic exercise is often recommended in people with neurodegenerative conditions (e.g., vigorous exercise at >75% HR max), current findings suggest that PwSCA may face challenges sustaining these workloads. These observations highlight the importance of prescribing exercise intensity to match individual capabilities, particularly when programs require sustained physiological effort. Fatigue is a common and disabling symptom for PwSCA (Brusse et al., 2011; Lai et al., 2024). Previous studies have reported that PwSCA who reported fatigue on the Fatigue Severity Scale had increased ataxia severity, mood disorders (e.g., depression), and decreased quality of life throughout their disease course. (Lai et al., 2024) Although awareness for fatigue in PwSCA is gaining traction, research into its underlying neural mechanisms remains in stages of early development (Casamento-Moran et al., 2023). Emerging literature suggests that decreased cerebellar excitability leads to decreased perceptions of physical fatigue and worsened motor control (Casamento-Moran et al., 2023). which impact initial metabolic responses to acute forms of exercise in PwSCA (Barbuto et al., 2020; Lai et al., 2024; Macpherson, 2024). These factors were accounted for in this pilot randomized controlled trial with a ramped exercise protocol that used individualized target heart rate zones and ratings of perceived exertion. Balance training programs were also tailored to the individual based on ataxia severity and functional impairments (Barbuto et al., 2020; Lai et al., 2024). However, despite these efforts our findings suggested that although exercise may play a role in mitigating fatigue at lower intensities, decreased perceptions of physical fatigue may have contributed to over-work at higher intensities of and served as a rate limiting factor for exercise tolerability. As proposed in work by Casamento-Moran et al., (2023), fatigue and performance-related processes may compete for the same cerebellar resources, suggesting a need to better integrate fatigue awareness and mitigation strategies when engaging in higher intensities of exercise, or with intensive multimodal training programs (Casamento-Moran et al., 2023) In the past decade, greater emphasis has been placed upon increasing intensity across exercise type to promote neuroplasticity (Hortobágyi et al., 2021; Miyai, 2012; Rogge et al., 2018; Sparrow et al., 2016). However, a growing body of evidence is showing that perhaps the highest intensities of exercise are not necessary to induce the greatest gains in function and neuroplasticity (Hugues et al., 2022; Macpherson, 2024; Statton et al., 2015; Steib et al., 2018; Wanner et al., 2020; Youssef et al., 2024; Zoladz et al., 2014). The premise of the current study stemmed from motor priming literature, which often uses high intensity aerobic exercise prior to intensive task specific training (e.g., balance training) to elicit changes in motor learning and behavior (T. A. Moriarty et al., 2019; Sivaramakrishnan et al., 2022; Stoykov & Madhavan, 2015). However, individual differences, disease severity, decreased physical capacity, and presence of fatigue are important considerations for clinicians and researchers as they may not only contribute to the development of secondary comorbidities, but also an inability to safely tolerate higher forms of exercise intensities. Emerging literature is revealing that effects of motor priming can not only be achieved with moderate intensities of aerobic exercise (T. A. Moriarty et al., 2019) in people with Parkinson disease (Steib et al., 2018) and adult control populations (T. Moriarty et al., 2022), but that moderate and high intensities may play different roles for skill acquisition and retention (T. Moriarty et al., 2022; Statton et al., 2015; Swarbrick et al., 2020). As PwSCA in our study had greater tolerability of moderate intensity aerobic exercise compared to high intensity aerobic exercise, future work should explore moderate intensity aerobic exercise prior to balance training as a form of motor priming. Limitations Our sample included adults with early to middle stages of ataxia severity (Scale for Assessment and Rating of Ataxia scale 8–25/40), with known impairments in gait, balance, and functional mobility. Digital outcomes may be better served among those in prodromal or early disease states (Scale for Assessment and Rating of Ataxia scale <8/40), while changes in later stages may be better assessed with use of clinical measures such as the Scale for Assessment and Rating of Ataxia. In addition, the sample size for this pilot randomized controlled trial was underpowered to determine preliminary intervention efficacy. Although sample size estimates were informed a priori from our pilot trial (Macpherson, 2024), our post-hoc sensitivity analysis revealed that our results were significantly smaller than expected, and that we did not have at least 80% power to detect effects without Type II error. Conclusion This study supports the feasibility of a telehealth-delivered exercise intervention for PwSCA. Low intensity exercise prior to progressive balance training was feasible and may improve disease-specific outcomes and fatigue in PwSCA, while high intensity exercise appeared to influence fatigue, thus limiting intervention tolerability. Future research should explore the use of low versus moderate intensity exercise prior to intensive balance training and aim to optimize exercise prescription and mitigate fatigue in PwSCA.