Dual-Task Balance Training in People With Multiple Sclerosis: a Randomized Feasibility Study

BACKGROUND People with multiple sclerosis (PwMS) often show reduced walking and balance performance when they perform a cognitive task at the same time (dual-task, DT). The neural basis of this motor-cognitive interaction is still unclear, and there is little evidence on how rehabilitation affects its behavioral and neural correlates. Recent meta-analytic data suggest that balance training including DT exercises can improve cognitive functions in PwMS.

DESIGN This is a single-center, randomized, controlled, feasibility study. Thirty PwMS, recruited mainly among participants of a previous cross-sectional phase of the same project, are randomly allocated 1:1 to the intervention group or to a no-treatment control group. The allocation sequence is computer-generated with permuted blocks of 6 by a team member not involved in recruitment or assessment, and is concealed until the baseline (T0) assessment is completed. The sample size (15 per group) is in line with recommendations for pilot and feasibility studies and is not based on a formal power calculation.

INTERVENTION The intervention group receives 14 in-hospital sessions (60 min, 2 sessions/week, about 7 weeks) at the Neurorehabilitation Unit of AOUI Verona. Each session includes warm-up (5 min), static balance exercises (15 min), dynamic balance exercises (15 min), dual-task training combining dynamic balance exercises with cognitive tasks (20 min), and cool-down (5 min).

The allocation sequence will be computer-generated before the start of the enrollment, using permuted-block of 6.

Experimental details - Rehabilitation protocol (intervention group):

i) Warm-up (5 min): joint mobilization and gentle stretching of trunk and lower limb muscles; ii) Main training block (50 min); iii) Cool-down (5 min): slow walking and breathing exercises.

1.1. Main Training Block

The main 50-minute training block will be structured as follows (with moderate flexibility in terms of phase duration according to each participant's performance and fatigue):

1. Static balance exercises (15 min)
2. Dynamic balance exercises (15 min)
3. Dual-task training: dynamic balance exercises + cognitive tasks (20 min)

Exercises are designed to target different resources required for balance postural stability - according to Horak's framework - that are relevant in the context of Multiple Sclerosis.

1. Static Balance Exercises (15 min)

   Static balance exercises will be performed in standing position or will include postural transition tasks (e.g. sit-to stand), and will target limits of stability, anticipatory postural adjustments and multisensory integration. Exercises in static standing will require the participant to maintain the standing posture and balance with progressively challenging constraints regarding a) the base of support, b) sensory input, and c) concurrent upper limb movements.

   Complexity progression will be performed manipulating these three elements as follows (according to participant's functional level):

   1. Base of support: wide symmetrical → narrow symmetrical → tandem → single-limb stance.

   2. Sensory information: different combinations of visual, somatosensory, and vestibular information manipulation will be explored. Progression will be performed within each sensory modality as follows:

      • Visual input: eyes open → eyes closed;
      • Somatosensory input: standing on a firm surface → standing on a compliant surface;
      • Vestibular input: head movement: no → yes. First, each sensory modality will be manipulated individually; then different combinations will be explored according to each participant's functional level.
   3. Concurrent upper limb movements: bilateral arm raising → forward and lateral reaching → ball handling.

   Progression criterion: three consecutive repetitions performed successfully (i.e. without the need to reach for support or change the base of support) will be needed to progress to the next difficulty level. The target duration of each condition will be 30 s.

   Progression modality: first, each intervention factor (base of support, sensory information, and concurrent upper limb movement) will be manipulated individually, to identify the appropriate level of difficulty. Then, different combinations will be performed (e.g. standing with narrow symmetrical base of support, on a firm surface, with eyes closed and bilateral arm raising), adapting the difficulty level of each factor to the participant's functional level and following the same criterion for difficulty progression.

   Sit-to-stand will then be performed. Progression will be based on the height of the chair (standard → low), the use of armrests (yes → no), and simultaneous upper limb raising (no → yes).

2. Dynamic Balance Exercises (15 min)

   Dynamic balance exercises will target the modification of the base of support, walking in different directions, and multisensory integration during locomotion, and will include march in place, walking in different directions, obstacles courses and stair climbing and descending. Conditions involving walking with eyes closed will be performed in close proximity to the therapist and with parallel bars available if needed.

   The following progressions will be followed:

   1. Base of support (during walking): standard → narrow → tandem;

   2. Walking direction: forward straight walking → slalom walking → sideways walking

      → backward walking;

   3. Obstacle height: progressively increased;
   4. Sensory information: the same progression described for static balance exercises will be applied to dynamic exercises (when feasible and safe). Progression criterion: performance of three consecutive repetitions without the need to reach for support to maintain the balance.

   Progression modality: similarly to the static exercises block, first, each intervention factor (base of support, walking direction, obstacle height, and sensory information) will be manipulated individually, to identify the appropriate level of difficulty. Then, different combinations will be performed (e.g. walking in slalom with narrow symmetrical base of support, on a compliant surface, with eyes open), adapting the difficulty level of each factor to the participant's functional level and following the criterion for difficulty progression.

3. Dual-Task Training (20 min) Dual-task training will include the simultaneous performance of a motor and a cognitive task. Motor tasks will be the same as those used in the dynamic balance block, with progression adjustment according to the dual-task performance. The cognitive task will include serial subtraction of 3s and 7s, reciting alternate alphabet letters, fluency exercises (e.g. naming all animals/cities whose names start with a given letter), and will be varied within each session. The digit span will not be included, as it is reserved as an outcome measure. Difficulty will be set aiming to maintain performance in the cognitive task approximately to 70% rate of correct responses during dual-task conditions. Difficulty progression will follow the same levels described in the dynamic exercises. Progression criterion: performance of three consecutive repetitions without the need to reach for support to maintain the balance and with a minimum of 70% rate of correct responses.

Progression modality: like in the previous two blocks, first, each intervention factor (base of support, walking direction, obstacle height, and sensory information) will be manipulated individually. Then, different combinations will be performed, adapting the difficulty level of each factor to the participant's functional level and following the criterion for difficulty progression.

ASSESSMENT (T0 and T1) A comprehensive clinical, functional, cognitive functions, quality of life assessment will be carried out includin: Motricity Index, Dynamic Gait Index, Mini Best Test, Multiple Sclerosis Functional Composite, Hospital Anxiety and Depression Scale (HADS), Fatigue Severity Scale, Activities-specific balance confidence (ABC), the Manual Ability Measure-36 (MAM-36), Brief Pain Inventory (BPI), Multiple Sclerosis Quality Of Life-54, Five Levels EQ-5D, Paced Auditory Serial Addition Task (PASAT), number of falls in the last month.assessment batter

Participants will also perform a DT paradigm with six conditions in random order: single-task TUG (5 trials), single-task 2MWT, single-task backward digit span (2-minute version and 5 x 15-s version), DT-TUG, and DT-2MWT.

Cortical activity is recorded with an 8-channel gel-free wearable EEG headset (Mindtooth Touch, prefrontal and parietal channels) and gait with three inertial measurement units (Movella DOT, lower back and shanks). NASA-TLX and a fatigue numeric rating scale are collected after each condition. Clinical scales are administered by an assessor blinded to group allocation. Participants who interrupt the training for more than one week, or who have a relapse between T0 and T1, are withdrawn.

OUTCOME MEASURES Primary endpoints (feasibility)

• Process metrics: recruitment rate, retention rate, session adherence (sessions completed out of 14).
• Resource/management metrics: assessment burden (time per session), protocol compliance, data completeness.
• Scientific feasibility: ability of the selected outcome measures to detect change, to inform sample size of a future trial.
• Patient-reported: Global Impression of Change, ABC scale, MSWS-12. Secondary endpoints (preliminary effect)
• Change (T0→T1) in motor DT costs for TUG and 2MWT, compared between intervention and no-treatment group.
• Change in IMU-based gait kinematic parameters.
• Change in EEG-based prefrontal activation and TLI during the ST motor conditions.
• Change in Clinical, functional, cognitive functions, quality of life outcomes: Motricity Index, Dynamic Gait Index, Mini Best Test, Multiple Sclerosis Functional Composite, Hospital Anxiety and Depression Scale (HADS), Fatigue Severity Scale, Activities-specific balance confidence (ABC), the Manual Ability Measure-36 (MAM-36), Brief Pain Inventory (BPI), Multiple Sclerosis Quality Of Life-54, Five Levels EQ-5D, Paced Auditory Serial Addition Task (PASAT), number of falls in the last month.

Expected outcomes

• The protocol will prove feasible in terms of recruitment, retention and adherence.
• The intervention group will show reduced prefrontal activation and mental workload (TLI) during TUG and 2MWT in the single-task condition.
• Reduced DT cost, expected primarily for TUG (and less for 2MWT, since endurance is not specifically targeted by the training).
• Preliminary effect-size estimates to power a future randomized controlled trial.

STATISTICAL ANALYSIS. Feasibility outcomes (primary outcomes) are reported descriptively as counts, proportions or means with 95% confidence intervals. Secondary outcomes are analyzed as exploratory, with no formal hypothesis testing of efficacy. Dual-task costs are computed as the normalized performance decrement in DT compared to single-task (ST). EEG features include theta and alpha band power spectral densities and the Task Load Index (frontal theta / parietal alpha). Within-group changes from T0 to T1 are tested with paired t-test or Wilcoxon signed-rank test according to data distribution. Between-group differences at T1 are estimated with repeated measure ANOVA with "group" effect and reported as adjusted mean differences with 95% confidence intervals. When ANOVA assumptions are not met, a rank-based alternative is used. Standardized effect sizes are calculated to inform the sample size of a future randomized controlled trial. Analyses follow the intention-to-treat principle, including all randomized participants with available data.