Skip to content

Perceptual-motor Interaction to Improve Bimanual Coordination After Stroke

Perceptual-motor Interaction to Improve Bimanual Coordination After Stroke

Status
Completed
Phases
Early Phase 1
Study type
Interventional
Source
ClinicalTrials.gov
Registry ID
NCT03755076
Acronym
Bimanual
Enrollment
79
Registered
2018-11-27
Start date
2019-04-19
Completion date
2024-01-03
Last updated
2025-08-14

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Stroke

Brief summary

Significant difficulty in incorporating the weaker arm in daily activities after stroke is, in part, driven by difficulty in engaging both arms interactively in a coordinated manner. The current study aims to determine the nature of bimanual coordination deficits after stroke and takes initial steps to test a novel theory-driven approach to improve interactive bimanual coordination in patients with stroke. This project will advance stroke rehabilitation by identifying novel, scientifically-based strategies to improve the engagement of the weaker arm in coordinated and interactive bimanual actions of daily life, thus improving quality of life in individuals after stroke.

Detailed description

Specific Aim: Determine the immediate effects of perceptual cues on bimanual coordination in stroke survivors. Rationale: Bimanual performance after stroke is impaired and demonstrate two principal deficits: (a) impaired spatiotemporal coordination between the two arms and (b) reduced engagement of the paretic arm. Perceptual cuing often overrides motor execution constraints to influence bimanual coordination in healthy individuals, and help improve performance. Our aim is to determine the immediate effects of specific perceptual cues on bimanual coordination in common-goal bimanual action in participants with stroke. The perceptual cue conditions are (a) indiscriminate: no specific feedback, (b) altered gain: paretic arm gain was reduced such that the paretic arm had to contribute more compared to the nonparetic arm, (c) coordination cue provided as feedback of the horizontal tilt of the common goal, (d) Dual: altered gain of the paretic arm combined with the coordination cue. Our working hypothesis is that time-lag feedback will constrain the two arms to move simultaneously, shorten the time-lag and increase CCr. Further, greater movement weighting applied to the weaker arm to increase the engagement of the weaker arm toward the bimanual action may impair bimanual coordination immediately. Overall materials and methods: The proposed experiments will follow the general pattern we have used in previous behavioral psychophysical studies in individuals with and without stroke. A controlled virtual environment that records kinematics of the arm and allows real-time movement interaction with a virtual motor task will be used. A program integrating motion capture system (Acsention Technology TrakSTAR) and MATLAB-based GUI-gaming environment allows tracking of the two arms, experimenter-control for specific manipulation such as position of the virtual brick, target gap, relative contribution of the arms and mapping of each arm movement to the movement of the virtual brick to provide real-time and post-response feedback. Participants will be seated in an adjustable chair facing a computer monitor with their trunk constrained to the chair. Their arms will be completely supported on a low friction table top and free to move in the horizontal (X-Y) plane with minimal resistance. An opaque screen will occlude participants' direct vision of their arms. Magnetic markers will be secured to their hands just proximal to the wrist joint and the position of the markers mapped to a virtual brick shown on the computer screen. The motor task is to move the virtual brick(s) to a virtual target gap(s) on the computer screen by moving both arms in 2D (X-Y) plane on the low-friction table. While real-time visual feedback of the brick is available, the brick cannot be felt haptically. From a perceptual perspective, in independent goal condition, each arm moves its own virtual brick to the target gap. In common-goal condition, a common brick was moved on the computer screen to a target window by predetermined weighting of each arm movement. Prior to beginning the experiments, participants will reach with the paretic arm in three different directions (135º, 90º and 45º relative to the horizontal) to record the maximum reaching distance for two trials in each direction. The minimum reach distance across the three directions will be used to calibrate the start and end position of the target gaps. The target window position will be placed at 90% of the maximum reach distance, oriented at 90º relative to the horizontal. Procedures: Participants will come to the laboratory for a baseline evaluation to determine eligibility to participate in the experimental protocol. During this baseline evaluation, we will perform the following tests: (1) Fugl-Meyer Examination, (2) Mini-mental scale, (3) Tests for hemineglect using a line-bisection test, (4) Western Aphasia Battery for patients with Aphasia (5) TMS safety questionnaire, (6) MRI safety questionnaire, (7) Box and Block test as well as (8) Penn Neurocognitive assessment. In the present study, we will test the effect of two different perceptual cues on bimanual coordination. Participants will be instructed to move a common virtual brick with both arms to the target windows at 90 degree position without tilting the brick within a target MT of 800 milliseconds- 1.2 seconds. Participants will complete four 60-trial blocks a pseudorandom order. Each block will consist 60 trials of a distinct perceptual cueing condition depending upon the nature of perceptual cues provided. Condition 1: Indiscriminate: This condition will be similar to the common-goal bimanual condition in Aim 1 where they will transport a common virtual block fixed in a horizontal position to three targets in pseudorandom order. The movement of the bar will be an unweighted average of the two arm movements; i.e., each arm will contribute to 50% of the virtual bar movement (50-50 weighting). Condition 2: Altered gain: This condition will be similar to Condition 1 (block fixed in horizontal position), but the arm weighting will be differential. The paretic arm will have reduced gain such that it will have to move farther to move the virtual brick with this cue than without it. Condition 3: Coordination cue: the weighting coefficients of the two arms will be equal (i.e. 50-50); and the virtual brick will tilt in the direction of the lagging arm proportional to the relative time-lag between arms. Operational definition of relative time-lag: Relative time-lag is different than absolute time lag. Relative time-lagis the time-lag between the relative timing of each arm within its trajectory. To illustrate, if the left and right arm are contributing to 70 and 30% of the brick movement, relative time-lag at mid-movement will be zero if the left and right arm have covered half of their respective trajectories. Therefore, relative time-lag is influenced by temporal as well as spatial component of the movement of each arm. Concurrent and post-response feedback about the tilt and path of the virtual brick will be provided after each trial. Condition 4: Dual cues combining altered gain with coordination cue: In addition to the coordination cue (i.e. tilt feedback about the relative time-lag), the paretic arm will have reduced gain such that it will have to move farther and faster to keep the virtual brick horizontal.

Interventions

BEHAVIORALPerceptual cuing

Participants were provided four perceptual cues (or feedback) in a pseudorandom order: (a) indiscriminate, (b) Altered gain, (c) Coordination, (d) Dual: altered gain and coordination

Sponsors

University of Pennsylvania
CollaboratorOTHER
Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD)
CollaboratorNIH
Albert Einstein Healthcare Network
Lead SponsorOTHER

Study design

Allocation
NON_RANDOMIZED
Intervention model
SINGLE_GROUP
Primary purpose
TREATMENT
Masking
NONE

Intervention model description

Participants will perform bilateral arm movements in a virtual reality environment under different perceptual conditions to identify a condition that improves coordination between arms.

Eligibility

Sex/Gender
ALL
Age
21 Years to 80 Years
Healthy volunteers
Yes

Inclusion criteria

1. Clinical diagnosis of unilateral stroke 2. ability to reach along a diagonal direction at least 80% of their arm length while fully supported on a frictionless surface and trunk constrained. 3. Mini-mental scale score \> 26, OR s score of 4 or above on auditory verbal comprehension part of the Western Aphasia Battery to ensure intact comprehension and following commands. (e) no evidence of hemispatial neglect tested by a line bisection test.

Exclusion criteria

1. bilateral stroke, 2. complete paralyses, 3. basal ganglia/cerebellar stroke, 4. pain or stiffness in upper extremity that will interfere with the task or inability to follow task instructions.

Design outcomes

Primary

MeasureTime frameDescription
Maximum Cross-correlation CoefficientAfter 20 minutes of training under each perceptual cue conditionCross-correlation between tangential velocity profiles of the two hands was used to index spatial and temporal coordination between hands. Cross-correlation measures similarities of two distinct time series as a function of the displacement of one relative to the other. Repeated correlations between the two hand velocity profiles were obtained as the velocity profile of one hand was successively lagged. The maximum cross correlation coefficient obtained gave a measure of similarity between the two profiles, indexing spatial coordination. The better the arms moved in space, the higher the score they got, called the cross-correlation coefficient (CCr). CCr values range from zero to 1; with values close to 1 means the arms were moving very well together.
Between Hands Time-lagAfter 20 minutes of training under each perceptual cue conditionTemporal coordination is about how well both arms move at the same time. Temporal coordination was quantified as the time lag at which the peak cross-correlation coefficient was obtained via cross-correlation analysis. If one arm moves a little later than the other, there's a time delay (or time lag). Shorter the delay, the better the timing between the arms. If the delay is positive, it means the weaker or less-used arm is moving after the stronger one.

Countries

United States

Participant flow

Recruitment details

Participants with stroke (n=50) and age matched controls (n=29) were recruited from Moss Rehabilitation Research Registry in Pennsylvannia, PA. Recruitment occurred between April 2019- January 2024.

Pre-assignment details

No participants were excluded from the study after enrollment.

Participants by arm

ArmCount
Neurotypical Control
Neurotypical healthy age-matched controls
29
Stroke Survivors
Stroke survivors with mild-to-moderate motor impairments
50
Total79

Baseline characteristics

CharacteristicNeurotypical ControlTotalStroke Survivors
Age, Continuous63.82 years
STANDARD_DEVIATION 10.91
61.73 years
STANDARD_DEVIATION 9.77
60.52 years
STANDARD_DEVIATION 8.94
Race and Ethnicity Not Collected0 Participants
Region of Enrollment
United States
29 Participants79 Participants50 Participants
Sex: Female, Male
Female
14 Participants34 Participants20 Participants
Sex: Female, Male
Male
15 Participants45 Participants30 Participants
Upper Extremity Fugl-Meyer Score; Motor impairment66 units on a scale
STANDARD_DEVIATION 0
56.82 units on a scale
STANDARD_DEVIATION 12.77
51.5 units on a scale
STANDARD_DEVIATION 13.45

Adverse events

Event typeEG000
affected / at risk
EG001
affected / at risk
deaths
Total, all-cause mortality
0 / 290 / 50
other
Total, other adverse events
0 / 290 / 50
serious
Total, serious adverse events
0 / 290 / 50

Outcome results

Primary

Between Hands Time-lag

Temporal coordination is about how well both arms move at the same time. Temporal coordination was quantified as the time lag at which the peak cross-correlation coefficient was obtained via cross-correlation analysis. If one arm moves a little later than the other, there's a time delay (or time lag). Shorter the delay, the better the timing between the arms. If the delay is positive, it means the weaker or less-used arm is moving after the stronger one.

Time frame: After 20 minutes of training under each perceptual cue condition

ArmMeasureGroupValue (MEAN)Dispersion
Neurotypical ControlBetween Hands Time-lagCondition 1: Indiscriminate: Baseline19.57 millisecondsStandard Deviation 24.04
Neurotypical ControlBetween Hands Time-lagCondition 1: Indiscriminate: End of Practice0.35 millisecondsStandard Deviation 0.95
Neurotypical ControlBetween Hands Time-lagCondition 2: Altered gain: Baseline15.92 millisecondsStandard Deviation 18.31
Neurotypical ControlBetween Hands Time-lagCondition 2: Altered gain: End of Practice0.42 millisecondsStandard Deviation 1.25
Neurotypical ControlBetween Hands Time-lagCondition 3: Coordination: Baseline28.64 millisecondsStandard Deviation 27.06
Neurotypical ControlBetween Hands Time-lagCondition 3: Coordination: End of Practice0.21 millisecondsStandard Deviation 1.13
Neurotypical ControlBetween Hands Time-lagCondition 4: Dual: Baseline107.85 millisecondsStandard Deviation 121.29
Neurotypical ControlBetween Hands Time-lagCondition 4: Dual: End of Practice0.071 millisecondsStandard Deviation 0.377
Stroke SurvivorsBetween Hands Time-lagCondition 4: Dual: End of Practice26.48 millisecondsStandard Deviation 76.15
Stroke SurvivorsBetween Hands Time-lagCondition 1: Indiscriminate: Baseline72.75 millisecondsStandard Deviation 149.08
Stroke SurvivorsBetween Hands Time-lagCondition 3: Coordination: Baseline49.15 millisecondsStandard Deviation 48.81
Stroke SurvivorsBetween Hands Time-lagCondition 1: Indiscriminate: End of Practice1.55 millisecondsStandard Deviation 7.59
Stroke SurvivorsBetween Hands Time-lagCondition 4: Dual: Baseline152.44 millisecondsStandard Deviation 188.3
Stroke SurvivorsBetween Hands Time-lagCondition 2: Altered gain: Baseline51.51 millisecondsStandard Deviation 64.83
Stroke SurvivorsBetween Hands Time-lagCondition 3: Coordination: End of Practice8.44 millisecondsStandard Deviation 19.49
Stroke SurvivorsBetween Hands Time-lagCondition 2: Altered gain: End of Practice0.04 millisecondsStandard Deviation 0.29
p-value: <0.001ANOVA
Primary

Maximum Cross-correlation Coefficient

Cross-correlation between tangential velocity profiles of the two hands was used to index spatial and temporal coordination between hands. Cross-correlation measures similarities of two distinct time series as a function of the displacement of one relative to the other. Repeated correlations between the two hand velocity profiles were obtained as the velocity profile of one hand was successively lagged. The maximum cross correlation coefficient obtained gave a measure of similarity between the two profiles, indexing spatial coordination. The better the arms moved in space, the higher the score they got, called the cross-correlation coefficient (CCr). CCr values range from zero to 1; with values close to 1 means the arms were moving very well together.

Time frame: After 20 minutes of training under each perceptual cue condition

ArmMeasureGroupValue (MEAN)Dispersion
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 1: Indiscriminate-Baseline0.989 cross-correlation coefficientStandard Deviation 0.014
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 1: Indiscriminate-End of practice0.995 cross-correlation coefficientStandard Deviation 0.006
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 2: Altered gain- Baseline0.985 cross-correlation coefficientStandard Deviation 0.016
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 2: Altered gain: End of practice0.998 cross-correlation coefficientStandard Deviation 0.003
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 3: Coordination: Baseline0.957 cross-correlation coefficientStandard Deviation 0.068
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 3: Coordination: End of practice0.998 cross-correlation coefficientStandard Deviation 0.002
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 4: Dual: Baseline0.873 cross-correlation coefficientStandard Deviation 0.184
Neurotypical ControlMaximum Cross-correlation CoefficientCondition 4: Dual: End of practice0.997 cross-correlation coefficientStandard Deviation 0.003
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 4: Dual: End of practice0.990 cross-correlation coefficientStandard Deviation 0.018
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 1: Indiscriminate-Baseline0.945 cross-correlation coefficientStandard Deviation 0.064
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 3: Coordination: Baseline0.933 cross-correlation coefficientStandard Deviation 0.065
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 1: Indiscriminate-End of practice0.996 cross-correlation coefficientStandard Deviation 0.005
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 4: Dual: Baseline0.861 cross-correlation coefficientStandard Deviation 0.097
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 2: Altered gain- Baseline0.951 cross-correlation coefficientStandard Deviation 0.05
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 3: Coordination: End of practice0.995 cross-correlation coefficientStandard Deviation 0.006
Stroke SurvivorsMaximum Cross-correlation CoefficientCondition 2: Altered gain: End of practice0.997 cross-correlation coefficientStandard Deviation 0.003
p-value: <0.001ANOVA

Source: ClinicalTrials.gov · Data processed: Feb 4, 2026