In-phase bilateral arm movements facilitate smoother movement in acute stroke: a kinematic and fMRI study

Introduction: Bilateral upper limb training (BT) is a potential stroke rehabilitation intervention.
Comparing unilateral and bilateral BT modes using kinematic assessments and neuroimaging allows
understanding of the cortical and behavioural mechanisms of action underlying BT. This study aimed to
compare the effects of unilateral, bilateral in-phase and anti-phase modes of a functional task on
paretic arm kinematics alongside functional magnetic resonance imaging (fMRI) in acute stroke.
Methods: 13 acute stroke individuals underwent 3D motion capture of the unilateral, bilateral inphase and anti-phase modes of a functional grasp task. Intralimb kinematics was assessed through
movement time, peak velocity, movement smoothness and movement directness. fMRI scanning was
undertaken using continuous wrist flexion-extension.
Results: Bilateral in-phase grasp task was performed with significantly smoother paretic arm movement compared to unilateral task mode (Z ¼ 2.510, p ¼ 0.012). Bilateral in-phase grasp task was
performed with significantly smoother paretic arm movement (Z ¼ 2.971, p ¼ 0.003) and more
direct movement (Z ¼ 2.761, p ¼ 0.006) compared to bilateral anti-phase task mode. Bilateral inphase wrist flexion-extension involved greater neural activity in numerous brain regions (inferior
frontal gyrus, superior temporal gyrus, middle temporal gyrus and inferior temporal gyrus of the
lesioned hemisphere) compared to bilateral anti-phase wrist flexion-extension in most participants.
Conclusion: Paretic arm movement became smoother during bilateral in-phase grasp task compared
to unilateral and bilateral anti-phase task modes. Therapists might consider using bilateral in-phase
functional UL tasks on stroke survivors with jerky movements. fMRI demonstrated significant differences in activation/deactivation of brain regions between bilateral in-phase and anti-phase wrist
flexion-extension in acute stroke, providing evidence of differential neural networks