Should Hands Be Restricted When Measuring Able-Bodied Participants to Evaluate Machine Learning Controlled Prosthetic...

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Neuroscience

Background:

• Able-bodied participants are often used to test prosthetic hand control, but their performance differs from individuals with upper limb absence (ULA).
• Restricting able-bodied participants' limb movements may reduce this performance gap, but its effect on electromyogram (EMG) features is unclear.

Purpose of the Study:

• To investigate EMG signal differences between unaffected and affected arms in individuals with ULA.
• To determine if restricting the unaffected limb's hand and wrist affects EMG separability and consistency.
• To assess the influence of arm posture on these effects.

Main Methods:

• Compared EMG signal separability and consistency between the unaffected and affected arms of 14 participants with acquired transradial limb loss.
• Evaluated two conditions: unaffected hand/wrist restricted and unrestricted.
• Examined the impact of arm posture (arm down, in front, or up) on EMG features.

Main Results:

• The unaffected limb demonstrated greater EMG signal separation than the affected limb.
• Restricting the unaffected hand and wrist reduced EMG signal separability, particularly when the arm was in the down position.
• Arm posture influenced the effects of limb restriction on EMG separability.

Conclusions:

• Limb restriction is a practical method to enhance the comparability of EMG data between able-bodied individuals and those with ULA.
• Future studies evaluating machine learning-based prosthetic hand control should consider restricting the hand and wrist movements of able-bodied participants.