Towards optimizing the non-invasive sensory feedback interfaces in a neural prosthetic control

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Engineering

Background:

• Non-invasive sensory feedback is crucial for prosthetic limb control.
• Somatotopic (SI) and non-somatotopic (NI) interfaces are common feedback methods.
• Differences in sensory characteristics and brain responses between SI and NI are not well understood.

Purpose of the Study:

• To investigate sensory differences between SI and NI.
• To propose an optimal encoding strategy for non-invasive sensory feedback.
• To objectively evaluate brain responses to SI and NI using EEG.

Main Methods:

• Seven amputees participated in tactile discrimination and object recognition tasks.
• Tactile sensitivity to position and intensity was compared between SI and NI.
• Electroencephalography (EEG) measured stimulus-evoked brain responses.
• A closed-loop prosthesis control system integrated feedback conditions.

Main Results:

• NI showed higher sensitivity to position, while SI was more sensitive to intensity.
• EEG data confirmed behavioral findings, with NI showing stronger alpha ERD for position discrimination and SI for intensity discrimination.
• Optimal performance (62% overall accuracy) was achieved when position feedback used NI and intensity feedback used SI.

Conclusions:

• NI and SI possess distinct sensory characteristics beneficial for different types of feedback.
• An optimized encoding strategy combining NI for position and SI for intensity enhances object recognition in amputees.
• This study provides a valuable strategy for improving non-invasive sensory feedback in prosthetic devices.