Spike-Based Neuromorphic Model of Spasticity for Generation of Affected Neural Activity.

Area of Science:

• Neuromorphic engineering
• Computational neuroscience
• Rehabilitation robotics

Background:

• Spasticity is a common motor disorder affecting muscle control and movement.
• Current models fail to capture all spasticity origins, particularly peripheral causes like hyperreflexia.
• Effective spasticity management requires understanding its diverse origins and informing therapeutic strategies.

Purpose of the Study:

• To develop a novel computational, spike-based neuromorphic model of spasticity, named NEUSPA.
• To incorporate peripheral factors, such as additive (ADD) and multiplicative (MUL) inputs, into the model to simulate velocity-dependent EMG responses.
• To validate the NEUSPA model using existing literature and patient data, and apply it to simulated real-world scenarios.

Main Methods:

• Developed NEUSPA, a neuromorphic model extending beyond basic spinal loops to include ADD and MUL inputs.
• Validated the model against classic experiments and clinical data from post-stroke patients.
• Simulated patient-relevant scenarios, including finger-pressing on a deformable object.

Main Results:

• NEUSPA successfully simulated spastic EMG responses driven by hyperreflexia from additional inputs.
• EMG onsets showed higher sensitivity to ADD inputs (r²=0.96) than MUL inputs (r²=0.92).
• Simulations indicated spasticity increased finger-pressing duration by approximately 16% compared to non-impaired conditions.

Conclusions:

• The NEUSPA model effectively synthesizes abnormal physiological data related to spasticity.
• The model's ability to incorporate peripheral causes offers a more comprehensive approach to spasticity modeling.
• NEUSPA shows potential for aiding decision-making and machine learning applications in neurorehabilitation.