Computational neurorehabilitation: modeling plasticity and learning to predict recovery

Area of Science:

• Neuroscience
• Computational Biology
• Rehabilitation Medicine

Background:

• Limited computational models exist for sensorimotor rehabilitation mechanisms.
• Progress in computational neuroscience and neuroscience necessitates new modeling approaches.
• Robotics and wearable sensors provide data for developing and testing computational models.

Purpose of the Study:

• Introduce Computational Neurorehabilitation, a new field focused on modeling neural processes for improved motor recovery.
• Discuss the integration of brain plasticity and learning systems into rehabilitation models.
• Advance understanding and prediction of individual recovery from neurologic impairment.

Main Methods:

• Review key aspects of neural plasticity and motor learning relevant to computational models.
• Discuss the relationship between computational neurorehabilitation models and current prognostic modeling.
• Analyze early computational neurorehabilitation models, particularly for upper extremity stroke recovery.

Main Results:

• Early computational models, even simple ones, offer novel insights for future research.
• Emerging technologies like robotics and wearable sensors enable feasible model development and testing.
• Computational models can integrate clinical imaging and real-world activity data.

Conclusions:

• Computational neurorehabilitation is crucial for a fundamental understanding of neurologic recovery.
• Mechanistic models informed by imaging and real-world data will drive future rehabilitation strategies.
• This field promises to enhance prediction and personalization of rehabilitation therapies.