Clinical modeling layer
Fast models designed to process rehabilitation data, track patient trajectories, and support research-stage risk analysis.
Technology
Dynamic neural models that evolve with the patient
NeuroTwin combines clinical modeling, computational neuroscience, and neuromorphic computing to develop patient-specific neural digital twins for rehabilitation monitoring and risk prediction.
Hybrid architecture: clinical models plus deep neuroscience simulation
The NeuroTwin architecture should be implemented as a staged system: practical clinical models for early research pilots, combined with deeper neuroscience simulation layers for offline research and long-term technology development.
Neural simulation layer
Patient-specific dynamic neural models focused first on motor function and spinal cord-related rehabilitation dynamics.
Neuromorphic acceleration layer
Long-term hardware-software co-design direction for efficient neural simulations using event-driven and in-memory computing principles.
Validation layer
Model outputs are compared with real rehabilitation outcomes to assess reliability, limitations, and clinical relevance.
Core workflow
- 1. Receive anonymized clinical data from a partner clinic.
- 2. Construct a patient-specific neural digital twin.
- 3. Update the model when new rehabilitation data becomes available.
- 4. Assess rehabilitation trajectory and deviations.
- 5. Highlight research-stage risk signals for clinician review.
Why start with motor recovery
NeuroTwin starts with motor recovery after stroke because it is a clinically meaningful and focused rehabilitation domain. The initial research direction emphasizes spinal cord and motor function modeling as a practical first step before expanding toward motor cortex modules and broader neural-system models.
- Closer connection to measurable rehabilitation outcomes
- Focused scope for early clinical validation
- Stronger alignment with motor rehabilitation workflows
- Incremental path toward broader brain digital twin capabilities
Development roadmap
Motor recovery digital twin
Initial focus on post-stroke motor recovery and spinal cord-related modeling.
Motor cortex modules
Expansion toward cortical contributors to motor rehabilitation dynamics.
Broader neural system modeling
Integration of additional brain regions and more complex neurological conditions.
Intervention simulation research
Simulation-oriented studies of rehabilitation strategies and potential therapeutic interventions such as electrical stimulation.
Toward energy-efficient neural simulation
Neural digital twins require scalable computation. Recent research comparing neuro-computational systems shows that dense digital AI workloads can consume substantially more energy per synaptic event than event-driven spiking and in-memory approaches. NeuroTwin’s long-term technology direction therefore emphasizes neuromorphic computing and memory-compute co-location as promising routes for sustainable neural simulation.
Critical Analysis of Energy Consumption in Neuro-Computational Systems
Interested in validating patient-specific neural digital twins in rehabilitation?
We are inviting neurorehabilitation clinics and research partners to explore retrospective and prospective collaboration opportunities.