Individual trajectories
Patients with similar initial symptoms may follow very different recovery paths.
Neurorehabilitation · Digital Twins · Neuromorphic Modeling
Patient-specific neural digital twins for post-stroke rehabilitation
NeuroTwin helps neurorehabilitation clinics and research partners explore dynamic neural models that evolve with each patient, supporting recovery monitoring, motor function prediction, and early risk identification.
Research-stage clinical decision support. Physicians remain in full control of all medical decisions.
Stroke recovery is dynamic. Current tools are mostly static.
Post-stroke rehabilitation varies significantly from patient to patient. Clinicians often need to interpret fragmented assessments, population-level risk factors, and delayed outcome signals while making decisions under uncertainty.
Limited prediction
Current clinical tools often provide static snapshots rather than dynamic, patient-specific forecasts.
Late risk visibility
Potential complications or deviations may become visible only after valuable rehabilitation time has passed.
A continuously updated neural digital twin for each patient
NeuroTwin is developing a research-driven platform that uses anonymized clinical data to construct patient-specific neural models. As new patient data becomes available, the model can be updated to help clinicians and researchers better understand recovery dynamics and potential risks.
Rehabilitation progress assessment
Rehabilitation progress assessment
Motor function recovery prediction
Motor function recovery prediction
Early identification of potential complications
Early identification of potential complications
Simulation-oriented research into intervention effects
Simulation-oriented research into intervention effects
NeuroTwin is designed to support clinical reasoning and research validation. It is not an autonomous diagnosis or treatment recommendation system.
From clinical data to patient-specific risk insight
Anonymized clinical data
A partner clinic provides regular anonymized data from post-stroke rehabilitation patients.
Patient-specific neural modeling
NeuroTwin builds a personalized neural digital twin focused first on motor function and spinal cord-related modeling.
Periodic model updates
The model is updated as new clinical observations, assessments, and rehabilitation data become available.
Trajectory assessment
The system compares expected and observed rehabilitation dynamics to identify meaningful deviations.
Risk highlighting
Clinicians receive research-stage decision support signals that may help prioritize attention and further assessment.
Starting with motor recovery after stroke
NeuroTwin begins with motor recovery because it provides a focused and clinically meaningful entry point for validation. The initial modeling direction emphasizes spinal cord and motor function dynamics, with a roadmap toward motor cortex modules and broader neural-system modeling.
Focused validation
Start with a measurable rehabilitation domain before expanding to broader neurological conditions.
Clinically relevant outcomes
Prioritize motor function recovery, patient trajectory monitoring, and risk alerts for rehabilitation teams.
Incremental expansion
Move from spinal cord and motor recovery modeling toward motor cortex modules and broader brain regions over time.
Built for research collaboration with neurorehabilitation clinics
We are inviting clinical partners to validate NeuroTwin through retrospective and prospective research pilots, co-publications, and grant collaborations.
Research pilots
Analyze anonymized retrospective or prospective rehabilitation data and validate patient-specific modeling approaches.
Co-publications
Develop joint scientific papers and new clinical-research methodologies around post-stroke recovery modeling.
Grant collaboration
Explore EU and UK funding opportunities for clinical validation, neuromorphic modeling, and rehabilitation research.
Neuromorphic modeling for scalable neural simulation
NeuroTwin combines clinical modeling with computational neuroscience and neuromorphic computing approaches. The long-term technology direction includes hardware-software co-design and energy-efficient neural simulation using next-generation computing substrates.
Recent benchmarking work across neuro-computational systems shows that event-driven and in-memory approaches are among the most promising routes toward energy-efficient neural computation.
Why organic memristors matter
Organic memristive devices are a research direction for synapse-like, adaptive, energy-efficient neuromorphic systems. Their conductance can depend on prior activity, which makes them relevant for modeling plasticity, learning, and memory-like processes in artificial neural systems.
Research-led team
NeuroTwin is built by a research-driven team combining computational neuroscience, neuromorphic systems, clinical collaboration, and technology commercialization.
Victor
Research · Organic memristive devices · Neuromorphic systems
Maxim
Research · Neural modeling · Computational neuroscience
Oleg
CBDO & CTO · Product strategy · Clinical partnerships · Technology commercialization
Updates
May 2026
NeuroTwin prepares clinical collaboration materials for neurorehabilitation centers.
We are preparing research pilot materials for clinics interested in patient-specific neural digital twins for rehabilitation monitoring.
LinkedIn post coming soonApril 2026
Energy benchmarking study published in IEEE Access.
A new paper compares energy consumption across classical AI hardware, digital spiking processors, memristive devices, and biological reference ranges.
Read the articleOctober 2020
Comparative analysis of memristive devices for neuromorphic applications.
A review of organic and inorganic memristive devices for neuromorphic applications was published in BioNanoScience.
Read the articleInterested in validating patient-specific neural digital twins in rehabilitation?
We are inviting neurorehabilitation clinics and research partners to explore retrospective and prospective collaboration opportunities.