Periodic Reporting for period 2 - CROSSBRAIN (Distributed and federated cross-modality actuation through advanced nanomaterials and neuromorphic learning)
Période du rapport: 2023-11-01 au 2025-04-30
CROSSBRAIN addresses the pressing need for minimally invasive, adaptive brain stimulation tools capable of tackling neurological disorders such as epilepsy. In an era marked by rising neurological disease burden and limited treatment options, CROSSBRAIN proposes a transformative approach by merging neuromorphic computing, advanced nanomaterials, and wireless microdevices. The project aims to develop µBots—tiny, implantable devices capable of sensing and modulating brain activity through multiple physical modalities (electrical, mechanical, optical, thermal) with unprecedented spatial and temporal precision. These µBots function autonomously in a federated network, enabling closed-loop, patient-specific neuromodulation. The integration of AI-based neuromorphic control and biocompatible nanomaterials supports long-term implantation with minimal immune response. CROSSBRAIN's technological innovations are expected to advance fundamental neuroscience, support new therapeutic strategies, and contribute to EU leadership in brain-computer interfaces and neurotechnology regulation. Social sciences and ethics are embedded to assess human relevance, regulatory acceptability, and societal impact. The project aligns with the EU’s strategic goals in health innovation, neurotechnology, and responsible AI.
During the second reporting period, CROSSBRAIN advanced significantly toward its goal of realising a minimally invasive, intelligent neuromodulation platform. A key achievement was the development and integration of functional nanomaterials—multiferroic nanoparticles, gold nanorod heaters, and conductive polymer hydrogels—that enable mechanical, thermal, and ionic stimulation of neural tissue. These were successfully interfaced with custom microelectronic components, including a second-generation low-power analogue front-end chip and magnetoelectric antennas for wireless operation. The µBots reached a functional prototype stage, demonstrating autonomous power harvesting, recording, and stimulation capabilities in vitro. A modular neuromorphic controller chip was designed and fabricated, with integration of ultra-low-power memristive synapses underway. This controller enables adaptive, closed-loop brain signal decoding and stimulation based on real-time analysis of neural activity.
Biocompatibility of key components—such as dummy µBots, nanomaterial coatings, and hydrogels—was tested in neuronal cultures and organotypic brain slices, showing high cell viability and normal neuronal activity.Furthermore the AI-driven closed-loop system was successfully tested in vitro to detect and suppress epileptiform activity, confirming the project's technical feasibility and clinical potential.
Biocompatibility of key components—such as dummy µBots, nanomaterial coatings, and hydrogels—was tested in neuronal cultures and organotypic brain slices, showing high cell viability and normal neuronal activity.Furthermore the AI-driven closed-loop system was successfully tested in vitro to detect and suppress epileptiform activity, confirming the project's technical feasibility and clinical potential.
CROSSBRAIN has delivered substantial advances toward next-generation neuromodulation, setting important foundations for wireless, intelligent brain interfaces. The project successfully demonstrated key building blocks of a µBot platform—including wireless power harvesting via magnetoelectric antennas, low-power neural recording with a custom analogue front-end chip, and the integration of nanomaterials enabling electrical, mechanical, and thermal stimulation. While the µBots have not yet reached the final miniaturised 100×100×100 µm³ form, early prototypes combining selected functionalities were tested in vitro and in brain slices.
The consortium also designed and fabricated a neuromorphic controller chip intended to process neural signals in real time, using spiking neural networks and ferroelectric memristive synapses. Initial in vitro experiments confirmed that the AI-based system can detect and modulate seizure-like discharges in neuronal cultures, demonstrating proof of concept for closed-loop intervention.
To realise clinical impact, further integration, in vivo testing, and system-level miniaturisation are needed. Scaling up manufacturing, validating long-term biocompatibility, and navigating regulatory and standardisation frameworks will be essential. The project has already contributed to shaping standards for wireless nano-implants and initiated IP protection, creating a strong basis for future commercialisation and translational development.
The consortium also designed and fabricated a neuromorphic controller chip intended to process neural signals in real time, using spiking neural networks and ferroelectric memristive synapses. Initial in vitro experiments confirmed that the AI-based system can detect and modulate seizure-like discharges in neuronal cultures, demonstrating proof of concept for closed-loop intervention.
To realise clinical impact, further integration, in vivo testing, and system-level miniaturisation are needed. Scaling up manufacturing, validating long-term biocompatibility, and navigating regulatory and standardisation frameworks will be essential. The project has already contributed to shaping standards for wireless nano-implants and initiated IP protection, creating a strong basis for future commercialisation and translational development.