Skip to main content
European Commission logo
français français
CORDIS - Résultats de la recherche de l’UE
CORDIS
CORDIS Web 30th anniversary CORDIS Web 30th anniversary
Contenu archivé le 2024-05-27

Information Processing by Natural Neural Networks

Livrables

Our market analysis clearly shows that we are still in the very beginning of the age of neural networks, and that many technological hurdles have to be overcome before a broader application range can be addressed, namely applications in the field of prosthetics and brain/machine interfaces. Nevertheless, many problems and issues mentioned by the interviewed academic and industrial institutions were already addressed in the INPRO project, which has to offer a variety of valuable solutions to problems that slowed down past progress in this field. The potential impact of our system in the field of pure and applied neuroscience is significant. This INPRO system provides solutions for unmet needs in the neuroscience community, prominently by its potential and ability to experimentally validate various old and new theories. The paucity of past experimental results stems from the fact that it is virtually impossible to easily perform high-resolution multi-site recordings in vitro. High spatiotemporal resolution recordings are necessary to study the evolution of neural network activity. Indeed, it is the transient nature of neural connections within the brain, known as synaptic plasticity, that is believed to underlie the poorly understood phenomena of learning, and hence memory. In the field of applied neuroscience, this device could be used to develop new algorithms with respect to information processing that may then be used by the IT and computing industry.
Breakthrough advances in the field of neural information processing may only be possible using a CMOS-based MEA system. This is because high spatiotemporal resolution recordings are necessary to completely explore the complex nature of neural information processing. The microelectronic system design approach in INPRO is based on a modular design. Each electrode of the microelectrode array (MEA) has an associated unit of circuitry including a fully differential band-pass filter for immediate signal conditioning, and a buffer for stimulation and mode selection. Implementing filters and buffers at each electrode offers important advantages in comparison to other CMOS MEAs: - The signal is amplified and filtered in close vicinity of the electrodes, which makes the design less sensitive to noise and distortion picked up along connection lines; - A stimulation buffer makes the stimulation signal independent of the number of activated electrodes; - The high-pass filter removes offset and slow drifts of the electrophysiological signals and, thereby allows for immediate signal amplification; - The low-pass filter limits the noise bandwidth and acts as an anti-aliasing filter for succeeding A/D-conversion. FPGA-based electronics interface to PC data aquisition and analysis using standard USB 2.0 protocols. Neurons from primary sources as well as from progenitor cell lines can be cultured on the INPRO chip for several months. Surface modification by biochemical adhesion and differentiation factors allow the definition of growth areas. The chip can be housed in a microfluidic chamber for nutrient exchange and application of neuroactive drugs. In summary, the INPRO platform has been tested and its functionality been validated using standard approaches and recipes in cell culturing and electrophysiology - demonstrating that upon commercialisation it may facilitate high-resolution investigation of learning paradigms by in vitro long-term multichannel electrophysiology without the need of changing well established protocols.

Recherche de données OpenAIRE...

Une erreur s’est produite lors de la recherche de données OpenAIRE

Aucun résultat disponible