MaDEleNA

Acronym: MaDEleNA

Typology: Large Project

Full title:

Developing and Studying novel intelligent nanoMaterials and Devices towards Adaptive Electronics and Neuroscience Applications

Duration: 01/09/2013 to 31/08/2017

Total costs: Euro 2,360,176.00

PAT contribution: Euro 2,360,176.00

Coordinating Entity:

Institute for Materials for Electronics and Magnetism of the National Research Council - IMEM-CNR

Project Manager: Prof. Salvatore Iannotta

Other participants:

• Bruno Kessler Foundation (Resp . Leandro Lorenzelli)
• University of Trento (Resp. Paolo Macchi)
• IBF-CNR Institute of Biophysics (Resp . Carlo Musio)
• Institute of Photonics and Nanotechnology IFN-CNR (Resp . Alessandro Chiasera)

Brief description of the consortium

The project partners offer consolidated knowledge at the best international levels in materials science (their synthesis and study), device science and electronics (design and realisation of electronic architectures, memristors), biophysics and biology (biocompatibility, neuronal tissues), all of which are fundamental sectors for achieving the final objectives.

OSRthematic area: Materials science: micro- nano- inorganic and hybrid technologies

Project objectives

Madelena proposes a highly innovative approach to study and propose solutions to relevant questions of modern science and technology, in fields as seemingly distant as electronics and neuroscience but with objectives of high productive, social and scientific impact.

Electronics is evolving towards nanoscale miniaturisation to produce computers with functions and performance increasingly similar to those of the human brain; however, the latter is endowed with patterns and an intrinsic adaptability that cannot be attacked by current electronic architectures. There is also a strong demand for models that reproduce precisely the behaviour of the human brain, in order to explore new horizons and open up new frontiers for research.

Madelena is moving precisely in this area of interface between electronics and neuroscience, with the dual objective of implementing new neuro-bio-inspired computing systems and creating hardware models (devices and systems) that mimic the human brain This will lead to new technologies and new methodological approaches, creating a centre of excellence in these fields in Trento.

State of the art and improvements that will be introduced by the project

The electronic architecture of a computer is based on a two-dimensional network with a high density of components, non-modifiable and sequential logic, with different zones for memory and computation. The brain, on the other hand, develops on three-dimensional networks, with parallel logic and is able to process data and learn in the same physical location. Considerable efforts have been made to develop neural networks based on traditional hardware, but as long as learning and computation are developed by software based on these architectures, satisfactory results are unlikely to be achieved.

What Madelena wants to achieve is to overcome this dichotomy between the seat of memory and the seat of computation, by developing 'neuromorphic' electronics that mimic neuronal systems, based on 'memristors', innovative adaptive electronic elements that change their state with the history of the events they have 'experienced' and that can interface with neuronal tissues in order to understand their mechanisms.

Work Organisation/Implementation

The five partners have expertise in materials science (IMEM, UniTN, IFN), electronic (FBK) and neuromorphic architecture design (IMEM), biological and neuronal systems (UniTN, IBF). The project is structured along five main lines relating to the synthesis and study of materials with memresistive properties, the realisation of logic devices and electronic architectures based on memristors, the development of neuromorphic networks, and the creation of hybrid interfaces formed by memristors and neuronal tissues. The research activities, whose time evolution follows a precise schedule, are monitored by specific bodies inside and outside the project and are supported by two external entrepreneurial partners, Biomat srl and ST Italia Spa.

Expected impact

The project aims first and foremost to promote scientific leadership and increase innovation and competitiveness in an emerging technology by creating a critical mass of state-of-the-art researchers with multidisciplinary skills at local level. This will lay the foundations for a technology transfer from scientific sources to local and national entrepreneurs, with careful management of intellectual property. From this point of view, the development of new architectures based on memristors for both memories and adaptive networks will be a real turning point for electronics. In addition, the potential of the original hybrid approach of memristor/neuron biointerfaces will have spin-offs that may go far beyond the local area.

Expected results

Will be developed:

• Deterministic electronic architectures but based on memristors, i.e. where logical operations take place with an intrinsic learning capability
• stochastic neuromorphic networks, based on a random distribution of individual memristors, to reproduce the network of synaptic connections and study learning in the brain
• hybrid biointerfaces between memristors and neuronal tissues for the bidirectional exchange of signals, overcoming current limitations due to the use of biocompatible electrodes and standard electronics.

Many of the efforts are aimed at achieving high reliability of the individual memristor elements, as well as of the architectures and biointerfaces, which is crucial for determining the success of the project. The hybrid approach is certainly the most original aspect and the greatest challenge of the project, but the possible future developments for understanding the human brain and creating adaptive architectures are disruptive in the fields of neuroscience and electronics technology.

Keywords:

Biomorphic adaptive electronics, nanomaterials, nanosystems, neuroscience, interfaces to the nervous system