NETT · Neural Engineering Transformative Technologies

Neural Engineering Transformative Technologies

NOTE: The below is the plain text version of the publishable summary. A pdf version is attached together with image files and URLs of web presence. NETT was formed to address some key challenges in Neural Engineering through inter-related projects combining the skills of mathematicians, physicists, neuroscientists and bioengineers. It has generated results in a wide array of application-driven fields, developing techniques and technology advancements in for example novel speech recognisers, neural-inspired laser networks for information processing, brain-computer interfaces (BCI), robots with cognitive skills and neural prosthetics for enhancing or repairing sensory-motor functions. NETT’s primary objectives were to: a) Bring together researchers from a variety of backgrounds that feed into Neural Engineering in order to pool resources and tackle current challenges with an interdisciplinary team trained in core skills in complexity science. b) Expose early career researchers to the opportunities the field offers in academia and industry, training them in the methodology employed in both, with input from experienced visiting researchers and corporate research teams. c) Progress the careers and employability of the fellows via internships with industry partners and a dedicated training programme in business, academic and social skills. d) Demonstrate the interrelatedness of the various project areas throughout the consortium and to the wider scientific community, industry and general public through study groups, workshops, public lectures and web presence. NETT has employed 21 fellows who collectively made impressive advances within and beyond the seven scientific work packages. All results and outputs from NETT were achieved by the fellows under guidance from the supervisors. In Adaptive Control Methods, a dramatic improvement of the effective sampling size for stochastic processes between two noisy observation points was achieved, a reduction of one order of magnitude of the particle number needed. An Echo State Network was also trained to simulate a stochastic mechanical system, including a model of Echo State learning with precise spikes in neural networks. In Synthetic Cognition, a feed-forward two-layer spiking neural network model of the midbrain Superior Colliculus was delivered, alongside growth of primary rat neurons in an integrated Multi-Electrode Array microfluidics device and modelling of assemblies of spiking neurons to study different collective phenomena as synchronisation and communication between brain areas as well as travelling waves of electrical activity in the hippocampus. An upstream model of the auditory pathway within a Bayesian framework to model information transmission in the pathway was constructed, and a cellular automaton for neural growth in cultures and constructed bespoke numerical schemes for delayed neural field models was designed. For Human-Robot Interaction, extensions of a dynamic neural field model were developed to learn the serial order and timing of complex behavioural sequences in shared tasks, allowing it to cope with varying movement times of different effectors, as well as extending a neuro-inspired planning model previously designed and implemented in the anthropomorphic robot platform ARoS, towards developing human-like synchronous movements of two high degree of freedom robot arms. A novel dynamic field model of a combined spatial and parametric working memory capacity was developed between UMinho and UNOTT that offers new perspectives for more efficient human-robot collaboration. For Neural Inspired Information Processing, the group at UPC have carried out studies of a mesoscopic brain model in the presence of noise. A paper on cross-frequency transfer in this model has been published in Frontiers in Computational Neuroscience (the #1 most cited and #1 largest open-access publisher in the category of Neuroscience). Neural Coding has used recordings of dendritic spine signals (inputs from neurons) to deliver novel two-photon targeted patch-clamping robotic technology (which is a specific instance of optogenetic BCI technology with significant prospects for high throughput in vivo drug characterisation as well as reverse engineering neural circuits), a system for optimal sequential cellular recording in an inertial laser scanning microscope (having developed a new scanning strategy called Adaptive Spiral Scanning, with successful simulations now being validated through experiments) and with Cortexica, new theoretical/computer vision approaches for the analysis of large scale optical neural recordings. In Emergent Neurodynamics, fellows have analysed the dynamics of diluted neural networks and have produced a software package (SPIKY) for spike train analysis and used this to analyse neuronal data. This package is open source and available in Matlab. Several published papers have appeared from the CNR group about SPIKY, in both high profile Neuroscience journals and Engineering journals. For Neural Rehabilitation, EEG decoders were developed to detect gait anticipatory potentials, relevant for rehabilitation after stroke or incomplete spinal cord injuries as well as trigger events in BCI. In addition to the commonly used features, phase has been proposed as an alternative representation with better signal to noise ration and decoding performance. The method was evaluated both on healthy subjects and chronic stroke patients. Work on detecting mind wandering episodes, relevant for neuro-rehabilitation therapies that require high adherence, was carried out and an integral framework was developed for planning and evaluation of this type of study overcoming some of the limitations that have contributed to the so-called reproducibility crisis in science. This was implemented for brain decoding, particularly BCI, and is available as an open-source package on github. All the scientific contributions have been or will be published (six journal papers, four accepted and two under review, plus six contributions to conferences). The training provided by NETT has been well-received by fellows, with all events attended by industry partners who have provided group training in e.g. networking, commercialisation, academic writing, and more specialist scientific and soft-skills needed for their project work. Fellows come from a wide range of scientific and geographical backgrounds and, but integrated well and the benefits of this interaction manifested through unplanned fellow-fellow collaborations and links to new institutions across the world. Secondments for the fellows further expanded their horizons as well as providing a career boost, and many fellows that have left NETT have gone on to strong positions in academia (e.g. at Graz University of Technology, Aix-Marseille University) or industry (e.g. Apple, Simplxr, UniCredit Bank). Research highlights in NETT include: • Characterisation of the microscopic and macroscopic dynamics of sparse pulse coupled neural networks (Physical Review E & Chaos, Solitons and Fractals) • Development of SPIKY: Matlab tool for monitoring spike train synchrony (open source - http://wwwold.fi.isc.cnr.it/users/thomas.kreuz/Source-Code/SPIKY.html) and PySpike (www.pyspike.de) • Establishing the effect of transmission delays in the communication between brain areas (PLOS Computational Biology) • Development of a novel algorithm for multi-photon scanning of neural tissue (Proceedings of IEEE Chicago EMBS) • Probed scale interaction in the brain through synchronisation (Philosophical Transactions of the Royal Society B) • Development of two-photon targeted patch clamping robot (presented at SFN 2013) NETT has contributed to a new generation of multidisciplinary researchers trained in neural engineering. The methods and technologies developed through, and as a result of, NETT stand to have an effect on almost every aspect of daily human life, from increasing mobility and aiding diagnosis and treatment of disease, to transforming the way in which we interact and communicate with machines.

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз