The objectives of the research and technological development (RTD) work packages within BRAIN BOW are divided into 6 work packages.
WP 1. Engineered interconnected neuronal assemblies
GOAL: Construct and characterize innovative brain models by using 2D/3D patterning techniques.
The aim of this WP is to build-up neuronal networks with a topological architecture of increasing anatomical complexity. In order to measure their electrophysiological activity we will use both electrical extracellular recordings and optical imaging. More specifically, we will make use of Micro-Electrode Arrays (MEAs) and calcium imaging methodology.
WP2: Integrated experimental set-up and multichannel data analysis
GOAL: development of tools for large-scale electrophysiology and for data analysis of multichannel signals.
The project will require the development of a flexible experimental set-up allowing the multi-site recording and stimulation of different neuronal preparations. Automated dedicated micro fluidics and environmental controls will be used to nourish the cells, and to deliver drugs and/or chemical agents at controlled concentrations. A software interface will be developed to control chemicals delivering and environmental parameters (e.g. temperature, pH). Moreover, starting from the sw platform SpyCode, developed in the context of a collaboration between IIT and UNIGE, an additional software package for off-line signal processing will be realized, in order to manage data of different nature (i.e. electrophysiological signals, calcium imaging/VSD signals).
WP3: Neurodynamics and information transmission in interconnected neuronal assemblies
GOAL: Extract the input/output function of our neuronal system.
Within this WP we will work in an open loop modality, by characterizing the input/output function of the neuronal system, either in standard condition or in presence of a lesion. The neuronal system will be constituted by the hierarchical modular structures developed in WP1 (2D and 3D cultures) and the in vitro whole brain of a small mammal, (IWB). Characterizing the input/output function of neuronal circuits require the possibility to record and stimulate the activity of neurons. Depending on the different neuronal system under study (in-vitro 2D/3D networks, ex-vivo whole brain), we will use different techniques.
WP4: Large-scale sw computational models to substitute the function of a biological sub-network
GOAL: Development of large-scale neuronal networks made up of point neurons with the aim at building hybrid networks with real neurons by means of a real time paradigm.
To model the dynamics of large-scale networks, two different spatial configurations (2D and 3D) will be taken into account, following the same pathway of the experimental activity. To describe the dynamics of each neuron, reduced models will be considered: a good trade-off between computational efficiency and capability of reproducing rich firing patterns exhibited by actual neurons is obtained by the Izhikevich (IZ) and integrate-and-fire (IF) neuron models.
WP5: A neuromorphic chip mediating the communication within cell assemblies in a pathological neuronal system
GOAL: Implementation of large-scale computational models into one hardware system to bi- directionally communicate with a biological neural network
The main goal of this WP is to design a hardware system which communicates with the biological neural network. To reach that point, we plan to reproduce the artificial neural network developed in WP4 into the hardware system. Finally, this hardware platform will communicate bi-directionally with the biological element (i.e. either an engineered network or a portion of an intact brain).
WP6: Realization of bi-directional neuroprostheses
GOAL: Perform experiments with the developed system to test the sw and hw replacement of neuronal connections and cell-assembly.
This WP represents the core of BRAIN-BOW. We will realize a bi-directional communication between the biological and the artificial element. First, we will connect the biological network to the ‘software’ artificial model developed in WP4, in a second phase to the ‘hardware’ computational model implemented in WP5. For each of the two experimental designs, we will take advantage from the results of WP3 and WP4 to extract and convey information in case of pathology (i.e. a focalized lesion at the level of the brain, which disconnect one assembly or interrupts the communication between two connected assemblies).