Dimitri Henniquau's thesis

"Design of a functional interface for communication between artificial and biological neurons
for applications in the field of neuroscience".

Thesis defence: 14 December at 10 a.m.
IEMN Amphitheatre - Central Laboratory (LCI) - Villeneuve d'Ascq

Jury :

Cécile DELACOUR, Research Fellow, Institut Néel, University of Grenoble, Rapporteur
Sylvie RENAUD, University Professor, IMS Laboratory, University of Bordeaux, Rapporteur
Serge BERNARD, Research Director, LIRMM, University of Montpellier, Examiner
Marc PANANCEAU, Senior Lecturer, NeuroPSI, Université Paris-Saclay, Examiner
Jean-Pierre VILCOT, Research Director, IEMN, University of Lille, Examiner
Virginie HOEL, University Professor, IEMN, University of Lille, Thesis supervisor
Christel VANBESIEN-MAILLIOT, Senior Lecturer, IEMN, University of Lille, Thesis supervisor
Alexis VLANDAS, Research Fellow, IEMN, University of Lille, Thesis supervisor

Summary:

Neuromorphic engineering is a new and fast-growing field that draws on skills in electronics, mathematics, computer science and biomorphic engineering to produce artificial neural networks capable of processing information in the same way as the human brain. In this way, neuromorphic systems not only offer more effective and efficient solutions than current information processing technologies, but also open up the possibility of developing novel therapeutic strategies for pathological brain dysfunctions.
The Circuits Systèmes Applications des Micro-ondes (CSAM) group at the Institut d'Electronique, de Microélectronique et de Nanotechnologies (IEMN), where this thesis work was carried out, has contributed to the emergence of these neuromorphic systems by developing a complete toolbox of artificial neurons and synapses. In order to integrate neuromorphic engineering into the treatment of pathological neuronal dysfunctions, artificial neurons and living neurons need to be interfaced in order to ensure real communication between these different components. In this context, and using the innovative tools developed by the CSAM group, the aim of this thesis work was to design and produce a functional interface enabling a bidirectional communication loop to be established between artificial neurons and living neurons. The artificial neurons developed by the CSAM group use CMOS technology and are capable of emitting biomimetic electrical signals. The living neurons are derived from differentiated PC12 cells.
The first stage of this work consisted in modelling and simulating this interface between artificial and living neurons; the second part of the thesis was dedicated to the fabrication and characterisation of neurobiohybrid interfaces, as well as to the growth and characterisation of living neurons, before studying their ability to communicate with artificial neurons. A neuronal membrane model representing a living neuron interfaced with a planar metal electrode was developed. Using this model, it was shown that it is possible to stimulate living neurons using the biomimetic signals from the artificial neuron model while maintaining low excitation voltages. The use of low excitation voltages would improve the energy efficiency of neurobiohybrid systems incorporating artificial neurons and reduce the risk of damaging living tissue. Next, the neurobiohybrid used to interface living neurons and artificial neurons was designed and produced. Experimental characterisation of this interface validated the approach, which consists of exciting a living neuron via a planar metal electrode. Finally, live neuronal cells derived from PC-12 cells were cultured and differentiated in the neurobiohybrids. Experimental proof of the capacity of the biomimetic electrical signals produced by the artificial neurons was thus provided using the calcium imaging technique.
In conclusion, the work presented in this manuscript clearly establishes the proof of concept of the excitation of living neurons by a biomimetic signal under our experimental conditions and thus supports the first part of the bidirectional communication loop between artificial neurons and living neurons.

Abstract:

Neuromorphic engineering is a new and rapidly growing field of study that calls upon skills in electronics, mathematics, computer science and biomorphic engineering in order to produce artificial neural networks capable of processing information in the manner of the human brain. Thus, neuromorphic systems not only offer more powerful and efficient solutions than current information processing technologies, but also allow the development of novel therapeutic strategies for pathological brain dysfunctions.
The Circuits Systems Applications of Microwaves (CSAM) group of the Institute of Electronics, Microelectronics and Nanotechnologies (IEMN) where this thesis work was carried out has contributed to the emergence of these neuromorphic systems by developing a complete toolbox of artificial neurons and synapses. In order to integrate neuromorphic engineering in the management of pathological neuronal dysfunctions, it is necessary to interface artificial neurons and living neurons in order to ensure a real communication between these different components. In this context, and using the innovative tools developed by the CSAM group, the objective of this thesis work was to design and realize a functional interface allowing to establish a bidirectional communication loop between artificial and living neurons. The artificial neurons developed by the CSAM group are made of CMOS technology and are capable of emitting biomimetic electrical signals. The living neurons are derived from differentiated PC12 cells.
A first step of this work consisted in modeling and simulating this interface between artificial and living neurons; a second part of the thesis was dedicated to the fabrication and characterization of neurobiohybrid interfaces, as well as to the growth and characterization of living neurons, before studying their capacity to communicate with artificial neurons. Thus, a neuronal membrane model representing a living neuron interfaced with a planar metal electrode was developed. The exploitation of this model allowed us to show that it is possible to stimulate living neurons using the biomimetic signals from the artificial neuron model while maintaining low excitation voltages. The use of low excitation voltages would improve the energy efficiency of neurobiohybrid systems incorporating artificial neurons and reduce the risk of damaging living tissue. Then, the neurobiohybrid allowing to interface the living neurons and the artificial neurons was designed and realized. An experimental characterization of this interface allowed to validate the approach consisting in exciting a living neuron through a planar metallic electrode. Finally, live neuronal cells derived from PC-12 cells were cultured and differentiated in the neurobiohybrids. An experimental proof of the capacity of the biomimetic electrical signals produced by the artificial neurons could thus be provided by the calcium imaging technique.
In conclusion, the work presented in this manuscript clearly establishes the proof of concept of the excitation of living neurons by a biomimetic signal under our experimental conditions and thus supports the first part of the bidirectional communication loop between artificial and living neurons.