Antoine PIROG soutiendra sa thèse intitulée "An embedded system for the multiparametric analysis of biological signals:application to the pancreatic biosensor of insulin demand", le 1er décembre 2017, dans l'Amphi JP. DOM du Laboratoire IMS.

Extracellular recording of electrical activity is a widespread technique in neurosciences, but only recently has it been applied to pancreatic islets and beta cells. The ease of use of Microelectrode Arrays (MEAs) has opened new perspectives for the electrophysiology of pancreatic cells, including screening methods for clinics and biosensor approaches for the artificial pancreas.
This thesis is a contribution to the design and characterization of a hybrid biosensor composed of pancreatic cells cultured on an MEA and dedicated processing electronics. This device is the product of multi-disciplinary projects conducted at IMS (Bioelectronics group), partnered with CBMN (Cell biology and Biosensors team), both at the University of Bordeaux.
Projects also involved the endocrinology service of university hospitals in Bordeaux, Montpellier, Grenoble, and Geneva.
The covered projects include:
- ISLET-CHIP (French ANR 2013-PRTS-0017), investigating a method of evaluating the quality of a preparation prior to its transplantation in type-I diabetic patients.
- BIODIA (EU FEDER), characterizing islet electrical response to glucose, hormone, and drug stimuli for screening, cell differentiation, and closed-loop approaches.
- DIAGLYC (French regional grant 2017 1R30 226), investigating the use of the hybrid biosensor as an artificial pancreas front-end sensor.
The thesis tackles the biosensor in both its electronic and biological aspects, its integration in applicative experimental setups, and experimental results. It also addresses the modeling of a closed regulation loop for type-I diabetic patients.
First, the electronic processing platform is described. It is a custom board performing multichannel acquisition and digital signal processing. It is built around an FPGA (Field
Programmable Gate Array) that makes its processing architecture versatile and evolutive. It is capable of measuring, displaying and storing multichannel data. Computation was optimized for low-processing latencies compatible with closed-loop configurations. Both its processing
algorithms and architecture are detailed.
Then, experimental results using this system and its algorithms are shown to illustrate islet response to glucose, drug, and hormone stimuli. Islet activity is quantified by analyzing Action Potentials (APs), and more importantly Slow Potentials (SPs), a novel electrical signature exclusively measured on pancreatic islets. These measurements in both steady state and dynamic regime help characterize the biosensor response, but also shed light on the endogenous
algorithms of islets of Langerhans.
Finally, approaches for integrating the biosensor in an artificial pancreas are investigated. The measured glucose and hormone responses are modeled and simulated in a full-body glucose-insulin system. This concept is novel in that the sensor in charge of measuring insulin demand in the body is not only sensitive to glucose, but also to blood hormones.

Key words
Biological signals, FPGA, Diabetes, Bioelectronics,