A study of pancreatic β-cell coupling for improved glucose sensing

This paper presents a bio-inspired methodology for improving glucose sensor performance for control of diabetes through modelling of the beta-cells of the pancreas. Specifically we utilise the noise shaping capability of gap junction coupling which is capable of synchronising β-cell behaviour to a varying glucose sensor input and reduce random noise. We show that through gap junction coupling of up to 8 cells we can improve sensor performance through a reduction in standard deviation of a glucose sensor signal by 74%. Additionally we present simulations to show how gap junction coupling of silicon β-cells can improve implementation in CMOS through reduction in variation due to mismatch in devices. These improvements will allow more accurate glucose sensing for use in a bio-inspired artificial pancreas.