Detection of <12 µVRMS extracellular action potential and local field potential by optimum design of a single pixel electrolyte-Oxide-MOSFET interface in CMOS 28 nm

Microelectrode-Arrays (MEAs) allow monitoring thousands of neurons/mm2 by sensing: extracellular Action Potentials and (in-vivo) Local Field Potentials. MEAs arrange several recording sites (or pixels) in a spatial grid, planarly and capacitively coupled with in-vitro cell cultures and/or integrated in electrocorticography grids. This paper focuses on Electrolyte-Oxide MOS Field-Effect-Transistors (EOMOSFET) MEAs for cell-level recording and presents a complete model of the neuron-electronics junction that reduces to a single electrical scheme all the biological (the neuron) and physical layers (the electrolyte, the Diffuse/Helmoltz capacitances, the oxide and the MOS transistor) composing the interface. This allows to predict the noise power coming from biological environment (electrolyte bath) and to optimize all electrical parameters with the main aim to minimize the final sensing Noise Figure and thus enhance the acquisition Signal-to-Noise-Ratio. Frequency domain simulations from the proposed model demonstrates that there is an optimum design point for all parameters involved in the building EOMOSFET pixel that allows to perform &gt;9 dB Signal-to-Noise-Ratio at &lt;12 µVRMS extracellular neuro-potentials power at the electrode node. This will finally enable high-resolution recording of ultra-weak neuro-potentials signals flowing by the electrolyte cleft that have not been never explored adopting planar capacitive coupling interfaces.