Reduced Order Modelling of a Neuron-Electrode Interface

The function of the human brain and its interaction with implanted neural interfaces is highly complex and still not fully understood. Numerical models are an efficient way to study these interactions. However, the high complexity of numerical models leads to long computational times and memory requirements. In this work, we introduce model order reduction as a method to reduce both. We consider a simplified finite element model of a neuron-electrode interface and build the reduced order model using Krylov-subspace based model order reduction. We compare the electric potential distribution as a function of time from reduced and full order model and quantify a speed up in computational time. Finally, we set up an initial system level simulation, by connecting the reduced order model to an ideal current source.