Towards neuro-silicon interface using reconfigurable dynamic clamping

Dynamic clamp emerges as an important apparatus to study the intrinsic neuronal properties through close-loop interactions between models and biological neurons. Modelling large-scale neuronal networks in software will result in significant computational delay that becomes a bottleneck to apply dynamic clamp for more complicated systems. In this paper, we present a real-time dynamic clamping system based on field programmable gate arrays (FPGAs) to accelerate the necessary computations. It also provides a flexible platform to reconfigure various model parameters and topologies. Realtime neuronal and synaptic models were implemented in FPGA, and interconnected with the stomatograstric ganglion (STG) nervous system to exemplify the real-time dynamics. Results show that our method can be effectively configured to mimic various biological neural networks and is two orders of magnitude faster than software approach using desktop computer.