This report explores the capabilities and limitations of the SpiNNaker neuromorphic computing platform by extending the current sPyNNaker API to implement a compartmental neuron model capable of distinguishing between multiple excitatory inputs. The implemented model successfully achieves this and a novel approach to implement more complex learning rule is proposed that utilises a modulating signal to initiate learning. These novel contributions take initial exploratory steps to inform the development of next generation neuromorphic hardware and software.

This report was completed in May 2017 and only uses the state of the SpiNNaker system and sPyNNaker API that were publicly available at the time of writing.