In this study, we describe the implementation of six functional networks on the Spikey neuromorphic hardware system: A synfire chain, a balanced random network, a soft winner-take-all network, a cortical layer 2/3 attractor model, an insect antennal lobe model and a liquid state machine combined with a tempotron. This study demonstrates the flexibility of the Spikey hardware system. Moreover, these networks can serve as functional building blocks of a large neuromorphic computing system.

Neuroinformatics member and project leader Michael Schmuker contributed the insect antennal lobe model in this collaboration with the Electronic Vision(s) group at Kirchhoff-Institute for Physics, Heidelberg University.

The paper is open access and available online (preliminary version until typesetting is complete).