Optimal Learning Rules for Discrete Synapses
2008

Optimal Learning Rules for Discrete Synapses

publication Evidence: moderate

Author Information

Author(s): Barrett Adam B., van Rossum M. C. W.

Primary Institution: Institute for Adaptive and Neural Computation, University of Edinburgh, Edinburgh, United Kingdom

Hypothesis

How does the storage capacity of discrete, bounded synapses compare to that of unbounded, continuous synapses?

Conclusion

Discrete synapses can have a storage capacity similar to unbounded synapses under certain conditions.

Supporting Evidence

  • The study calculates the storage capacity of discrete synapses in terms of Shannon information.
  • It finds that below a critical number of synapses per neuron, storage is similar to that of unbounded synapses.
  • The research suggests that discrete synapses do not necessarily have lower storage capacity.

Takeaway

This study looks at how brain connections store memories and finds that even with limited options, they can remember just as well as if they had unlimited choices.

Methodology

The study used analytic mathematics, computer programming, and literature review to analyze the storage capacity of discrete synapses.

Limitations

The analysis was restricted to about 8 states per synapse, and the implications for biological systems are not fully explored.

Digital Object Identifier (DOI)

10.1371/journal.pcbi.1000230

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication