Presynaptic stochasticity improves energy efficiency and helps alleviate the stability-plasticity dilemma

Abstract
Data availability
Article and author information
Metrics

Abstract

When an action potential arrives at a synapse there is a large probability that no neurotransmitter is released. Surprisingly, simple computational models suggest that these synaptic failures enable information processing at lower metabolic costs. However, these models only consider information transmission at single synapses ignoring the remainder of the neural network as well as its overall computational goal. Here, we investigate how synaptic failures affect the energy efficiency of models of entire neural networks that solve a goal-driven task. We find that presynaptic stochasticity and plasticity improve energy efficiency and show that the network allocates most energy to a sparse subset of important synapses. We demonstrate that stabilising these synapses helps to alleviate the stability-plasticity dilemma, thus connecting a presynaptic notion of importance to a computational role in lifelong learning. Overall, our findings present a set of hypotheses for how presynaptic plasticity and stochasticity contribute to sparsity, energy efficiency and improved trade-offs in the stability-plasticity dilemma.

Data availability

Code for experiments is part of the submission and is published on GitHub (https://github.com/smonsays/presynaptic-stochasticity).

The following previously published data sets were used

(2020) Data from: Unified pre- and postsynaptic long-term plasticity enables reliable and flexible learning
Dryad Digital Repository, doi:10.5061/dryad.p286g.

https://doi.org/10.5061/dryad.p286g

Article and author information

Author details

Simon Schug

Institute of Neuroinformatics,, University of Zurich and ETH Zurich, Zurich, Switzerland

For correspondence
sschug@ethz.ch

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0002-5305-2547
Frederik Benzing

Department of Computer Science, ETH Zurich, Zurich, Switzerland

Competing interests
The authors declare that no competing interests exist.
Angelika Steger

Department of Computer Science, ETH Zurich, Zurich, Switzerland

Competing interests
The authors declare that no competing interests exist.

Funding

Swiss National Science Foundation (Ambizione Grant,PZ00P318602)

Simon Schug

Swiss National Science Foundation (CRSII5_173721)

Frederik Benzing
Angelika Steger

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Copyright

This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.