This experiment asks which input-rate interval preserves enough digit information after synaptic and membrane filtering to justify using it in a variable-rate PING training run.

This experiment will find the lowest Poisson rates that preserve MNIST class evidence. An artificial neural network (ANN) will classify static features formed by passing each pixel’s spike train through an AMPA synapse and non-spiking membrane, then averaging its voltage over the presentation.

A published result often needs a conversation before it needs another revision. Hypothesis adds that conversation directly to a demolab web page: select a passage, attach a comment, and reply in a shared thread. The annotation stays outside the scientific record, while the experiment, figures, and PDF remain reproducible and unchanged.