Data for “When Coherence Resonance Does Not Predict Computation: Target-Weighted Capacity Predicts Noise-Optimal Readout Performance” Files

Description

This dataset contains the final analysis-ready numerical outputs supporting the article “When Coherence Resonance Does Not Predict Computation: Target-Weighted Capacity Predicts Noise-Optimal Readout Performance.” It includes spontaneous coherence-resonance results, primitive capacity curves, observed mixed-target performance, prospectively committed additive predictions, noise-optimum and regret summaries, baseline and robustness analyses, data underlying the main figures and tables, and metadata associated with the prospective SHA-256 commitments. The deposit contains the final numerical data used for the reported analyses rather than full time-step membrane-voltage and spike trajectories. Source code is not included.

Institutions

• The University of Sydney

  New South Wales, Sydney

Categories

Licence