Toy simulations to try and better interpret peak bandwidth in power spectra.
See the Makefile for experiments to date, as well as the ipynb folder.
- fakespikes: https://github.com/voytekresearch/fakespikes
- brian2: https://brian2.readthedocs.io/en/stable/
- numpy, scipy, etc (i.e. install conda).
- to run the experiments in the
Makefileyou'll also need gnu parallel: https://www.gnu.org/software/parallel/
- Changes to drive (input), and
- changes to oscillatory burst length probably have identifiable bw effects in real data.
- But the more interesting variations, where oscillator number changes, or phase entrianment chages, don't have strong (or any) BW effects.
- Also worth noting that adding any noise to WC models leads to very sharp noisy peaks in the power spectra. It's a far stronger effect than I'd have predicted.