v0.2 PNAS revision

The main changes I've made, other than wording things and figure tweaks are:

• I figured out how to more concretely distinguish between "informativeness" and "compressibility": informativeness ends up being about time whereas compressibility is about space. (I've written up a more detailed explanation in response to the reviewers' comments; see response letter). This is the biggest advance IMO, and it's something I had been struggling with describing intuitively. But I think I finally "get" it at a deeper level. So...phew!
• I've added some simulations and analyses in Figure 1, illustrating how we can vary informativeness and compressibility independently, and how those changes affect decoding curves
• I've added some new stats throughout to test some of our key claims (all in response to specific reviewer requests)
• I added a new permutation-based analysis to check whether the number of nodes in each network was influencing our "higher order networks are more informative and compressible" finding. This was requested by a reviewer, and was a huge pain to carry out because the analysis takes a long time, and I had to re-run the full pipeline for each permutation. The final result is (as I think was obvious even without the analysis) that our results cannot be explained solely by the number of nodes in each network. Still, at least we have a definitive answer for the reviewer!

v0.1 (Initial release)

First public release; corresponds to bioRxiv submission