How to organize a neural network outside of my brain to help me learn?

I find that with stuff like this drawing it out by hand is the best. When you're forming connections in your head, the interaction between your hand, the paper, and your brain is helpful. Seeing the big picture easily is helpful. Being able to express it exactly the way you picture it is helpful. Being able to access it immediately is helpful. Futzing around with software and working around the program's limitations just uses up your brain's bandwidth and is not helpful. I would get one of those big rolls of paper they sell for kids and put up a big piece of paper on the wall. Or a cheap non-precious sketchbook you can carry around.

Have a look at "thebrain" (for Mac/Windows - it's a mind map that reorganizes depending on from which perspective you want to look at it and uses father/brother/son relationships. It grows with you and you can delve into specific views. A bit hard to explain but very worth it.

This is something I often think about, and have never found a good way to represent digitally. In graduate school I poured all my notes into http://www.devontechnologies.com/products/devonthink/devonthink-pro-office.html, which has a concordance AI that tries to make connections between your files for you. But it's nowhere near as smart as it needs to be, and it can't organize those connections in any way other than through file structure. If you're less interested in discovery than in organization, mindmaps might work just fine (http://freemind.sourceforge.net is opensource and not bad). But the thing with them is that at larger scales, it becomes very difficult to accidentally make connections that you don't know you want to make in the first place. That is to say, at least for me, I've found it's hard to discover connections without you yourself realizing, when you're deep in one branch of a mindmap, that this bit looks like this other bit in some other (usually collapsed) branch of your mindmap. You still have to kind of curate and organize your connections. This may differ for STEM subjects, and it may be different for you. Speaking for myself, as a historian, in the years since I began and left grad school, I've found really no good replacement for the brain making the connections itself. This, through stubborn and dogged reading, re-reading, rethinking, circulating, backtracking, neural-sparking a-ha moments after the fifth, sixth re-read. Certainly no good digital visual way of keeping track of these things. I've literally devolved/evolved into implementing an http://en.wikipedia.org/wiki/Commonplace_book in DevonThinK, and keeping notebooks of ideas in Evernote. YMMV: the historical understanding and connections I'm trying to make and keep appear to be a fairly straightforward function of time, effort and memory; the nature of the connections are best clarified and solidified through writing, talking and teaching. Which is to say, not easily represented visually. I get a lot out of hand-drawing mind-maps of books I'm reading, but that's really much more for the process than the storage, and it would be impossible for me to connect my bookmaps up with each other in any meaningful way. But everyone's different, and I'd be the first to get excited if anyone has any suggestions for digital solutions.

A personal problem is that I am terrible at organizing physical papers and computer files. If I use a notebook or annotate a PDF, I'm not sure that I will ever find something specific again. I've been dumping ideas into Evernote but it's becoming a disorganized pit. The best thing I have found so far with extensive Google searches is this http://calnewport.com/blog/2009/05/11/how-to-build-a-paper-research-wiki/ on creating a research wiki using strict hierarchies. It's pretty simplistic but seems like a decent basis for something like this. Amtho - I don't know that quick data entry would be a problem if you took things one at a time. You could enter things as you read almost as taking a normal note. I can see if you are trying to dump everything you know or all your notes at once that it would be a nightmare.

Yes, I think structure, hierarchies and categorization are key, and my DevonThink database implements Cal Newport's structure for each project. For me, his structure has been really useful in terms of being able to find stuff and to be able to hyperlink between them. And yes, @amtho, one frequently does try to think in terms of labelling connections. But in terms of the intellectual work that those hierarchies, structures and labels for connection are doing: I would say it's very, very hard to do that digitally. The problem is that the very nature of scholarship consists in the changing and relabelling of those categories, all the time, over time. I still find http://archivingthecity.files.wordpress.com/2011/01/mills_on_intellctual_craftmanship.pdf, and why it's probably misguided to think you can start with a bunch of categories -- say, "constructs" or "impermeability", or even "may have caused" -- and think that will continue to do all the categorization work for you at all times in the future. As Mills puts it,the way in which these categories change, some being dropped and others being added, is an index of your intellectual progress … As you rearrange a filing system, you often find that you are, as it were, loosening your imagination.for this, I've never found a good digital proxy.

So far it's disappointingly simplistic, but Google Docs is trying to accomplish something like this when you turn on the "Research" sidebar. As long as you keep all your notes and files in Google Drive, it will continuously search them for you as you write in a Google Doc, and supposedly pops up related information (I guess using keywords that it creates by indexing your files). You can also let it give you information from the web in general, rather than just your own documents, but I find for narrow subjects like academic research, you'll get suggestions that are too broad unless you keep it focussed on your own files. Anyway, it seems like the sort of thing that could eventually do what you want, but so far, I don't think it is indexing files in detailed enough ways. For example, if I'm writing a paper about verb endings in a dialect of English, I would love it to throw up snippets from papers I've saved about verb endings, my own previous writing on that dialect of English, grammatical descriptions of verbs in other languages, etc. But instead it's more likely to give me anything that has the word "English" in it, and anything relating to the keyword "linguistics".

Memail me...

Here is another free mind mapping tool: http://cmap.ihmc.us/. In my grad program we used it to map relationships between various theories, approaches, methods, etc.

When I was completing my masters project I used http://wikidpad.sourceforge.net/ to store summaries of my readings and create links between different ideas.