I use everything in pic related, which sums up recent learning and memory research.
Most of it consists of using flashcard software (I use anki, which is free, there are others you can use). This takes advantage of various psychological mechanisms that help with memory retrieval (so I don't have to look at notes):
Testing effect: en.wikipedia.org/wiki/Testing_effect
Spacing effect: en.wikipedia.org/wiki/Spacing_effect
Interleaving effect: scientificamerican.com/article/the-interleaving-effect-mixing-it-up-boosts-learning/
Flashcard software tends to take advantage of all three, it forces you to retrieve things from memory (testing effect), they usually have an algorithm that spaces out cards depending on how well you know them (spacing effect), and if you keep them all in one deck it takes advantage of the interleaving effect by mixing up questions (making it harder for your memory to retrieve, which strengthens the memory if you can do it).
As to anki. I put in various questions (the input fields for decks are the question and the answer). My university degree was in philosophy and mathematics, so most of my deck is like this (along with some areas outside this).
Broadly, I would say I divide the questions in the flashcard decks up into definitional/conceptual/propositional questions and exercises/procedural.
Definitions are, "what is x?" Or asking a question where you are drawing/tagging a diagram. Or asking for arguments/counter-arguments for some concept. Or asking for equivalence of mathematical identities. Or finishing some conditional in some definition/theorem, e.g. if P, then … (then write whatever Q is in the answer. I'll also write questions for the contrapositive as well). Or asking a historical question (to gain intuition/historical knowledge about things), e.g. what was the problem that this conceptual tool was originally invented/created for?
Exercises/procedural questions are basically cut and pasting exercises and proofs into the questions, as well as programming stuff. For philosophy and math these include basic stuff like algebraic manipulation (solve for x), to graphing, to doing truth tables or semantic tableux. Or a question might be, "prove the following." Or it might be a programming question, e.g. solve this integral in MATLAB.
This is pretty much the only thing I do with respect to study and research. The only other thing I do is keep a read/to-read list where I place books/papers from a variety of areas.