As a grad student (Computer Science), I always feel like I lack the tools needed to make progress with my research. One solution I thought might be helpful would be to take more courses, especially mathematics courses. However, I also feel like of all the courses I have done in undergrad and masters, very little sticks in my mind, and very little is relevant to my actual work. How to decide whether (and which) courses would be a worthwhile investment?
In my experience, in mathematics, in the U.S., a course's instructor matters enormously, probably decisively.
I was never a fan of going to sit on a hard chair at pre-ordained times to hear a monotone, flawed recapitulation of a small snippet of what was in a book that I'd already read. Yes, that sort of class can benefit people who would otherwise not read the book at all, passively absorbing something by sitting in the room.
In some cases a course/instructor covers material that is simply not available anywhere else, etc. If it's of interest, that can be great...
And, sometimes, an instructor's viewpoint on even seemingly mundane things can be very, very insightful, far beyond very-conventional textbooks. This can be not only profoundly informative but inspirational.