I pick the thing on my to-do list that sounds most appealing to work on and work on that, even if it's not the most urgent or highly needed thing. If writing a single line of code seems intolerable, then I do non-code work, like writing documentation or filling out my expense report. If I can't find the motivation to write tests, I'll bang on that cool new feature for a while – etc, etc. Ideally, you can pick something small and (relatively) fun or interesting while still doing work that needs to happen eventually.
I timebox this stuff at 1 or 2 hours max (so I don't lose the whole day to a thing that's not top priority), but I usually find that once I break the mental block on "working" and can feel like I've accomplished something, tackling the other stuff doesn't seem so daunting anymore.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.