Why not just track keystrokes?
It's the obvious first idea, and it's the wrong one. Research has already demonstrated forgery attacks against keystroke-timing verification, and found that timing data alone carries no real information about who authored the text. It proves a human was present, not that they wrote this. Verified Writing tracks something harder to fake instead: whether a manuscript grew progressively, in a specific unbroken order, over many recorded sessions, and whether that recorded history has been rewritten since.