I'm not totally sure I'd call this sufficient evidence since functional damage != many-to-one mapping but it would shave some points off the probability for existing tech and be a pointer to look for the exact mode of functional memory loss.