Very interesting; that reminds me of a phenomena I learned about in a college neuroscience class, from studies of people who are cortically blind, as in their visual cortex is dead/nonfunctional, but their eyes and the rest of their brain work (e.g., from a brain injury localized to the back of their head where the visual cortex is located).

You can also present them with a forced choice test, such as a panel with big stripes either horizontal or vertical, and require them to say if it is horizontal or vertical. Of course the initial response is "I can't see it, so I can't say". But with a forced choice, they get it right something like 85% of the time — obviously far better than the chance results we'd get with blindfolded subjects.

It turns out that (at least the working understanding at the time) is that the brain has different circuits that use info from the optic nerve to tell the eyes where to track and look. E.g., this will get the eyes to track moving objects, or focus along the edges, etc. When those are still working, the brina can still somehow subconsciously access some of that knowledge from those different brain areas.

I wonder if this is relevant to the phenomena you described?