[wdutch]

I have aphantasia. I relate to what you say about how "feel" a face when I remember it. I'd compare it to a neutral nets embedding. I'm not storing every "pixel" of an image but instead a "vector" that corresponds to the key features (although I'm not consciously aware what those features are). I couldn't describe on demand the faces of my colleagues I see every day but I know them when I see them.