There are lots of books "about fonts" but for the most part they're just collections of "bunches of them," sometimes with claims about what the designer was trying to accomplish but with little about whether it worked.

That probably just suggests that most of the designers have been "artists" rather than "scientists." (?)

The very scant information on what makes a font more or less legible does suggest that for "dense type" as in a newspaper or magazine, a serif typeface is more legible since the serifs add information. The classic Times New Roman was designed specifically for newspapers, (including the ability to be "justified" so that both sides of the column are mostly straight) and it or very similar faces have been prefered by many for a very long time.

For a more "open type" the sans serif fonts may be more comfortable to read. The relatively low resolution (96 dpi?) for most optical display screens, compared to stuff printed on paper (1200 dpi for good book priners?) might make the sans serif styles better for your computer.

For dyslexics, it would be only a guess but a font with fairly strong "directionality," as with "bottom heavy" or a suggestion of right/left "lean" or "weighting" (lopsidedness?) might be a help, although I haven't seen more than casual mention of any strong opinions.

A lament frequently expressed by the designers is that, according to the books, you cannot copyright or patent the shapes of the characters (the design) but you can copyright the name you use for it. This leads to multiple versions of typefaces so closely similar (or identical) to each other that you can't really tell them apart, but with each version called by a different name.

As the web seems to have a support group for just about everything, there almost certainly must be sites for discussion of nearly any vision impairment that's been named, and those might be places to look for comments on what's helpful; although even on those kinds of sites finding why tit is better than tat is probably unlikely to produce much that's very well documented.

John