Paul Rubin wrote:
Reuben Fine claimed there were two kinds of world champions: A) those
who weren't interested in anything except chess (Fischer was a classic
example); and B) those who had other interests as well. He said type
A were generally crazy and type B were not. He discussed this in
detail in his book "The Psychology of the Chess Player". On the other
hand, that book seemed pretty crazy itself in many ways.

Yes, I have that egregious little book (if at 74 pages it can be
called a book). Written in 1956, it does not mention Fischer, but
surely would have were it written later.
The actual terms Fine used for his two types were "heroes" (among
whom he numbered Morphy, Steinitz, Capablanca and Alekhine) and
"non-heroes" (Staunton, Anderssen, Lasker, Euwe, Botvinnik). For the
latter chess was merely one of several important interests, for the
former it was pretty much the only interest.
Fine says his four heroes all entertained "fantasies of omnipotence"
and showed "considerable emotional disturbance" either during their
chess careers (Capablanca, Alekhine) or soon thereafter (Morphy,
Steinitz). Frankly, I'm not at all persuaded about Capablanca; to my
knowledge, his troubles stemmed more from physical sources (high blood
pressure) than psychological. Much of what Fine says about the others
is anecdotal, without supporting evidence, and some of it is now known
to be apocryphal. The amount of factual error and worthless hearsay in
Fine's little tome is amazing.
As far as hero types generally being crazy, again I don't think Fine
had much of a case, Fischer notwithstanding. Other great masters who
(as far as I know) have been pretty much obsessed with chess included
Pillsbury, Marshall, Blackburne, Janowski, Spielmann, Tal, Korchnoi,
Portisch, and Kasparov, and while many of them had their quirks, I
don't think any is considered psychotic.
He might have made a case for the hero type being at least more prone
to "emotional disturbance," but the sample he uses is too small to have
statistical significance, and too selective to be logically valid.