Karl C. Garrison, Psychology of adolescence, 6th edition, Prentice-Hall, Englewood Cliffs, N.J., pages 199-200

An interesting characteristic of mathematically gifted adolescents was their independence with regard to how they spent their out-of-class time. ‘Though they played some individual sports and some musical instruments, they completely resisted any regimented activity in the way of planned recreation. In fact, irregularity would seem to have been the rule with a high drive level continually displayed and an occasional spurt of frenzied mental activity.’

... It seems likely that mathematical talent depends on the neurological and chemical organization of the brain. Hence the personality traits descried in the foregoing passage very likely derive also, directly or indirectly, from neurological or mathematical factors. This is interesting because it suggests that neurological or biochemical factors, rather than psychological factors such as childhood experiences and so forth, account for my own imperative need for complete personal autonomy, for doing things on my own initiative, for not being part of the system. Why don’t other research mathematicians rebel as I did?

Suppose because they have satisfied their need for autonomous action by retreating to a fantasy world -- i.e. the world of mathematical abstractions. Mathematics is probably the last area of scientific research where the “lone wolf” investigator still predominates. Thus it is excellent for one who needs to exercise autonomous initiative, provided he is willing to have as the principal concern of his life a body of abstractions unconnected with the practical aspects of his daily existence such as the food he eats, the clothes he wears, the people, animals, and physical objects around him, etc. Where I differ from other mathematicians is in having refused to accept a life in a world of abstractions and in having insisted on the opportunity for autonomous action on my immediate personal environment ...