That tired (and sexist) old aphorism about the way to a man's heart being through the stomach is almost true. But it's not the heart, it's the brain.

Psychobehavioral researchers agree in principle with a growing body of popular nutritional dogma holding that what you eat can affect how you behave. Where the establishment researchers differ from so-called alternative therapists is on which foods affect which behaviors, and how.

The relationship of diet to behavior drew some 30 specialists, researchers and clinicians to a 2 1/2-day symposium in Arlington last week, sponsored by the American Medical Association, the Nutrition Foundation and the International Life Sciences Institute.

In a sense, this first major meeting on the subject marks an attempt by the scientific community to reassert jurisdiction over nutrition -- a void they had left for decades. The wide-open field has been subsumed by an informal consortium of nutrition therapists, holistic health and wellness practitioners, mega-vitamin proponents -- ranging from some traditionally trained physicians to California-style mail-order "experts" and out-and-out quacks peddling untested nostrums.

From this group has come an outpouring of food-related anecdotes, some fostered by health food stores and pop-health magazines, many growing out of the '60s-based concept that anything natural was perforce good.

No speaker for the most widely accepted, albeit untested, food concepts -- such as sugar's link to violent behavior -- was asked to to be on the program, but several were present and vocal in their displeasure. Conference spokesmen, however, declared their purpose as trying to find ways to distinguish proven scientific fact from health anecdotes. Anecdote is a dirty word in medical parlance. Sometimes, scientists will concede, anectodal data may be the basis for medical knowledge -- but only when verified by the scientific method, the antithesis of anecdote.

Specifically, most criticism at this conference was aimed at the widely held view that sugar can cause antisocial behavior, including hyperactivity in children and violence in adolescents and adults. The so-called "Twinkie" defense of Dan White -- convicted of manslaughter, jailed and recently paroled for the 1978 killings of San Francisco Mayor George Moscone and Supervisor Harvey Milk -- was a recurrent topic of conversation.

Dr. Richard Wurtman, neuroendocrinologist at the Massachusetts Institute of Technology and a pioneer in discoveries linking foods to changes in brain chemistry and subsequent behavior, said the symposium was designed "to assess what we do know, because of the speed with which public policy is made based on food anecdotes. I think it is one thing for a mother to decide in her kitchen that she thinks that maybe this food or that may make Johnny more hyperactive and decide not to give it to him. But it is another thing for that mother to sit on a jury, say that the man who shot the mayor is not guilty of murder because he ate a lot of Twinkies."

Wurtman's federally financed work, much of which is being confirmed by other scientists, demonstrates that carbohydrates, including those in sugar or potatoes or any other starch, put into effect a series of metabolic events that raises the brain levels of an amino acid called tryptophan, a precursor for the brain messenger serotonin. Serotonin, in turn, is known to be involved in calmness and in sleepiness -- a finding opposite to the anecdotes.

In a study growing out of Wurtman's work, psychologists Bonnie Spring and Harris R. Lieberman compared the afternoon mood and behavior of two groups of adult volunteers. One group was fed a high-carbohydrate lunch and the other a protein lunch. Their results, not previously published, showed that, although effects were subtle, those who ate carbohydrate lunches were sleepier and slightly less alert than the protein eaters.

Other findings reported at the conference:

* Some of the studies purporting to substantiate that removal of sugar from prisoner diets can control violence were found to be flawed because honey, molasses and orange juice were substituted for sugar in the diets. There is no substantive difference between those foods and plain sugar.

* Biochemical studies link deficiencies of iron, iodine, zinc and copper in pregnant mothers to brain defects in their babies. Zinc, suggests a new study from Turkey, may be linked to neural tube defects such as spina bifida and to anencephaly -- babies born with virtually no brain.

* New but tentative results from studies at the National Institute of Mental Health on children whose parents say they "go off the wall" in response to sugar are showing no evidence of sugar reactivity and may suggest the opposite.