NEW YORK (FORTUNE) -- If we are what we eat, why do we eat what we do? That question has puzzled scientists - not to mention big food companies - for decades, but new research continues to shed light on this fascinating mystery of human behavior.

Much of the work in this area has been done at the Monell Chemical Senses Center, a Philadelphia-based non-profit whose researchers study the senses of taste and smell. The scientists there have found that our earliest experiences, even prior to birth, can establish flavor preferences and eating behaviors. At a recent gathering of the Institute of Food Technologists in Orlando, delegates learned of the latest advances in this field.

Monell's work dovetails nicely with prenatal sensory studies showing that babies become familiar with their mother's voice while in the womb, and can even recall the theme song to a soap opera their mother watched while pregnant.

For example, one landmark study in 2001, headed by developmental psychobiologist Dr. Julie Mennella found that food preferences - and, by extension, cultural flavor principles - are established long before the introduction of solid foods.

Mennella's team randomly divided pregnant women into three groups. One group drank water during the last trimester of pregnancy and then carrot juice during the first two months of lactation; a second group drank carrot juice during pregnancy, and water during lactation; and a third group drank just water during both pregnancy and lactation.

A few weeks after the babies began eating cereal but before they had ever been exposed to the taste of carrots, the researchers videotaped the infants as they were fed cereal prepared with water, then cereal prepared with carrot juice.

They found that infants who had been previously exposed to carrots exhibited fewer negative facial expressions while eating the carrot-flavored cereal compared with the plain cereal, whereas infants in the control group - whose mothers drank only water - displayed no difference. Moreover, those babies exposed to carrots prenatally ate more of, and were perceived to enjoy, the carrot-flavored cereal compared with the cereal and water.

"Their take on things is contrary to the older dogma, which says our preferences are shaped later in life," says Dr. Barry Green of the Yale University School of Medicine. "They push it back into the womb."

Green questions, however, whether these flavor predilections carry forward later in life, which would require long-term studies that could take decades. "The nature versus nurture argument has gone on for many years," he says.

Other studies in this arena have shown that children who are exposed to a variety of vegetables - say, peas, squash, potatoes - are more likely to find palatable a totally new vegetable, like, say, carrots. Also, children who eat healthy foods early in life tend to continue those habits later on, which is good news for obesity and public health specialists.

More recently, Mennella and other researchers have tried to decipher what role genetics plays in taste sensitivities of children and adults. A 2005 Monell study found that variations in a taste receptor gene accounted for a large part of individual differences in bitterness perception in both children and adults, as well as predicting children's preference for sweet-tasting beverages and foods. And no wonder: studies have shown that even infants will suck harder on a sucrose-flavored nipple than a latex nipple.

"Children seem to live in their own sensory world," says Monell research associate Catherine Forestell.

Monell is not the first, nor the only, group to explore flavor learning. Much earlier work has been done on animals, in particular rats. The concept of so-called "sensitive periods," a time during early development when an organism is primed to receive and perhaps permanently encode important environmental information, dates back to ethologist Konrad Lorenz in the 1960s. Even Charles Darwin dabbled in this area more than a century ago, showing that preferences for sour tastes is heightened during childhood.

For all that has been learned, much research remains to be done in this field. "This is becoming more complicated yet more interesting at the same time," says Yale's Green.