The study of seven-and-a-half-year-olds by scientists in Detroit found lower IQ scores, and memory and problem-solving difficulties among those who had had low-level exposure to alcohol during pregnancy.
The research, which the Department of Health says it wants to consider, appears to challenge current British Government advice that pregnant women can safely consume one to two units of alcohol a week.
It has long been established that serious problems such as severe learning difficulties and physical abnormalities can occur when women drink large amounts of alcohol during pregnancy, leading to foetal alcohol syndrome (FAS).
Now American scientists say more research is needed to look at the damage caused by prenatal exposure to alcohol at lower levels.
Julie Croxford, from Wayne State University in Detroit, said: "In the past, much focus was placed on studying full-blown FAS. More recent research has considered those individuals damaged by lower levels of exposure. This is an important focus."
The study, published in the journal, Alcoholism: Clinical and Experimental Research, looked at 337 African-American children who were exposed to moderate to heavy levels of alcohol in the womb.
It found they were able to perform memory, number and other tasks as well as other youngsters when these tasks were simple, such as naming colours.
But when the children were pressed to respond quickly while having to think about the response, their processing speed slowed down significantly.
Researcher Matthew Burden, from Wayne State University, said: "Prenatal alcohol exposure is often associated with slower reaction times and poorer attention in infancy and some of these deficits may be at the core of poorer academic performance and behaviour problems often seen later in childhood.
"In cases of FAS, lower IQ scores are common, often reaching the level of retardation. This is because alcohol consumed by the mother has a direct impact on the brain of the foetus.
"However, full FAS is not required to see this impact it is just less obvious to detect across the array of exposures found in foetal alcohol spectrum disorders (FASD), which include effects of prenatal alcohol at lower drinking levels."
Ms Croxford said: "This is likely to mean that these children may be more and more challenged the older they get by the demands placed on them within the school system and within their day-to-day social interactions. This reinforces the current public health message that women should not drink alcohol during pregnancy."
It is estimated that one baby in every 3,000 suffers from the full effects of FAS, others are struck with milder forms of symptoms.
Campaigners in the UK have said that the only definitely safe level of alcohol during pregnancy is no alcohol at all, saying that recommended limits are too high.
The Department of Health said: "Current DoH advice is that women who are pregnant or who are trying to get pregnant should not drink more than 1-2 units of alcohol per week.
"This guidance was reviewed as part of the Government's Alcohol Harm Reduction Strategy in March last year and was found to be safe."
"We would be interested to see any further research into this area but current evidence does not justify changing our advice."
The research echoes warnings from British scientists in the past year.