WASHINGTON - Toddlers rescued from orphanages and placed in good foster homes score dramatically higher on IQ tests years later than children who were left behind, concludes a one-of-a-kind project in Romania that has profound implications for child welfare around the globe.

The boost meant the difference between borderline retardation and average intelligence for some youngsters.

Most important, children removed from orphanages before age 2 had the biggest improvement - key new evidence of a so-called sensitive period for brain development, according to the U.S. research team, which included a Temple University scientist.

"The longer they stay in the institution, the worse their IQ," said Charles Nelson III of Harvard Medical School, who led the study being published today in the journal Science.

"What we're really talking about is the importance of getting kids out of bad environments and put into good environments," Nelson said.

The research already is credited with influencing child-care reform in Romania, and UNICEF has begun using the data to push numerous countries that still depend on state-run orphanages to start shifting to foster-care-like systems.

"The research provides concrete scientific evidence on the long-term impacts of the deprivation of quality care for children," UNICEF child protection specialist Aaron Greenberg said. "The interesting part about this is the one-on-one caring of a young child impacts . . . cognitive and intellectual development."

That orphanages are not optimal for child development comes as no surprise. Earlier studies have found that thousands of children adopted during the 1990s from squalid overseas orphanages in Eastern Europe, China and other nations continued to face serious developmental problems even after moving to affluent new homes with doting parents.

But the new study is one of the first scientific investigations to pin down answers to persistent questions, including how much damage early neglect does. Using a unique approach, U.S. researchers randomly assigned 136 young children in Bucharest's six orphanages to either keep living there or to go live with foster parents who were specially trained and paid for by the study. (Romania had no foster-care system in 2000 when the research began.)

The team chose apparently healthy children, and repeatedly tested brain development as they grew, including tracking those who ultimately were adopted or reunited with family. For comparison, they also tested the cognitive ability of similar children who were never institutionalized.

The researchers found that by age 41/2, youngsters in foster care were scoring almost 10 points higher on IQ tests than the children left in the orphanages. Children who left the orphanages before age 2 saw an almost 15-point increase.

Then Nelson compared the ages at which children were sent to foster care. Every extra month spent in the orphanage, up to almost age 3, meant roughly a half-point lower score on those later IQ tests.

Children raised in their biological homes still fared the best, with average test scores 10-20 points higher than the foster-care children.

What does that mean as these children grow up?

IQ tests don't determine how successful people are in life, Nelson stresses, and he has only now begun testing these children again as they turn 7 and 8. It's possible he will find they have caught up.

"What a parent should expect is that the older the child is when they leave the institution, the more likely that child may have some developmental problems and the more difficult it may be to ameliorate those problems," he said.

The Romanian government requested the study, and began its own foster-care program shortly thereafter.