Computers shed light on mental mysteries At Kennedy Krieger, brain 'dissections' are done by database


Venturing millimeter by millimeter, computer scientists and physicians are probing the human brain in startling new ways that rely on software instead of scalpels. They're dissecting it mathematically.

Researchers at Baltimore's Kennedy Krieger Institute are part of an effort to expand how computers can measure brain structures -- and uncover how differences reveal disease and shape human behavior.

What they're finding helped solve the puzzle of identical twins who turned out remarkably different.

AOne of the 16-year-old girls has a normal IQ, and some social and emotional problems. Her twin is mentally retarded.

DNA studies showed that both girls inherited a fragile X gene, but with slightly different mutations. Yet no one could see exactly what was happening inside their brains as a result.

Thanks to the software developed at Kennedy Krieger, the answer became apparent.

Images from numerous brain scans were loaded into the computer, translated into numbers and manipulated mathematically.

Researchers analyzed the brains of many healthy people to define for the first time what was normal.

Then they were able to compare the normal brains with those of the twins and others with Fragile X syndrome.

Ultimately, the computer showed precisely how and where the twins' brains differed.

"It's bringing us closer to what went wrong or what's going

wrong," said Dr. Allan L. Reiss, director of Kennedy Krieger's behavioral neurogenetics and neuroimaging and the Department of Psychiatry.

"Until very recently, the method of analysis has been performing a scan, going to your nearest light box and looking at it."

That method taps the physician's expertise, but it is crude compared with a system that takes pictures of 124 individual slices of the brain, from the nose to the back of the head, each XTC slice just 1.5 millimeters wide.

"It used to be we only saw the brain if a person was injured or dead," said Michael Huerta, who directs the Human Brain Project at the National Institute of Mental Health, the lead organization.

"These new techniques are providing windows on the essence of human beings."

Project expands

The project, which began in 1993, now underwrites 12 grants at cost of $5.5 million a year.

Five federal agencies -- from the National Institute of Health to NASA and the Department of Energy -- are sponsors, and all coordinate discussion among the researchers.

The work is similar to the infrastructure that computer scientists had to develop to track and categorize the Human Genome Project, the federal effort to map the entire human genetic sequence.

In that case, scientists were dealing with -- at its core -- a fairly simple information problem: a set of four key letters in the genetic alphabet, in different lengths and combinations.

But the Human Brain Project is exponentially more complicated. It deals with how different groups of cells communicate with each other, how molecules interact with receptors, and how all of that underlies complex behavior.

Since neuroscience is one of the hottest areas in research today, with hundreds of findings reported each month, computer programs that can pull the information together in a comprehensive way are crucial.

Kennedy Krieger received a $1,043,775 grant over five years, and its researchers are working on the brain images of children and adults, although their primary focus is children's diseases.

Sharing software

About a year ago, researchers from the Baltimore hospital put their software on line, so other scientists can use the techniques.

In one such technique, computer scientists have bypassed an excruciatingly long process of determining exactly what part of the brain is affected by adrenoleukodystrophy, or ALD, the rare disease that was the subject of the movie "Lorenzo's Oil."

By electronically planting "seeds" in the computerized image of the brain, not in the brain itself, scientists can watch those seeds grow until they highlight the exact area where the protective sheath around nerve fibers has been destroyed.

Through the computer developments, Dr. Reiss is also tracking how Kennedy Krieger patients with ALD are responding to two drugs.

Pictures of the children's brains are taken at the beginning of the clinical trial, and then at three, six and nine months, so researchers can assess how the affected parts respond to treatment.

The study is being conducted in collaboration with Dr. Hugo W. Moser, who directs mental retardation research at Kennedy Krieger.

Dr. Reiss can also use the computer technique to compare scans of abnormal brains with scans of hundreds of normal brains, which are being filed into a huge database.

Defining normality

This is a potent juxtaposition because it allows researchers to establish a normal range for the structures that make up the brain.

While brains are basically the same, each is like the human face: one person may have ears that rest a little higher on the head, another may have pouty lips, yet a third has a distinctive nose.

In the same way, each brain has a particular pattern of wrinkles on its thin covering.

That kind of individuality makes determining which deviations are abnormal a formidable job.

Yet the Kennedy Krieger scientists are figuring out how to measure those wrinkles and compare them to the wrinkles of another brain through advanced statistics and algorithms.

"It's a much more complicated picture of what is going on. What they're able to do now is capture the complexity of the human brain in a way that's fairly close to reality," said Dr. Reid Lyon, a neuropsychologist and director of research in learning disabilities at the National Institute of Child Health and Human Development in Bethesda.

"This is certainly a great step forward," said Dr. Lyon, who is overseeing Dr. Reiss' work.

Nature vs. nurture

Physicians say the techniques will shed light on many of the severe neuropsychiatric conditions and diseases that afflict children, such as Turner's syndrome and Tourette's syndrome.

The methods also may provide another path to understanding learning disorders such as dyslexia, Dr. Lyon said.

Because such disorders are influenced by many genes, they are difficult to sort genetically.

They also lack a straightforward marker, such as a lesion or a tumor, to highlight where the problem is in the brain.

Dr. Reiss hopes to sort which behaviors and symptoms are caused by genetics, and which are caused by environmental factors.

At Kennedy Krieger, the work has had an unexpected side effect: it's driving a new way to approach research, said Dr. Gary W. Goldstein, president and chief executive officer.

From now on, they'll be looking at the images to show how the brain connects genetics to behavior.

Said Dr. Goldstein: "This is a model of how we'll study everything."

Copyright © 2019, The Baltimore Sun, a Baltimore Sun Media Group publication | Place an Ad