Magnetic brainwaves

Delft Outlook, 2009.1

The mental disorders that often accompany old age can largely be attributed to the erosion of connections in the brain. Researchers at TU Delft are now developing new ways of mapping the decaying brain. “This project is of great social importance.”

Download as .pdf

Hemmed in between three high-rise apartment blocks lies the unimposing grey building of the Ommoord Health Centre. From the outside, nothing suggests that this is where the Rotterdam Study (see text box at page 22), a major epidemiological study of health in old age, is being conducted. Ommoord, a district in northwest Rotterdam, was selected because the area was deemed a representative model for the Dutch population as a whole, and perhaps also because Ommoord is situated close to the Erasmus Medical Centre (Erasmus mc), the Netherlands’ largest academic medical centre.

At the Ommoord Health Centre, a steady flow of people aged 45 and over comes in to enrol for the population survey. Upstairs they undergo bone scans, eye tests, and ultrasound tests of their cardiovascular systems and blood vessels. Downstairs is where the neurological tests are done for cognitive alertness and motor skills, plus a brain scan using the centre’s mri (magnetic resonance imaging) machine. The neurological aspect of the Rotterdam Study is the responsibility of Dr Monique Breteler, professor of neuroepidemiology at Erasmus MC.

Essential

In a recent article in NeuroImage, a journal of brain function, researchers at Erasmus mc wrote that a special imaging technique known as dti (diffusion tensor imaging – see text box at page 22) could shed new light on how white brain matter changes as the brain ages.

A section of the brain showed grey matter on the brain’s outer part, and white matter on inner part. The grey matter mainly consists of the cell bodies of nerve cells, whereas the white matter is comprised of connections between brain cells — extensions of nerve cells sheathed in a layer of protective white myelin, which explains the white colour. Good connections are essential for the brain to function well. A loss of white matter is associated with Alzheimer’s disease, dementia, and schizophrenia. The question now is what causes the loss of white matter to commence.

“We know that various diseases and ageing cause myelin to break down. The result is that axons – the nerve cells’ extensions – are dismantled,” says Dr Wiro Niessen, who as professor of medical image processing works at both Erasmus MC and TU Delft. “This causes a fibrous microstructure of nerve bundles to change into a more amorphous structure. Something changes inside the brain’s nerve channels at a microscopic level and this alters the mri signal. The current thinking is that we may be able to use dti techniques to measure the damage to the white matter at an early stage.” It is hoped that early detection of dementia could help slow down the progression of the disease. Niessen: “This knowledge will certainly allow us to better manage patients. If we could delay the onset of dementia by a year, it would save society billions. This project is of great social importance.”

“Do you know what diffusion is?” lecturer Dr Frans Vos asks, when asked to explain the principle of diffusion tensor imaging, or dti. This new mri technique is currently being developed further by the Quantitative Imaging research group of TU Delft’s Faculty of Applied Sciences’ department of Imaging Science & Technology. To explain anisotropic diffusion, Vos uses a piece of tissue paper and a sheet of newspaper. A drop of ink spreads evenly (isotropically) on the tissue, forming a wet circle; however, on a newspaper, the result is an ovalshaped spot, because diffusion takes place more rapidly in one direction than in the other. In other words, anisotropic diffusion occurs. To explain the diffusion that occurs in nerve strands, Vos is ready with an analogy: “Isotropic diffusion is water that can go anywhere. Now, suppose you drop a bunch of spaghetti into water. The water can flow along the length of the spaghetti, but not at right angles to it. The same happens in white matter — water molecules inside white matter strands can move easily along their length, but with difficulty at right angles to them, because they are stopped by the cell wall and myelin layer.”

DTI is an imaging technique that so far has been used mainly in academic medical centres. Neurosurgeon Aaron Filler invented the technique in 1991. dti adds a sensitivity for biological microstructures to mri, which is a technique suitable for making images of the body’s various biological tissues. This specific sensitivity renders the technique highly suitable for creating images of nerve bundles, as mri images are often too coarse for this. More generally, dti is considered a suitable method for brain scans, allowing for clear images of strokes, white matter, and interconnections inside the brain.

Even so, dti technology is far from fully developed. “Making images is only the first step. After that comes the interpretation stage, and that’s not simple,” Niessen says. “There is a lot of noise, we’re dealing with complex anatomy, the nerve strands are narrower than the resolution of the scanner, and the nerve strands are all criss-crossed. We end up with an enormous amount of data about nerve strand topology, but is requires modelling to convert that data into a useful image. And a medical degree can’t help you there: you need engineering, mathematics and information technology. That’s where TU Delft comes in.”

Using dti to study the development, diseases and degeneration of the brain’s white matter is an active research area: since 2005, more than 2,500 research papers have been published on the subject.

Less damaging

““We are always on the lookout for new ways to extract additional information from the data,” says Matthan Caan, a phd student who, like his supervisor Dr Frans Vos, also works at both TU Delft and the Amsterdam Medical Centre (amc). At the amc, Caan researched the side-effects chemotherapy had on the brains of young cancer patients. To do this, he used a form of dti analysis, called tractography, to track the paths of the nerve extensions. Put simply, this amounts to picking a starting point and then following the direction of greatest diffusion, assuming that this corresponds with the direction of the nerve strand. This results in an image of gossamer-thin traces that represent the tracks of nerve strands inside the brain.

For his research, Caan had two sets of six brain scans from children being treated for cancer with chemotherapy. Half of the patients had received a high dose, while the other half a lower dose. Could Caan tell the difference? To measure the diffusion, he decided to focus on the corpus callosum, which is the thick nerve bundle that connects the two halves of the brain. He collated the data for all the young patients to get average values, and then, for a certain length of the nerve strand, calculated the deviation from the average value. Caan: “Normally, you’d be looking at each individual point in turn, and you wouldn’t notice much of a difference from the average. We combine measurements across a much wider range to see whether the anisotropy decreased across the entire bundle. This better matches the hypothesis for the side-effects, which is that chemotherapy affects the white matter.”

Caan was able to demonstrate that reduced doses of chemotherapy cause considerably less damage to the brain. The brain damage in the images produced by Caan also matched the clinical data — the worst affected patients had on average a lower iq and took longer to understand a joke. Other research has shown that the lower dose was just as effective in terms of fighting the cancer, meaning the treatment could be safely adapted.

One reason Caan chose to study the corpus callosum is that its nerve bundles largely run parallel to each other. Once nerve bundles diverge, their paths become much more difficult to trace. This is the job of yet another phd student, Ganesh Khedoe. He will expand the tractography (i.e. the tracing of nerve strands) to include crossing strands that cannot be differentiated using a standard dti. Khedoe: “I want to identify specific strands inside the brain. They could come from different brains, or from the same brain but at different times. I will look for differences between them.””

Biomarker

“We’re concerned with the clinical added value of images showing the degeneration of nerve strands,” says Niessen, who is involved in the Rotterdam Study. “It’s not about a single image, but about statistical analysis. We collect hundreds, thousands of images of people’s brains, and we track what happens to them. This provides us with a gallery of images and a library of life stories and problems. Statistics and the way Alzheimer’s disease affects the aged have led us to believe that dti images could be used for early diagnosis of neurodegenerative disorders. We are now in the process of finding out whether that is true.”

Research like this requires a long-term population study, because researchers are looking for brain scans that long predate the moment when a person visits a neurologist. Among the roughly 5,000 people participating in the Rotterdam Scan Study, a significant number of them will ultimately develop Alzheimer’s disease; however, by that time there will be scans showing their brains before and after the onset of neurological complaints. This constitutes unique material for scientific research. Niessen: “We will use statistical methods to look for differences between people who do and do not develop Alzheimer’s. Will we then be able to use brain scans to predict how the disease progresses? Thanks to our collaboration with TU Delft, we are now better able to quantify the images, and ultimately answer the question as to which number best predicts the course the disease will take.”

“That is the holy grail of this research,” Vos acknowledges. “We are looking for what we call ‘biomarkers’, specific indicators of a certain disease. A biomarker allows you to recognise whether a person is destined for normal old age, or for one that involves Alzheimer’s. We believe such a biomarker is hidden among the diffusion data.”

According to Niessen, discovering an Alzheimer biomarker is mainly a question of time. The dti scanner tests in Rotterdam started in 2005. “In two to three years’ time we will have a follow-up covering five to six years,” Niessen says. “A number of people will have then been scanned a second time. I expect that we will have conducted the first tentative studies by then, and we can look forward to a solid statistical basis in five years’ time. Whatever the case, we’re certain to get some results.”

The question remains as to how this knowledge will be used. After all, you can hardly perform preventative scans on every Dutch person aged 45 years and over. Niessen’s view is that an mri scan would be reserved for people with an increased risk of Alzheimer’s disease, assuming, of course, you can indeed identify this group.

Niessen: “Initially we hope to discover more about what takes place during dementia. Such knowledge would certainly help improve the treatment of patients. The obvious course would be to apply it to early diagnosis and preventative therapy. But how exactly everything would work out depends on the accuracy of the prognosis, among other factors. There is still a lot of research required for this.”