Forget MRI not – mapping neurodegenerative diseases

6 March 2015



While imaging technology has helped researchers to map the brain and understand many of its processes, memory has, until recently, remained a mystery neuroscientists have predominantly only been able to speculate about. Chris Godfrey speaks with Professor Emrah Düzel from the University Hospital Magdeburg to find out how his team have used high-resolution MRI scans to unlock the secrets of an important human memory circuit and assist our understanding of neurodegenerative diseases.


Memory is an enigma that has long puzzled science professionals. Though plenty of models exist that broadly outline the process, little is known about its finer intricacies. Why does our brain elect to store some things and not others? What causes us to forget? Why do neurodegenerative diseases affect some and not others? There are plenty of theories, but for the most part, physiological evidence remains elusive.

Still, fuelled by technological developments, our understanding of the brain and its capacity to store memories has progressed. Advances in magnetic resonance imaging (MRI) have enabled us to see the brain like never before - we can now detect brain tumours in their relative infancy and with far greater accuracy, or monitor which areas are responsible for processing language, pain and learning. We can even communicate with coma patients in vegetative states, measuring their brain activity when they're asked questions or instructed to imagine playing tennis.

One of the latest developments may have finally made crucial inroads into exposing the secrets of the memory - and with it, our understanding of Alzheimer's disease.

Wonder down memory lane

Professor Emrah Düzel, a director at the Institute of Cognitive Neurology and Dementia Research (IkND) at the University Hospital Magdeburg, Germany, was part of the team that pioneered a study to determine the precise location of where memories are created. With the help of some specially developed, highly accurate MRI technology, the group hopes to enable a greater understanding of how memories are processed and the neurological effects of diseases like Alzheimer's.

"Throughout my career, I've been very interested in cognitive function and cognitive neurology, trying to understand the brain network that enables certain functions such as memory," says Düzel. "The drawback with imaging studies so far is that we've been unable to look at circuits. Although the brain is organised along particular circuit-like pathways, which have quite clear input and output structures, we haven't been able to tell these apart."

Though certain types of surgery have helped with understanding which regions of the brain are important to memory, such as the hippocampus or entorhinal cortex, neuroscientists could only speculate about their precise functions.

The previous limitations in MRI technology prohibited us from knowing the reasons why these areas became active or how they process information at all.

"We know from functional imaging studies that when we're shown something new that needs to be memorised, certain regions become activated - but that was really it," explains Düzel. "We could only say that the memory task has activated a region, but we couldn't really look at this as a circuit that processes inputs or outputs. Consequently, we couldn't say what goes in or what comes out of a certain brain region."

Neuroscientists recognised that input and output structures in the brain tend to be segregated into different layers of the cortex or other regions, but separating these layers requires imaging technology capable of producing an incredibly high resolution, one that would present a dramatically different way of interpreting brain activity maps.

"What we thought was that if we can basically look at different layers, the deep or superficial regions for instance, then we could understand where memory-related information goes in or comes out," says Düzel. "The ability to look at input and output layers separately allows us to start looking at what goes in to these structures and what goes out. Rather than saying this region is active, we will be able to say that this region provides an output or receives an input about something."

Novel finding

The thickness of the enthorinal cortex is 3.2mm, but typical MRI technology usually has only been able to offer resolutions between 3.0 and 3.5mm, covering the entire cortical patch; Düzel's team used a resolution of 0.8mm. Though conventional MRI has a magnetic field strength of up to 3T, they used one capable of 7T. As a result, they could look at the superficial and deep third of the cortex, and see which of these corresponded to input and output layers, determining what they represented.

Analysing the data provided by such images posed another challenge for Düzel's team. Deciphering MRI data is already a complex task, involving the careful alignment of scans for each corresponding region in participants to create an aggregate. The higher the resolution, the harder this becomes.

"Advanced MRI technology and the ability to carry out more complex analyses have helped neuroscientists to identify the presence of a ‘novelty signal’ in the memory system."

"It works if your resolution is not so high, but not if you have such an ultra-high resolution as we have, as you cannot align so perfectly across participants," says Düzel. "You have to use a new approach where you really consider each minute brain region from each participant separately. Then you do separate statistical comparisons within individual subjects, followed by comparisons across subjects."

This advanced MRI technology and the ability to carry out more complex analyses have helped neuroscientists to identify the presence of a 'novelty signal' in the memory system - a process that effectively activates the memory circuit whenever something new is experienced, deciding what's to be remembered and what forgotten.

"If you look at the superficial part of that cortex, we see it carries information about everything that's novel; if you then go deeper, you see that it carries information about what is remembered," says Düzel. "Say you look through a selection of new photographs: every new photograph will activate the superficial layer, but you will get activation at the deep layers only for those photographs that you will be later able to remember.

"Each new photograph will activate the superficial layer, but only a few will activate the deep layer," he continues. "There has been a debate, for instance, if the output from the memory circle is itself a novelty signal, but we know that cannot be the case. What comes out really is a signal that tells the cortex, 'This can be remembered', 'This will be remembered' or 'This must be remembered', and not 'This is novel'."

Commit to memory

As well as providing a more accurate model of the human memory circuit, it has the potential to focus on any system to look at directionality of information flow - motor regions, sensory brain areas and others - as well as various types of diseases. Düzel's team hopes to use it to understand the cause of neurodegenerative conditions, starting with Alzheimer's.

"You can get a more circuit-level mechanistic understanding of brain activity," says Düzel. "You move away from basically saying a brain region is active or not active to saying a brain region produces an output or receives information. This is really true for any system that shows this type of layer-specific organisation, and most systems in our brain do that."

While it's already known that Alzheimer's causes specific regions in the brain, like the entorhinal cortex or hippocampus, to shrink, there's little evidence showing exactly why this happens and how the disease progresses across different parts of the brain.

"We can really start to understand at what level something is going wrong," says Düzel. "In principle, we can now start to understand whether the input into the memory circuit is generated before the output. So we can look at the superficial part of the entorhinal cortex and see whether the information that goes into the hippocampus is still intact or whether it's already corrupted, or alternatively, whether that information is still intact and what comes out is basically corrupted."

Düzel and his team now plan to combine scanners from multiple centres and look at larger populations, granting them more accurate, reliable statistical data. There are also plans to use the MRI technology to explore the impact of aging and its relationship to Alzheimer's.

"One important question is whether Alzheimer's and aging are really analogous; are they just different phases of the same process or are there qualitative differences about the disease? Is there a difference at all?" asks Düzel. "Now we can really start looking at this at a circuit level. We were starting with two centres and hoping to proceed to more. This is really our next goal."

It may be years or even decades before neuroscientists have a complete physiological understanding of the memory circuit, how it works and why it works. We may even never know.

But the work of Düzel and his team will certainly advance the learning process. An aging population will see increasing numbers of sufferers of Alzheimer's and related neurodegenerative diseases; any research that might help mitigate the impact would prove memorable.

Professor Emrah Düzel is director of the Institute of Cognitive Neurology and Dementia Research at University Hospital Magdeburg, Germany. His work uses multimodal imaging to study the functional anatomy of human episodic memory networks, neuromodulatory circuits, their alterations in aging and neurodegeneration, and their scope for plasticity.
MRI technology has traditionally been limited to broad-scale ‘on/off’ activation analysis.


Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.