Insights into the brain by means of artificial intelligence and “rapid” measurement methods
Bonn/Germany, November 26, 2024. Using artificial intelligence (AI) and accelerated measurement routines, Prof. Alexander Radbruch aims to expand the existing capabilities of magnetic resonance imaging, gain new insights into the brain and improve the diagnosis of neurodegenerative diseases. The Bonn-based radiologist and neuroscientist heads DZNE’s new research focus “Human Imaging”.
Magnetic resonance imaging (MRI) has long been an integral part of medical care and science – this is particularly true for brain research. Because this technology allows for detailed insights into the heads of humans by means of magnetic fields and radio waves. “Unlike many other diseases, in Alzheimer’s, Parkinson’s or ALS, we cannot easily remove diseased tissue and examine it in the laboratory. The human brain is, so to speak, a high-security area that it is better to keep out of,” explains Alexander Radbruch. In addition to his research at DZNE, he also directs the Department of Neuroradiology at the University Hospital Bonn (UKB). “However, MRI is non-invasive and therefore a key tool for both neuroscience and routine medical care.”
Early diagnosis of dementia
However, MRI scans are expensive, time-consuming and it often takes many weeks to get an appointment for an examination, the Bonn scientist points out. “But I am convinced that AI and optimized measurement methods can help to make MRI more widely available and cost-effective. I expect this to foster the early diagnosis of dementia and other neurodegenerative diseases. This would also be significant for supervising therapies. In view of this, I anticipate that in the future, brain imaging will play a bigger role in the diagnosis and monitoring of neurodegenerative diseases. That is what we are researching for. Our aim is to see the methods we develop used in science and clinical practice. We consider ourselves as enablers for novel opportunities.”
About ten times faster
Under the umbrella of the new research focus “Human Imaging”, existing resources at DZNE were pooled and expanded. Radbruch’s own research group has been complemented by expertise in the development of MRI measurement routines and know-how in the analysis of MRI data using AI. “Presently, an MRI scan of the brain can take up to almost an hour. Many people feel uncomfortable in the narrow tube of the brain scanner and this also applies to the striking noises that accompany the examination for technical reasons. Our aim is to reduce the current measurement time to around one-tenth. This way, we want to make the examination more patient-friendly,” says Radbruch.
Mobile scanners
The researchers are taking advantage of the fact that the operation of an MRI scanner offers many possibilities: Via software, the duration, sequence and strength of the magnetic fields and radio impulses used can be adjusted in a wide variety of combinations. Yet, to generate meaningful images of the brain, complex physical phenomena must be considered. “DZNE has extensive experience in the development of such MR sequences, as these measurement routines are also called,” explains Radbruch. “Moreover, we are investigating the use of portable MRI scanners. These have only recently become available and can be moved around like an ultrasound device. Unlike conventional MRI scanners, the patient does not lie inside a narrow tube with the entire body. Only the head is surrounded by the device. Furthermore, these novel machines are cheaper to purchase and operate.”
AI-based “sharpening”
Shorter scanning times and mobile MRI scanners come with a caveat, though: lower resolution. “By optimizing the measurement protocols and post-processing techniques that we apply to the data and which make use of AI, we aim to get the maximum out of such examinations. In particular, I want to investigate whether AI can make these mobile MRI scanners achieve an image quality that is good enough to be useful in any neurological practice. Then, unlike presently, you wouldn’t have to wait as long for an appointment for an examination. That would be a significant advance.” AI is based on pattern recognition, for example with regard to the shape of specific brain areas. To acquire this ability, AI algorithms need to be trained on huge amounts of data. For this purpose, Radbruch’s team will be leveraging MRI data from clinical trials and the “Rhineland Study”, a population study run by DZNE in the Bonn area. On the other hand, the research work benefits from the close exchange with the UKB.
Medical data treasure
There is also a link to the Center for Medical Data Use and Translation (ZMDT) at the University of Bonn. “Here, research is being done from various perspectives into how medical data can be used for the best benefit of patients,” says Radbruch, who is also the director of the ZMDT. “Medical data is a treasure trove that we should utilize, but also must handle carefully. The city of Bonn has a good chance of becoming a pioneer in this field in Germany. On the one hand, we are very well positioned in medical research, and on the other hand, institutions such as the Federal Office for Information Security and the Federal Commissioner for Data Protection are located here. I would like to promote Bonn’s development into a hub for the use of medical data. Neuroimaging plays a prominent role in this, because this technology is very much data-driven.”
Cleaning processes in the spotlight
In addition, Radbruch’s team is investigating how the brain disposes of harmful metabolic products. “In healthy individuals, this so-called brain clearance happens while sleeping. That is one of the reasons why sleep is so important for our nervous system,” says the neuroscientist. “A dysfunction of these cleaning processes can lead to the accumulation of toxic substances and damage to neurons. In Alzheimer’s and other neurodegenerative diseases, certain proteins are deposited in the brain. This could be due to reduced brain clearance.” The Bonn researchers are therefore developing special MRI procedures to measure the flow of the so-called CSF in the brain, because this fluid can help to remove toxins. “My vision is to detect disorders in brain clearance at an early stage,” says Radbruch. “To this end, we are cooperating with a European consortium. The brain’s cleansing mechanisms have long been neglected. Research in this area is still in its infancy, but could contribute to better therapies in the long term. I see a lot of potential here.”