Researchers present technique for brain imaging in ‘super-resolution’

15 Jun 2021

Image: © samunella/Stock.adobe.com

Researchers have reported successful preclinical studies in enhancing resolution of PET imaging by simultaneously tracking the head movements of subjects.

Research presented at the Society of Nuclear Medicine and Molecular Imaging’s (SNMMI) 2021 virtual annual meeting has the potential to address early diagnostic limitations of brain imaging.

The SNMMI is an international scientific and medical organisation focused on advancing nuclear medicine and molecular imaging in order to hone precision medicine for diagnostics and treatment.

Positron emission tomography (PET) is one such area of research. PET works by detecting gamma rays emitted by a radionuclide. These tracers are placed on a biologically active molecule, such as sugar, which are then administered into the body.

By tracking the subsequent emissions, it is possible to see where the molecule has travelled and what activity might be taking place.

Currently, researchers are limited by participants moving during the scanning procedure. These movements reduce the resolution of the subsequent images.

By using external motion tracking devices to measure head movement during these scans, it is possible to account and compensate for these drops in resolution.

Support Silicon Republic

Importantly, researchers demonstrated that there was an increase in resolution using these combined methods, even when compared to images of a still subject. By leveraging the increased sampling information associated with the images of the moving participants, the preclinical study achieved an overall superior resolution.

Coined as ‘super-resolution’ by the researchers, the group hopes to extend their research beyond preclinical studies and involve human participants.

Looking to clinical applications, researcher Yanis Chemli, PhD candidate at the Gordon Center for Medical Imaging in Boston, highlighted neurodegenerative diseases as one domain of particular interest.

“Alzheimer’s disease is characterised by the presence of tangles composed of tau protein. These tangles start accumulating very early on in Alzheimer’s disease – sometimes decades before symptoms – in very small regions of the brain,” he said.

“The better we can image these small structures in the brain, the earlier we may be able to diagnose and, perhaps in the future, treat Alzheimer’s disease.”

Other research presented at the meeting included the use of machine learning for PET image denoising and the use of PET in examining animal models of fragile X syndrome.

Sam Cox is a journalist at Silicon Republic covering sci-tech news

editorial@siliconrepublic.com