Computer Vision News - March 2023

15 Duygu Tosun-Turgut imaging and functional MR imaging to identify the underlying pathophysiological changes that might be related to their symptoms. We're actually slowly seeing it come to clinical practice and clinical trials, not just seeing research ideas or research findings. But some of my work is actually on alternatives to PET imaging, which is extremely expensive. It costs around $5,000 to $7,000 for imaging modality, even in resource settings. It's difficult to find those imaging modalities in every clinical setting. So, we've been focusing on some other imaging approaches: not molecularly specific, but still might serve as a surrogate marker of these functional and structural changes in the brain, like MRI imaging, and developing what I call computational biomarkers. It's not a true pathological or molecular biomarker, but it captures how the brain has been affected over the years or how the brain is affected by the insult of these pathologies or syndromes. We interpret, or we accessible and scalable for everyone, especially in Alzheimer's disease. Not like having state-of-the-art biomarkers, which require special equipment or expensive imaging technology, so it can only be accessed by certain countries or by certain specialized centers. Biomarkers have been playing a huge role in us defining what is really Alzheimer's disease and what kind of pathological changes are happening. In the past ten years, my research has been dedicated to developing computational approaches to discover imaging biomarkers that can be used in clinical trials to identify participants and also to track how they're responding to treatment. Does this mean that we can detect Alzheimer's disease and chronical depression in medical images or in blood test? Both things are happening right now. We do have very special molecular imaging approaches. These are positron emission tomography (PET) approaches where we have specially engineered or designed radiotracers that specifically bind pathologies or protein changes in the brain. We've been using that for Alzheimer's disease as well as depression. The way we conceptualize Alzheimer's disease is as a proteinopathy, meaning there are abnormal protein accumulations. And these molecular imaging modalities allow us to detect and see how these proteins accumulate in individuals while they're alive before autopsy. We can identify these individuals veryearly, evenbefore they start having a cognitive impairment or cognitive decline. Similarly, depression, for example, is conceptualized as a synaptic dysfunction in particular brain regions. We've been using particularly glucose metabolism PET ... science is definitely more accessible, which is great!