Researchers at Cardiff University have borrowed ideas from popular dating apps (including Tinder) to train volunteers to find low-quality brain scans. The latest technology, which is being tested by scientists to improve the overall quality of brain imaging in research, will help in the study of diseases such as dementia, according to the BBC.
Researchers at Cardiff University borrowed ideas from popular dating apps to train volunteers to identify poor-quality brain scans. With the new application, thousands of scans from a study can be analyzed much faster, which will help a number of studies.
“The human eye is extremely sensitive to subtle differences in size, shape, color and appearance and this can be useful in research,” explains Dr. Judith Harrison of Cardiff University’s Center for Brain Imaging – Cubric. “Knowing that most people view different images from day to day, for example when using dating apps, we wondered if the same principle could be applied to filtering brain scans.”
The sheer scale of the various studies means scientists have to take tens of thousands of scans – and not all of them are good. In the same way that we can take a bad picture, if our hands are shaking or a finger gets caught in part of the lens, patients can move in the scanning machine at the most inopportune moment. The effect is the same – the image is of little use. It can take an individual examiner several hours to review each scanned image and decide if it is good enough for research use.
This is where a new app developed by Cardiff University and the National Academy of Software comes to the rescue. Using real-life images of the brains of Alzheimer’s patients, the app trains volunteers to recognize the correct shape of the fornix in the scanned image.
The app is called Neuroswipe and is still under development. So far, the project is only available as a web application.
“In recent years, the process of filtering images has been largely automated, but training an AI program to detect poor quality scans is difficult,” explains Dr. Harrison. “We want to see if people can help us weed out some of the images that haven’t been processed correctly and become scientists of their own kind.”
The application can be expanded for use in large-scale studies involving thousands of patients. Thousands of scans from the survey can be quickly analyzed.