How Neuroswipe Works
Medical researchers and physicians collect large volumes of data from people to diagnose conditions and advance research. Analyzing all of that medical data takes significant effort. Until now, the process of assessing medical scan image quality was a bottleneck that slowed down research and medical care.
Neuroswipe may change how brain scans and related medical data are processed. In a large scale study, researchers may generate thousands of images. Unfortunately, some of these images will be of low quality. That's where the app comes to play. Users can swipe to the right if the scan looks good or swipe to the left to reject an image based on a few quality criteria.
Knowing What to Look for in Brain Scan Images
Volunteer users do not need to understand the brain inside out to use the app. Instead, the researchers decided to focus on a single structure. Since Alzheimer's disease tends to have a significant adverse effect on memory, app users are asked to focus on the fornix. This brain structure is considered to be critical in the creation of new memories.
Improving the Quality of Alzheimer's Research
By focusing on this brain structure in scans, app users can quickly judge the scan's quality. Over time, the app trains users to spot which scans have problems. As a result, brain scan users will become more proficient in the task.
AI Isn't Always the Answer
A large amount of artificial intelligence (AI) and machine learning development has focused on image processing in recent years. Despite that development, researchers have found that specific tasks are not a good fit for AI. Today, filtering out low-quality scans is not a task that AI can do effectively.
Support Brain Research From Home
Volunteering for medical research traditionally involved visiting a lab and interacting with researchers in person. Due to the COVID-19 pandemic, in-person medical research is becoming more challenging to administer. That's why Neuroswipe is so significant. Volunteers can contribute to supporting medical research from home.
Even better, volunteer involvement does require a large amount of time. A volunteer can assess one, a few, or many images. This level of flexibility means that people stuck at home with little to do thanks to COVID-19 can contribute to medical research.
Available as a web app, Neuroswipe starts by training users to distinguish between three types of images. The first set of images are ideal quality (i.e., "What a Good Fornix Looks Like") — this is what researchers aim for in their studies. The next tier of image quality could be called border-line quality (i.e., "Not perfect-looking, but OK"), according to the website. Finally, the researchers also have provided examples of images that have to be rejected.
Saving Money With Volunteers
In almost all scientific research efforts, there are different kinds of tasks. Senior scientists direct the project and oversee the work of more junior scientists. In some cases, technicians execute the details of specific experiments or gather evidence. Asking highly trained scientists to spend hours assessing image quality is wasteful.
According to Salary.com data, a medical research scientist's median salary is $95,000 in the United States. That means that an hour spent reviewing images for quality effectively costs close to $50. That's just part of the cost. Spending time on image quality review means less time and energy are available for developing creative ideas.
Other Remote Medical Research Projects
Leveraging volunteer contributions at a distance is not a new idea in the medical research world. Stanford University's Folding @ Home project also uses spare computing capacity from volunteers. Like the Cardiff researchers, Folding @ Home focuses on the challenge of Alzheimer's disease.
In 2020, the Stanford University group has redirected their resources to contribute to COVID-19 research. Computer owners can download the Stanford University app and download unused computing capacity to support the program. Folding @ Home originally started close to twenty years ago. This distributed computing approach was an alternative to much more expensive computing options, like using a supercomputer.
About the Author
Bruce Harpham is an author and marketing consultant based in Canada. His first book "Project Managers At Work" shared real-world success lessons from NASA, Google, and other organizations. His articles have been published in CIO.com, InfoWorld, Canadian Business, and other organizations. Visit BruceHarpham.com for articles, interviews with tech leaders, and updates on future books.