A Machine Learning-Based Early Diagnosis of Alzheimer’s Disease
Alzheimer’s disease (AD) is one of the most common neurodegenerative diseases in the world. Though there is no cure, an early detection of AD (when signs of mild cognitive impairment (MCI) are shown) may be critical to allow for treatment that may further prolong the patient’s brain function. In this study, Jack presents a generalized classification model using a three-dimensional DenseNet architecture for AD diagnosis through whole-brain magnetic resonance imaging. He then evaluated his classifier on two different datasets to assess both its performance as well as the validity of MCI as a precursor stage for AD. Jack found that his model achieved high accuracies upwards of 80%, indicating that the classifier succeeded in providing a preliminary diagnosis of AD from magnetic resonance imaging.
Jack wrote a paper that was published in the Columbia Junior Science Journal. He also created a website showcasing his project along with more information/graphs on the dataset he used.Visit his website!
Jack is a 17 year-old high schooler from Cypress, TX.
Zach was quite possibly the best teacher/mentor I've had so far, at least in how much he was able to teach me in the short period of time that a Polygence project lasts. His explanations are almost always accompanied by visualizations/drawings that he makes during the meetings (which helps a lot with comprehending the topics), and he makes sure that we actually understand what he's talking about. Zach is also very knowledgeable about these topics and was able to answer all the questions that I had during the meetings. I started out knowing very little about machine learning, neurodegenerative diseases, or even Python in general. But with Zach's guidance and all the resources that he spent his time to compile for me, I was able to learn so much about these vast topics, to the point where I'm now comfortable writing a full research paper on what I know about these topics.