AI and Alzheimer’s: A Tale of Two Brains

September 29th, 2021. By Audrey Tam '22

Why STEM? It allows us to improve our quality of life. A brief examination of a new Alzheimer’s identification algorithm reveals how applied science helps us both prevent and mitigate today’s problems, and as a result, reduce human suffering.

As the world’s most frequent cause of dementia, early diagnosis is critical in treating Alzheimer’s. Timely intervention not only delays the onset of cognitive impairment, but also enables patients to gain access to information, plan for the future, and benefit from supportive treatments.

However, the process of diagnosis presents some complications. Identifying signs of onset dementia proves difficult because mild cognitive impairment, or MCI, doesn’t leave any physical traces. Moreover, MCI doesn’t always correlate with Alzheimer’s disease. Simply put, the complexities of the brain and its changes make detection difficult.

Medical professionals currently use a variety of methods to identify signs of Alzheimer’s, including the use of eye-tracking and voice analysis technology. But no system of detection is more promising than the use of artificial intelligence, which provides the crucial benefit of speed. With access to vast databases and the capability to progressively learn and apply information, neural networks diagnose diseases much quicker than we ever could.

How exactly does this look and work in practice? In most methods, researchers collect data from functional magnetic resonance image (fMRI) brain scans. Machines then process these images before feeding them into a deep learning-based algorithm. This algorithm is often a convolutional neural network, a type of neural network commonly used for image identification. The system finally splits the fMRI image input into six categories, ranging from cognitively normal to Alzheimer’s disease.

The implementation of artificial intelligence as detection programs is not a novel practice, but researchers have developed a new algorithm with near 100% accuracy in detecting potential indicators of Alzheimer’s. To improve its reliability, over 78,00 fMRI scans were analyzed to train and validate this proposed model. Even more exciting, this model “performed better than other known models in terms of accuracy, sensitivity, and specificity” (Odusami).

While the title of this article states that this case study tells a tale of two brains, in actuality, it shares the story of three. Research illustrates the fragility of the human brain under Alzheimer’s and the power of artificial brains when put into action, but more importantly, it displays the extent of our curiosity and innovation.

Despite our innate susceptibility to cognitive decline and limitations to speedy identification, we humans create the tools necessary to progress. Through STEM, we forge for ourselves the ability to overcome challenges and step by step, pave the path towards a better future for all.

Works Cited:

ADI - Importance of Early Diagnosis. Accessed 18 Sept. 2021.


Nield, David. “New Algorithm Can Identify Pre-Alzheimer’s Brain Changes With Over 99% Accuracy.” ScienceAlert, Accessed 16 Sept. 2021.

Link: New Algorithm Can Identify Pre-Alzheimer's Brain Changes With Over 99% Accuracy

Odusami, Modupe et al. “Analysis of Features of Alzheimer’s Disease: Detection of Early Stage from Functional Brain Changes in Magnetic Resonance Images Using a Finetuned ResNet18 Network.” Diagnostics 11.6 (2021): 1071. Crossref. Web.

Link: Analysis of Features of Alzheimer’s Disease

Rasmussen, Jill, and Haya Langerman. “Alzheimer's Disease - Why We Need Early Diagnosis.” Degenerative neurological and neuromuscular disease vol. 9 123-130. 24 Dec. 2019, doi:10.2147/DNND.S228939

Link: Alzheimer's Disease – Why We Need Early Diagnosis