Artificial Intelligence and Alzheimer’s Research

Artificial intelligence (AI) is poised to play an important role in the diagnosis and treatment of Alzheimer’s disease and other dementias.

In general, AI refers to computer programs designed to learn and/or perform tasks in a manner similar to the human mind. The field has now advanced to the point that AI is becoming increasingly capable of solving problems in ways that researchers cannot predict. This means that it can potentially provide solutions to long-standing medical mysteries, like the origin of Alzheimer’s disease.

Why Do We Need AI?

Society has a fixation on the potential dangers of AI (thanks, Terminator). While there are certainly some legitimate concerns about how things can go wrong, we would be remiss to ignore the benefits that AI can offer. Plus, since we are so aware of the apparent threat of an AI-based apocalypse, it is likely that many safeguards are being put in place by developers (fingers crossed).

With the doom and gloom out of the way, we can now look at the bright side of AI.

The fact that we can design computer programs to solve problems and complete tasks is exciting alone, but the most astounding development is the evolution of AI systems that can outperform the human mind. Some famous examples include the supercomputer Deep Blue beating world chess champion Garry Kasparov back in 1996, and Watson demolishing “Jeopardy!” legend Ken Jennings in 2008.

Far from the public eye of computer vs celebrity battles, researchers have been working on ways to use the superior capabilities of AI to battle complex medical conditions like Alzheimer’s disease.

AI Could Break the Alzheimer’s Code

Alzheimer’s disease is a particularly troublesome illness in many ways. Its symptoms are severe and irreversible. From a research standpoint, there has been little to no success in preventing, stopping, or even slowing its development. Diagnosis efforts have also been frustratingly stifled.

Currently, the best options for detecting Alzheimer’s disease early are cognitive tests, like the SAGE exam on which the BrainTest® app is based. There are no reliable biological markers to aid in an early Alzheimer’s diagnosis, so it cannot be detected by a blood test, brain imaging, or other existing medical methods before the symptoms have taken hold.

The barriers that are preventing some of the brightest human minds from reliably addressing Alzheimer’s disease could soon be overcome by AI. Studies are already starting to emerge demonstrating how it can be used to diagnose Alzheimer’s disease up to several years earlier than is currently possible.

Some newer AI methods focus on the cognitive symptoms that are currently used for diagnosing Alzheimer’s, while others have been able to deliver accurate diagnoses using novel approaches, like analyzing speech patterns and eye movements. It will take time and effort to replicate these results on a large scale, but the early findings are highly encouraging.

Digging Deeper

A cutting-edge application of AI is a collection of techniques known as deep learning. Other “traditional” AI programs can usually only be improved by making changes to the program itself. Deep learning methods, on the other hand, are designed in a way that allows them to continue to improve as more data is added.

For example, a deep learning program that is trained to diagnose Alzheimer’s based on proteins in the blood will continue to get better at its task as long is it is provided with a constantly growing database of blood test results from people who develop the condition.

With deep learning currently being applied to many questions related to Alzheimer’s and dementia, it is only a matter of time until we witness some major breakthroughs.