Television shows like Star Trek feature artificially intelligent characters like Data and The Doctor who make health diagnoses using scanning devices, advanced software and data analysis. Popular science fiction writers fuel the imagination by blurring the lines between the technology of today and what the future might hold for advancements in medicine.
While we are a long, long way from achieving any real form of artificial intelligence (AI), advanced machine learning is contributing to a variety of promising software tools, scanning devices, and data analysis algorithms that are helping doctors make early diagnoses relating to dementia and Alzheimer's disease.
The use of machine learning applications by doctors and researchers in diagnosing Alzheimer's is an exciting new development in medicine that promises to help overcome the difficulties of early diagnosis, disease severity progression measurement, and new drug research.
What's at stake?
Currently, the Alzheimer's Association reports 5.7 million Americans are living with Alzheimer's or dementia. Surprisingly, Alzheimer's is the 6th leading cause of death in the U.S. To put this in perspective, Alzheimer's (or another form of dementia) kills more Americans each year than breast cancer and prostate cancer combined. The number of people with the disease is expected to rise to 13.9 million Americans by 2069 according to the CDC. In just this year, the financial impact of dementia is projected to cost $277 billion, and by 2050 the cost is expected to rise to $1.3 trillion.
Take the Alzheimer's Quiz and test your knowledge of this degenerative brain disease.
What's so encouraging about the applications of machine learning?
Alzheimer's is hard to diagnose. The difficulty for doctors is that there is no single test you can take, and brain scans alone can't determine if someone has or is likely to get the disease. Currently, doctors rely on a variety of indicators including a patient's medical history and anecdotal reports from family members and co-workers. Understanding this challenge, researchers began making real breakthroughs using advanced machine learning applications.
The use-cases describing the improved prediction of dementia fall into four primary categories: speech monitoring, medical image analysis, visual indicators, and genetic analysis. In each of these diagnosis categories, machine learning is being utilized to help doctors make diagnoses significantly sooner than traditional methods while also becoming up to 85% more accurate in their diagnoses, researchers found.
What are some examples of machine learning breakthroughs?
- Speech Monitoring: As it relates to earlier and more accurate diagnosis of dementia and Alzheimer's, companies like Tech Emergence, Winterlight Labs and Canary Speech are using advanced machine learning and algorithms to analyze speech patterns that monitor and detect the progression of dementia and Alzheimer's disease. These companies have products that scan through hundreds of variables relating to unusual word repetitions, speech pauses, word replacements and more. These variables are recorded and matched against archived datasets of the speech patterns of thousands of patients with varying degrees of the disease.
- Medical Image Analysis: Companies like Avalon AI claim they have products that use machine learning to scan brain MRIs to identify signs of brain degeneration linked to diseases like dementia and Alzheimer's. Advanced algorithms were developed after studying some 70,000 brain scan images and are now able to recognize tiny differences in visual characteristics of the brain based off of traditional scans. Research scientists from Douglas Mental Health University Institute's Translational Neuroimaging Laboratory at McGill have also developed algorithms that can identify amyloid proteins in PET scans to help predict the onset of dementia two years earlier than conventional methods.
- Visual Indicators: Research is showing that the speed, direction and patterns of eye movements may provide valuable insight into how well the brain is functioning. Based on this research, companies like Neurotrack claim to use computer vision to measure eye movement patterns that can indicate memory health. In a short five-minute test, computer vision algorithms from companies like Neurotrack can help provide key brain function data to doctors that can help them make more accurate predictions about the likelihood of patients contracting various forms of mental illnesses relating to reduced memory function.
- Genetic Analysis: Machine learning algorithms created by companies like Aequa Sciences are learning how to predict the onset of Alzheimer's by looking for potential indicators of the condition by contrasting the genetic data of healthy patients against patients with Alzheimer's. The algorithms organize the signs of Alzheimer's into neural networks to group distinct marker categories in ways that help doctors more accurately quantify the risk of developing the disease.
While we are not yet regularly traveling through space, science and medicine have advanced to the degree of harnessing one of the earliest footstones of AI, machine learning, to help doctors detect Alzheimer's earlier than ever before. Early detection is the key to improving preventative care, initial treatments, new drugs, and disease progression management. Improvements in these critical can have a beneficial impact on the collective health of our communities.
Learn how the neurologists at the Swedish Neuroscience Institute are helping ease the challenges posed by Alzheimer’s disease and other neurological disorders. Read the Swedish Neuroscience Institute blog.
Optimal Aging from Swedish connects families with trusted service providers who can assist with cooking, cleaning, errands, transportation and other services for seniors.
Subscribe to the Swedish blog and we’ll bring the health news to you.
Recommended for you:
This information is not intended as a substitute for professional medical care. Always follow your health care professional's instructions.