Scientists are working on a machine learning (ML) model that could potentially make early detection of Alzheimer’s dementia possible through a simple smartphone tool.
This tool would be able to distinguish Alzheimer’s patients from healthy individuals with a 70-75% accuracy rate, a promising figure considering more than 747,000 Canadians suffer from Alzheimer’s or some other form of dementia.
Problem
Detecting Alzheimer’s dementia in the early stages can be challenging, as initial symptoms are often subtle and can be mistaken for typical age-related memory issues.
Traditionally, detecting brain changes indicative of Alzheimer’s has required expensive lab work and medical imaging, which is not generally carried out at the early stages of the disease.
Solution
The machine learning model in development could be incorporated into a mobile app, providing a convenient early indicator of potential Alzheimer’s dementia.
The ability to begin treatment and interventions earlier could potentially slow the disease’s progression.
Furthermore, a mobile screening tool would offer a convenient telehealth solution for those who may have geographical or linguistic barriers to accessing healthcare services.
How It Works
The model doesn’t focus on specific words but rather examines language-agnostic acoustic and linguistic speech features.
For instance, it considers that Alzheimer’s patients typically speak more slowly, with more pauses or disruptions, and often have reduced intelligibility in their speech.
The model can screen for these speech features, potentially transcending language boundaries and having universal applicability.
Next Steps
The user experience with the developed tool would be straightforward: a person talks into it, analyzes the speech, and predicts whether or not the person has Alzheimer’s.
The resulting information could then be taken to a healthcare professional for further investigation and action.
The team behind this tool, the computational psychiatry research group at the University of Alberta, has also developed similar AI models and tools for detecting psychiatric disorders such as PTSD, schizophrenia, depression, and bipolar disorder.
They believe that any technological advancement that can assist in managing diseases earlier and at a lower cost is beneficial.
The team’s machine learning model is detailed in a paper set to appear in the ICASSP 2023 Signal Processing Grand Challenge, where they ranked first in North America and fourth globally.
If you care about brain health, please read studies about vitamin D deficiency linked to Alzheimer’s and vascular dementia, and higher magnesium intake could help benefit brain health.
For more information about brain health, please see recent studies about antioxidants that could help reduce dementia risk, and coconut oil could help improve cognitive function in Alzheimer’s.
The study was published in ICASSP 2023—2023 IEEE International Conference on Acoustics, Speech and Signal Processing.
Copyright © 2023 Knowridge Science Report. All rights reserved.
