Simple Memory Tests Beat Expensive Scans in Predicting Alzheimer's

The study included 401 people from a long-term Alzheimer's study. The AI model analyzed data from MRI scans, genetic information, and cognitive tests (plus demographics and spinal fluid). The results were surprising: genetic data contributed the least, while cognitive test results contributed the most. Combining MRI and genetic data improved prediction over using either alone, but adding cognitive tests to any combination was only as good as using cognitive tests alone.

This means that the current clinical procedure—gathering cognitive test results—can outperform predictions that use expensive genetic or spinal fluid data. For patients and doctors, this is good news: cognitive tests are cheap, easy, and non-invasive. They can be done in a regular office visit. Expensive tests like spinal taps or genetic screening may not be necessary for predicting Alzheimer's progression.

The study suggests that AI can help make better use of simple, widely available tests. This could lead to earlier interventions and better planning for patients and families. However, the study is based on a specific group of people and more research is needed to confirm the findings in broader populations. Still, it's a promising step toward affordable, accessible Alzheimer's care.