📄 At Home Alzheimer’s Detection Takes a Leap Forward
Imagine checking your brain health the way you check your blood pressure or sleep patterns—quickly, privately, and from the comfort of your own home. That vision is no longer futuristic speculation but the driving force behind PREDICTOM, a major European study that is quietly reshaping how Alzheimer’s disease might be detected long before symptoms appear.
Photo: Colourbox
At the heart of the project lies a bold scientific idea: bring early detection into people’s homes using AI and multimodal biomarkers, transforming what used to be a hospital bound process into a scalable, accessible digital pathway. The newly published methodology of the PREDICTOM study offers a rare look under the hood of this innovation.
Turning the Home Into a Testing Room
The first stage of the study asks more than 4,000 adults to complete a fully home based set of assessments. These include short memory and thinking tasks, a hearing check, and a brief eye tracking activity using an ordinary webcam. Participants also collect saliva and a tiny drop of blood from a finger prick—no clinic visit needed.
That small blood sample can be analyzed for p-tau217, one of the most promising early markers of Alzheimer’s related changes. This is a remarkable shift: a test that used to require specialized staff and equipment can now begin at someone’s kitchen table.
AI based stratification
Once all the at home information is gathered, the PREDICTOM’s platform stratifies participants based on the likelihood they are developing Alzheimer’s pathology. The developed algorithm considers digital measures, cognitive performance, lifestyle risk factors, self reported symptoms, and biological markers collected remotely. Instead of sending everyone to clinics for expensive or time consuming procedures, only those who appear most likely to benefit move on to in person assessments such as brain scans or detailed laboratory tests.
This practical approach addresses a growing challenge worldwide: traditional health systems cannot screen entire populations for Alzheimer’s. But a simple first step at home could make early detection more scalable and more realistic than ever before.
See figure 1. Patient information including demographics, medical history, and lifestyle factors is processed through a predictive model to generate individualized time-dependent risk trajectories for developing dementia. Risk factors are taken from the Lancet Commission on dementia prevention.
Figure 1. Schematic representation of a machine learning approach for predicting time-dependent dementia risk using routine primary care data
More Than One Type of Clue
What makes PREDICTOM stand out is how many different types of information it combines. The study looks at:
performance on digital cognitive tasks,
eye movement patterns,
hearing ability,
lifestyle and health questions,
saliva and finger stick blood samples, and
follow up clinical measures like MRI or EEG for those who need them.
By combining established and novel aspirational biomarkers, the study aims to build multidimensional “signatures” of early Alzheimer’s pathology. The paper outlines how machine learning models will be trained to identify patterns within this multimodal dataset—an approach that could push diagnosis far earlier than today's methods allow.
Feasibility—People Can Really Do This at Home
A major question was whether people would manage these tasks on their own. The answer so far is encouraging.
Of the more than 2,032 people screened early in the study (Read PREDICTOM Study Progress for consented recruitement), 630 completed the entire home testing package, a strong sign that the process is usable for everyday adults. The dropout rate was just 6%, unusually low for a remote study involving several steps and biological samples.
This shows that large scale, at home brain health testing isn’t just possible—it’s already happening.
Redefining the Start Line of Alzheimer’s Care
Early detection matters. Alzheimer’s disease begins many years before memory problems arise, leaving a long window where changes are quietly unfolding. The PREDICTOM approach opens a door to catching those changes early—before opportunities for prevention or intervention slip away.
The vision is simple: one day, checking your brain health could be as routine as monitoring your heart rate or blood sugar. A few online tasks. A quick sample from your fingertip. And potentially, years of extra time to intervene, plan, and adapt.
