New in OS X: Get MacRumors Push Notifications on your Mac

Resubscribe Now Close

New Study Aims to Determine Whether iPhone and Apple Watch Can Detect Early Signs of Dementia

Wednesday August 7, 2019 2:43 pm PDT by Juli Clover
Apple has partnered with pharmaceutical company Eli Lilly and health startup Evidation to determine whether data collected from the iPhone and Apple Watch can be used to detect early signs of dementia.

A research paper published this week and shared by CNBC lists researchers from Eli Lilly, Apple, and Evidation Health. The paper, called "Developing Measures of Cognitive Impairment in the Real World from Consumer-Grade Multimodal Sensor Streams," explores whether sensor data and activity info from smart watch devices can be mined for "physiological and behavior signatures of cognitive impairment."

The ubiquity and remarkable technological progress of wearable consumer devices and mobile-computing platforms (smart phone, smart watch, tablet), along with the multitude of sensor modalities available, have enabled continuous monitoring of patients and their daily activities. Such rich, longitudinal information can be mined for physiological and behavioral signatures of cognitive impairment and provide new avenues for detecting MCI in a timely and cost-effective manner.

In this work, we present a platform for remote and unobtrusive monitoring of symptoms related to cognitive impairment using several consumer-grade smart devices. We demonstrate how the platform has been used to collect a total of 16TB of data during the Lilly Exploratory Digital Assessment Study, a 12-week feasibility study which monitored 31 people with cognitive impairment and 82 without cognitive impairment in free living conditions. We describe how careful data unification, time-alignment, and imputation techniques can handle missing data rates inherent in real-world settings and ultimately show utility of these disparate data in differentiating symptomatics from healthy controls based on features computed purely from device data.
According to the abstract, 31 people with cognitive impairment and 82 without cognitive impairment were monitored over a 12-week period, with 16TB of data collected. The study claims that the data was able to be used to differentiate people with early signs of cognitive impairment from those who were healthy.

People who had symptoms of cognitive decline typed more slowly, typed less regularly, relied more heavily on support apps, and sent fewer text messages. The study did not reach long-term conclusions as more analysis is needed.

In a statement to CNBC, Evidation co-founder Christine Lemke said that data collected from the iPhone, ‌Apple Watch‌, and Beddit sleep monitors was used for the study. Apple acquired the company behind the Beddit sleep monitor back in 2017.
"With this research, we looked at how everyday behavior data, such as those captured by iPhones, Apple Watches, and Beddit sleep monitors, may be effective in differentiating between individuals with mild cognitive impairment and early Alzheimer's disease, and those without symptoms."
Early detection of dementia is important because an early diagnosis can allow for better management of symptoms and quality of life improvements even though the progression of the disease can't be stopped. According to the World Health Organization, 50 million people around the world have dementia, with close to 10 million new cases surfacing every year.

[ 66 comments ]

Top Rated Comments

(View all)

Avatar
Relentless Power
10 weeks ago
Crazy. I mean, to see the advancements that technology is making in the wearables sector is incredible (Which also explains it’s growth). EKG, fall detection, wheelchair capabilities on the Apple Watch and possibly detecting dementia, this is what’s making the future for wearables a ‘must have’ versus what once started primarily as a notification device.
Rating: 21 Votes
Avatar
BootsWalking
10 weeks ago
One way to detect dementia is by counting the number of times an owner uses the 'Find My Phone/Watch' feature every week :)
Rating: 12 Votes
Avatar
Neepman
10 weeks ago
Boy. Talk about a slippy slope. Ring the right algorithm bell and be in front of the death panel the next day.
Rating: 6 Votes
Avatar
Edsel
10 weeks ago
My mother had dementia and it is a weird disease. They can be perfectly normal one minute then think my wife has of 30 years was my girlfriend, my mother thought I was my Father. She stormed away angry that I had a “girlfriend”.

So, as an oldster now, I worry that I might get dementia. There are moments when I find myself thinking about the past. Will I get dementia? I will never know but my family will. Godspeed to the scientists and medical professionals for trying to root out this affliction that effects so many.
Rating: 5 Votes
Avatar
Naraxus
10 weeks ago
This would be amazing. Dementia is a horrible horrible disease that takes every last bit of dignity from you. The earlier the detection, the earlier the planning can begin or even cure it.
Rating: 4 Votes
Avatar
Suttree
10 weeks ago
I know nothing about medicine and have zero medical training and knowledge, yet even I know this study is bunk. My mom asked me why am I still living at home, depriving this world of my sage knowledge and skillset when I can be goodly remunerated and leave a positive impact rather than merely armchair quarterbacking. But it is obvious this study is plain bunk with only 31 subjects. Common sense. Simple as that.
Rating: 4 Votes
Avatar
GrumpyMom
10 weeks ago

One way to detect dementia is by counting the number of times an owner uses the 'Find My Phone/Watch' feature every week :)

Oh crap.
Rating: 3 Votes
Avatar
The Cappy
10 weeks ago

One way to detect dementia is by counting the number of times an owner uses the 'Find My Phone/Watch' feature every week :)

Crazy. I mean, to see the advancements that technology is making in the wearables sector is incredible (Which also explains it’s growth). EKG, fall detection, wheelchair capabilities on the Apple Watch and possibly detecting dementia, this is what’s making the future for wearables a ‘must have’ versus what once started primarily as a notification device.

Because...? If it said you were going to develop Alzheimer's, we've just ruined your life for no good reason. It's not like we have preventative medicines. The only thing you might do is what you're supposed to do anyway which is to be as healthy as possible in other respects so you are more resilient when the dementia gets its teeth into you.

But what will happen instead is that, armed with this cognitive death sentence, far more people will greet the early dementia with fatalism and surrender. This is why we shouldn't (and generally don't) do DNA tests on family members of Alzheimer's patients to see if they're at high risk.
Rating: 3 Votes
Avatar
69Mustang
10 weeks ago
{reads headline aloud}
"New... study... aims... to determine... whether... iPhone... and Apple Watch... can... detect... early... signs of... dementia"
-Aide

/paces nervously
/hurriedly picks up phone

Calls for missle strike on Apple HQ [via Twitter]
-Aide's boss
:p:D
Rating: 3 Votes
Avatar
macduke
10 weeks ago
Did they control for age? Because if not, congrats, you determined that those people are old, and like 99% of people with dementia are old.

Edit: Ok, this part from the paper gives me pause:

Even so, due to difficulties with recruiting symptomatic participants, the symptomatic cohort was an average of 3 years older than the healthy control cohort.


Ok, it looks like they did match ages, but not genders, and the sample size is kinda small.


In order to verify that the device-derived features were not de- tecting differences in behavior due to normal aging, we selected the nearest age-matched control within the healthy control cohort for every participant in the symptomatic cohort. Doing so produced two age-matched cohorts of 31 symptomatics and 31 healthy controls with an average age difference of less than six months2. We re-ran the full analysis on these age-matched cohorts and report the results in Table 3. After controlling for age via matching, there was a large drop in performance for the demographics only models (AUROC=0.519, AUPRC=0.536). However, device-derived features still show moderate performance, with AUROC decreasing from 0.771 to 0.726 on the full cohort, and AUPRC increasing from 0.628 to 0.758.


Seems like it still correlates, but not as strongly, when controlling for age? Hmmm. Don't have time to read the full thing in-depth but would love to hear from someone who did.
Rating: 3 Votes

[ Read All Comments ]