Early Signs of Dementia Can Be Detected By Tracking Driving Behaviors (newatlas.com) 93
An anonymous reader quotes a report from New Atlas: A fascinating new study from a team of US researchers has used machine learning techniques to develop algorithms that can analyze naturalistic driving data and detect mild cognitive impairment and dementia in a driver. The work is still in the preliminary stages, however, the researchers claim it could be possible in the future to detect early signs of dementia using either a smartphone app or devices incorporated into car software systems. The research utilized data from a novel long-term study called LongROAD (The Longitudinal Research on Aging Drivers), which tracked nearly 3,000 older drivers for up to four years, offering a large longitudinal dataset.
Over the course of the LongROAD study, 33 subjects were diagnosed with MCI and 31 with dementia. A series of machine learning models were trained on the LongROAD data, tasked with detecting MCI and dementia from driving behaviors. "Based on variables derived from the naturalistic driving data and basic demographic characteristics, such as age, sex, race/ethnicity and education level, we could predict mild cognitive impairment and dementia with 88 percent accuracy," says Sharon Di, lead author on the new study. Although age was the number one factor for detecting MCI or dementia, a number of driving variables closely followed. These include, "the percentage of trips traveled within 15 miles (24 km) of home ... the length of trips starting and ending at home, minutes per trip, and number of hard braking events with deceleration rates 0.35 g." Using driving variables alone, the models could still predict those MCI or dementia drivers with 66 percent accuracy. The new study was published in the journal Geriatrics.
Over the course of the LongROAD study, 33 subjects were diagnosed with MCI and 31 with dementia. A series of machine learning models were trained on the LongROAD data, tasked with detecting MCI and dementia from driving behaviors. "Based on variables derived from the naturalistic driving data and basic demographic characteristics, such as age, sex, race/ethnicity and education level, we could predict mild cognitive impairment and dementia with 88 percent accuracy," says Sharon Di, lead author on the new study. Although age was the number one factor for detecting MCI or dementia, a number of driving variables closely followed. These include, "the percentage of trips traveled within 15 miles (24 km) of home ... the length of trips starting and ending at home, minutes per trip, and number of hard braking events with deceleration rates 0.35 g." Using driving variables alone, the models could still predict those MCI or dementia drivers with 66 percent accuracy. The new study was published in the journal Geriatrics.
Lots of dementia coming (Score:4, Insightful)
and number of hard braking events with deceleration rates 0.35 g.
Considering the number of people in their 20s, 30s, and 40s I see every day slamming on the brakes at the last second, there's a whole bunch of folks who are going to get dementia.
Either that or they're looking at their phone and not seeing traffic stopped.
Re: (Score:2)
Young people seem to drive their cars as if real life was freakin' GTA5.
Re: (Score:2)
Young people seem to drive their cars as if real life was freakin' GTA5.
Ahh, this would explain why I see so many with dangling bumpers.
Re: (Score:2)
My dad, first time in a car with me driving: "You drive well, but you also drive like it's a computer game."
Yeah, I hit the apex and accelerate hard out of the corners. What, that's wrong?
Re: (Score:3, Insightful)
Considering the number of people in their 20s, 30s, and 40s I see every day slamming on the brakes at the last second, there's a whole bunch of folks who are going to get dementia.
If the hard braking goes along with hard acceleration you are probably safe (from Dementia at least). It's the snail like take off along with hard braking that would give it away.
Re: (Score:2)
a few years ago, we had a monitor on my wife's van--after a minor accident, it would save us over $1k/year if she passed.
So I ended up being emailed every event.
*Lots* of "sudden braking" reports--and practically all while dropping off our kids in the high school parking lot . . .
Re:Lots of dementia coming (Score:5, Informative)
and number of hard braking events with deceleration rates 0.35 g.
Everyone is getting this backwards. It's a badly written summary.
People with early MCI/dementia have fewer hard braking incidents.
From the study. https://www.mdpi.com/2308-3417... [mdpi.com]
Reported changes in older drivers with preclinical AD or early-stage dementia include declines in driving performance, such as increased incidence of getting lost in traffic [11], increased risk of failing a driving test [5,6] and reduced spatial navigation ability [13], and atypical driving behaviors, such as decreased driving exposure (e.g., fewer driving trips, driving days, driving destinations, nighttime driving and rush-hour driving) [7], restricted driving space (e.g., less freeway driving and more driving within 5–10 miles of home) [7,11], and reduced unsafe driving behaviors (e.g., fewer hard braking events and speeding events)
Drilling down in the references also shows that elderly people that have impairment from medication problems do have increased hard braking incidents.
Re: (Score:2)
I fear lockdowns will have compromised the usefulness of this data. Lots of people barely driving anywhere, and when you're just popping down to the local shops you're probably less likely to need "unsafe" driving events.
Which is another thing. Hard braking isn't an unsafe driving event. Going around that blind bend at 80 is an unsafe driving event.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Hmm, it sounds like they are simply detecting people who are retired
Exactly. Frequent short trips and rare long trips on
highways is called "not having to commute anymore".
Re: (Score:2)
Either that or they're looking at their phone and not seeing traffic stopped.
Popularity of dementia vs. personal distraction devices.
And that's not dementia. It's mass ignorance. Milder, but smells like shit and stupidity.
Detection is good, but... (Score:2)
...are we hopefully coming up with preventative measures? (I mean other than looking for a semi-unreasonable excuse to take someone's drivers' license away prematurely.)
It's one thing to be told you've got, or are going to have, dementia.
It's another thing entirely to be told there's not a damn thing that can be done about it. To the point of questioning the value of even knowing about a diagnosis like that.
Re: (Score:2)
Better to know why something is happening even if you can't do anything about it.
Re: (Score:2)
Better to know why something is happening even if you can't do anything about it.
Sure. There's value in that. For pretty much everyone except you.
For you, it could be anything but good news, which could be quite detrimental for the life you have remaining. That's kind of my point.
Re: (Score:2)
Or, you could change what your were planning to do with your life, now that you know you aren't going to have 25 years of 'good retirement' but you might only get 5. You can accelerate your plans to go on round the world holidays, spend more time with your family, etc.
My uncle died shortly after he retired from parkinsons, which progressed fairly rapidly. The hospital and doctors botched his diagnosis, robbing him of 3 years of quality life from his initial symptoms that he could have spent productively. II
Re: (Score:2)
Or, you could change what your were planning to do with your life, now that you know you aren't going to have 25 years of 'good retirement' but you might only get 5. You can accelerate your plans to go on round the world holidays, spend more time with your family, etc.
About the only positive thing that could come out of knowing that you were going to become a demented Alzheimer's patient, would be to set yourself up with a system to keep the healthcare system away from your money.
You'd divorce your SO if married, and give them as absolutely much of your estate as you could. You'd have a plan where you'd live with them as a pauper, so that when the medical system took you into their loving embrace, you'd have nothing for them to take.
The present system is a nightmare
Re: (Score:2)
Re: (Score:2)
The present system is a nightmare of keeping you alive long enough to drain your estate, then after a decade of life not knowing who you are, living in diapers, you'll finally expire, your SO might be bankrupt. That's a cruel system that stretches out death for as long a time as possible.
Much of what you said I agree with, but assuming you are talking about the US system there are some rules to prevent the spouse from going bankrupt. There's also some planning you can do before dementia gets advanced. I just went through this with my parents. Most of this has to do with getting enrolled in Medicaid. You'll need to spend down assets to qualify but things like a home owned as joint tenants, IRAs, and assets in the spouse's name only are not counted. Income such as social security and pensions mostly go the the spouse to live off of. The planning part involves moving countable assets out of the patients name. You need to do this really early as Medicaid will look back 5 years and claw those assets back. It's a complicated matter unfortunately.
It definitely is complicated. And it's a hella thing to end up saving and investing for your family just to have the nursing home get most of it. One of the most shocking things is the new drugs that stretch out dementia as long as possible. My mother in law went on these - it probably gave her a few extra years of her body being alive. Who knows where her mind was? Cruel AF.
Me? I'm not suicidal, but seeing how the system works, I'll check out manually if at all possible if it looks like I'm catching dem
Re: (Score:2)
Re: (Score:2)
I hear you. My Dad's Alzheimer's took 10 years to run its course with the last 18 months spent in a 'memory care' facility. Nearly killed my mom caring for him and worrying about their finances. The only relief came when hospice took over. They ditched the drugs and only gave him medication to make him comfortable.
Ugh - memory care for those who have none. Seems like Newspeak from 1984. It's torture for the family, as the inevitible is drawn out so much longer. When I was a kid, and people got "hardening of the arteries", the end was a lot quicker. It was sad and no doubt, but having had elders who took years to die, as opposed to my mother, who had a massive heart attack and was gone in a few minutes, I know what my choice would be. I mean - she beat the system.
Re: (Score:2)
Better to know why something is happening even if you can't do anything about it.
Sure. There's value in that. For pretty much everyone except you.
For you, it could be anything but good news, which could be quite detrimental for the life you have remaining. That's kind of my point.
This is why people who are at risk of invariably fatal diseases like Huntingdon's, ALS, and others often don't test for it. If a driving test showed I was going to sped the last 10 years of my life in depends, not knowing who I was except on the 1 good day maybe every couple months, while the "medical system" drainse my estate, I would be planning my date with the second amendment treatment plan - cures everything.
I wouldn't be too surprised if suicides increased a hundredfold. And the present day medica
Re: (Score:2)
Exactly.
I'd want to know, especially if there was nothing that could be done.
Re: (Score:2)
...are we hopefully coming up with preventative measures?
Dementia costs the US economy more than $300B each year.
It is not just the medical costs but also the cost of young people dropping out of the workforce to care for older relatives. As boomers age and smaller generations follow, this problem will worsen.
We should be spending far more on research to delay, prevent, and cure dementia. Economically, it is a no-brainer.
Re: (Score:1)
it is a no-brainer.
You forgot to say whether the pun was intended.
Re: (Score:2)
Economically, it is a no-brainer.
It's funny you mention this, and not just for the twisted humor.
Regarding the economics, I've been far too jaded by the corruption in the Medical Industrial Complex in the US to assume that these kinds of studies would not be abused as leverage to identify and start forcing certain societal restrictions on those who are part of the "diagnosed" generation.
Pre-diagnosed? Better take that drivers license away now. We only test for these things once every few years. The lawyers say we can't risk the liabilit
Re: (Score:3)
I would rather know than not know.
If nothing else I could get some of my affairs in order before I was too addled to to so.
As in wills, trusts, long-term planning, checking off a few bucket list items, etc.
And, possibly planning my own passing- a more graceful demise on my own terms, rather than wasting away in a nursing home for years and years in a dementia-addled fog.
I watched my grandfather go that way and it wasn't pretty. It wasn't "living" by any stretch of the imagination, it was more like watching
Re: (Score:2)
This. If I were diagnosed with dementia, there are some places I want to see. Beyond that, I'll end things on my own terms, thank you. I do NOT want my children watching me become mentally absent.
course, they may already think that - most young people tend to so think of their elders...
Re: (Score:2)
Not everyone with MCI progresses to Alzheimer's etc, but they need to manage their life carefully.
You want to know so that you can take measures to compensate while you can still make informed decisions. That is making sure you have someone you trust with power of attorney to manage your finances.
Not only to prevent theft, but just to manage the day-to-day things making sure the bills and rent are paid so you'll have a place to live as long as is possible. Also, your heirs/trustee need to know what bank and
Re: Detection is good, but... (Score:2)
8 hours of sleep per night.
Re: (Score:2)
As getuid() alludes to, there are things that can be done to delay or prevent the onset. There are drugs people can take that will delay the onset of dementia, and increasing the amount of sleep you get can actually prevent dementia, as lack of sleep has been linked to increased risk of dementia.
This requires machine learning? (Score:1)
Re: (Score:3)
Anyone else a little suspicious about this?
Yes, I am extremely suspicious of anyone claiming to have accurately trained an ML model using 31 data points.
Almost by the definition of machine learning (Score:2)
The output of machine learning is pretty much "pretty basic heuristics like just some plots". So yes, any decision made by machine learning can, in production use, be done with some pretty simple algorithms.
Machine learning with a decision tree algorithm outputs a bunch of nested if-then statements. Those if-then are then used in production. Can't get much simpler than that.
Machine learning with linear regression outputs a bunch of coefficients, the "some plots" you were talking about.
The value of machine
Data Science or Statistics != Machine Learning (Score:2)
Re: (Score:2)
I forgot to mention one other important aspect of machine learning. Adjusting the parameters over time. With ML, the machine not only finds the initial parameters based on the training data, it adjusts those parameters over time based on new data.
> Regression existed long before machine learning.
Regression by computer didn't exist long before machine learning. There was a machine learning checkers game in the early 1950s. It adjusted it's own parameters over time based on "experience". Pretty much as soo
Re: (Score:2)
What makes the study interesting is that they measured the volunteers amyloid plaque load before and at the end of the study, and they also chose people whose Clinical Dementia Rating scores of zero.
"It is worth noting that amyloid biomarkers appear to be associated with driving performance but not with global cognitive test scores, implying that the assessment of driving may be a useful strategy for the early detection of cognitive declines."
This is good news because many people don't want to have a spinal
Full self driving (Score:5, Interesting)
Yet many people are viciously against autonomous driving and active accident avoidance technology development. They demand a zero crash guarantee when humans are homiciding close to 1 million people a year. Nearly 40,000 per year just in the US. Yet fools do not want driverless cars even if they are 10 times safer. They do not care if their attitude murders 30,000 people many of them not even driving a car.
Re: (Score:1)
Yet many people are viciously against autonomous driving and active accident avoidance technology development.
Who exactly? Or this one of those 'they said' anecdotes?
They demand a zero crash guarantee when humans are homiciding close to 1 million people a year. Nearly 40,000 per year just in the US. Yet fools do not want driverless cars even if they are 10 times safer. They do not care if their attitude murders 30,000 people many of them not even driving a car.
Well I'd consider myself someone who is cautious towards automated driving, but not for any of the reasons you claimed. Automated driving removes some risk, but introduces others and these need to be well thought-out and thoroughly tested before people are killed. Is not a rational approach?
There is no doubt that the technology will eventually get there, but that doesn't mean you throw your brains out the window on product release 1.0.
Re: (Score:1)
Can't speak for anyone else, but I'm not against autonomous driving. When it actually works. From my informal YouTube watchings, it's a long way from general availability. A long way.
And I'm a computer guy. A lover of tech. Perhaps even a nerd.
Re: (Score:2)
The problem is that a zero deaths stance actually costs lives because many existing human drivers suck.
66 percent accuracy (Score:2)
Its going to have to get a lot better to be useful. I mean we get 50 percent with a coin flip, no data needed.
It's even easier... (Score:5, Funny)
Do you own a German car out of warranty?
Do you buy a new car every two years?
Do you own a Chrysler or Chevy?
All of these are signs of deteriorating metal capacity.
Re: (Score:3)
Not sure if deteriorating mental capacity or just trolling...
Re: (Score:2)
Insert Letterkenny "Fucking Embarrassing" gif.
Re: (Score:2)
Re: (Score:2)
he's been drpping lettrs from wrds ever snce he bougt that dodg :)
hwk
Re: (Score:2)
If you can afford it, a new car every two years isn't a sign of dementia anymore than going out for steak once in a while isn't irresponsible when you're middle class.
Re: (Score:2)
Do you own a German car out of warranty?
Just wondering, where does this one come from? BMW? Mercedes-Benz?
My wife's car is a VW Passat station wagon straight from Germany, getting 12 years old in a couple of months. It's quite literally the most care-free car I've seen so far, so much so that I've very recently decided to buy another VW. I guess having a gas engine rather than diesel helped a bit.
Note: I had a BMW 3 series coupé back before the kids, wonderful car to drive, but maintenance wasn't cheap.
Re: (Score:2)
My eight year old German car is doing fine too.
Although I'd suggest getting a Skoda rather than a Passat. Same car, same build quality, lower price.
Re: (Score:2)
lol - yeah, I could comment; having owned several VW's and BMW's and an Mercedes.
The Mercedes was a diesel, and that generation were pretty much bulletproof-reliable; I bought it with 300k miles, and did most of the work myself for the 3 years I owned it, and recouped my $5k investment when I sold it.
The classic VWs I've owned - is pretty much an apples-to-oranges comparison, and the modern one I owned, a diesel jetta, honestly was a complete pile of crap. The transmission and motor (ALH) was great, and I
Re: (Score:2)
I'd suggest getting a Skoda rather than a Passat. Same car, same build quality, lower price.
As a general rule, I'd sign this. Octavia rather than Golf, Superb rather than Passat, etc.
There's a couple reasons I didn't go that route: a) the Superb wasn't available as station wagon back when we were buying the Passat. b) Skoda doesn't have the Sharan (Seat would have the Alhambra). c) I now know the VW garage for its good service, so I'll rather stay with them than switch to a yet unknown Seat dealer.
I only wish there were more choice left in the "family van" range, pretty much everybody's moved over
Re: (Score:2)
Do you use your phone while driving? You're already too far gone. Off to the looney bin for you where you can't be a danger to yourself or others.
Re: (Score:2)
Do you own a German car out of warranty?
Do you buy a new car every two years?
[...]
All of these are signs of deteriorating metal capacity.
So IOW you have to replace your German car every three years to not be a worry? (That's how long their powertrain warranties last...)
Re: (Score:2)
Hmm, mine only had 3, not that it mattered by the time I got it
Guess they've expanded since I last looked. Go Korea! Maybe someday the Germans will catch up to your ten year warranties!
Not.
Oblig (Score:1)
There's an app for that. (Score:2)
Yes, it's probably possible to predict this stuff, but somehow I predict an app in the near future with a high margin of error and false positives.
Unrelated note. I wonder what disease people who turn their engines off at stoplights and fast food drive-thrus suffer from... I no longer have a car, so I've been riding a bicycle for a few years, and I see this a lot. There will be a red light, and as soon as it turns green, I hear their starter motor start their engine. Literally 5 seconds.. This was common
Re: (Score:2)
Nice presumption (Score:2)
I wonder what disease people who turn their engines off at stoplights and fast food drive-thrus suffer from...
That's a fairly obnoxious way to ask about the built-in automatic start-stop system [wikipedia.org] available from about two dozen auto manufacturers.
Re: (Score:3)
Also, the auto start-stop features are less about economy and more about passing tightening government emissions standards.
Re: (Score:2)
You were saying something about low UIDs?
Re: (Score:2)
A script alerts me whenever someone puts "low UID" in a post. Whut we talking about?
zzz snork zzz zzz Huh? Whut? zzz
Perhaps I missed something (Score:3)
Nice work (Score:2)
Re: (Score:2)
Very nice analysis - there was obviously no AI being used here. They may have had some machine learning in use, but no AI. Not that anyone has one, but a TRUE AI would analyze this sample, take the numbers, and declare "I can't predict shit with this crap"
It's not dementia (Score:2)
My car's front end alignment is shot.
Keep honking (Score:1)
Keep honking, I'm retired. Oh BTW, I drive a stick, by choice.
Re: (Score:3)
Keep honking, I'm retired.
Keep honking, I'm reloading. (just kidding, I'm kidding, really!)
Anyway, as I approach Certified Old Fogey status I rarely drive fast any more. Car time (what little of it I get these days) is "me time", and if I'm going too slow, use the steering wheel thingy and go around me.
I mean, I won't go 55 in the fast lane but I'm also not going to go 75 in the other lanes just because some jackass behind me is in a hurry. (Unless Matlock is on and I'm really late.)
The main reason I don't let anyone goad me into
Re: (Score:1)
Thank goodness I don't have to commute any more. What really made my blood boil is: the ones who *cannot* shut the damn phone off and forget that it exists. And also, the ones who never had Drivers Ed: making a turn right in front of you from 3 lanes over. You know they drilled it into us, to stay in your lane when you turn. *then* change lanes after the turn.
OK, I'm not *completely* retired yet but close enough to slow life down a bit. Getting spammed by AARP etc.
Re: (Score:2)
Thank goodness I don't have to commute any more.
Same here. I won't even consider a job that requires me to get up, shave, shower, get dressed, get in a car, spend time driving X number of miles to an office, drive back home, and then basically repeat that same mindless, time-wasting sequence day after day after day after day.
Not gonna do it, period.
short drives, local roads (Score:2)
Anybody notice how interesting it is to consider short local drives and indication if mental issues to come, given we (in California) have all been told for the last year to stay home and not take trips?
Oblig. (Score:2)
Re: (Score:2)
Senior life is more like this one.
https://www.youtube.com/watch?... [youtube.com]
A more accurate measurement (Score:1)
Are the turn signals always on?
Who needs AI? (Score:2)
Houston, TX (Score:1)
Stupid studies are a sign of dementia? (Score:2)
Honestly, it's strange that the researchers see these signs as predictive. The most correlated variables, according to the study, were age, race and the nature of trips the person takes (distance and length). Age and race are pure statistics that just predict what statistics normally can. Trip distance is probably a result of already existing conditions or diagnosis, such as trips to local doctors or when family recognises that the old person has problems and for example gets them to move closer and better
Nobrainer (Score:2)
heyy (Score:1)
Do you know this feeling when Slashdot headlines (Score:2)
sound like real life stealth insults?
In the past year... (Score:2)
Re: (Score:2)
the raw number of trips, lengths, or even ratios, probably doesn't mean much by itself.
A *change* in these, though, for the same person, could be indicative of something.
We've been able to do this for years (Score:2)
Look at driver:
Is driver on phone while driving -> Mentally handicapped idiot, likely suffering from severe late stage cognitive decline.
Is driver not on phone while driving -> probably okay.
Definitely dementiafiable (Score:2)
In-car computing is not only capable of identifying distracted drivers but emerging drivers who will go on to test positive for real diseases. Dementia but also emergent FSD diseases like backseat driving, drive apnea and DWI class impairment. Preventable fatalities!
SO its not just FSD is the answer. Its behaviors that point to pathologies that can be categorically associated to statistical fatality numbers. That will emerge as FSD begins to show the multitude of ways human drivers operate that contribute t
Demented researchers (Score:2)
So, let's see, based on training on 64 subjects, you are now going to lose your drivers license because the infallible AI, believed by idiots like the people in charge of drivers licenses, has said that you are demented.
Bogus statistics (Score:2)
They convert people to person-years as if multiple measurements on the same person are independent in order to get the statistically strong results they obtained. The reality is they are off by an order of magnitude on measuring confidence. Staplin et al which they quote established that you will need to study 4000 elderly drivers in a year in order to see 1 crash involving persons who are also suffering mild cognitive impairment. So in some years in 4000 people a result of 0 would be perfectly reasonab
Missing (Score:2)
Missing words in your can also be a sign.