Algorithm Can Identify Suicidal People Using Brain Scans (wired.com) 87
An anonymous reader quotes a report from WIRED: In a study published today in Nature Human Behavior, researchers at Carnegie Mellon and the University of Pittsburgh analyzed how suicidal individuals think and feel differently about life and death, by looking at patterns of how their brains light up in an fMRI machine. Then they trained a machine learning algorithm to isolate those signals -- a frontal lobe flare at the mention of the word "death," for example. The computational classifier was able to pick out the suicidal ideators with more than 90 percent accuracy (Warning: source may be paywalled; alternative source). Furthermore, it was able to distinguish people who had actually attempted self-harm from those who had only thought about it. In today's study, the researchers started with 17 young adults between the ages of 18 and 30 who had recently reported suicidal ideation to their therapists. Then they recruited 17 neurotypical control participants and put them each inside an fMRI scanner. While inside the tube, subjects saw a random series of 30 words. Ten were generally positive, 10 were generally negative, and 10 were specifically associated with death and suicide. Then researchers asked the subjects to think about each word for three seconds as it showed up on a screen in front of them. "What does 'trouble' mean for you?" "What about 'carefree,' what's the key concept there?" For each word, the researchers recorded the subjects' cerebral blood flow to find out which parts of their brains seemed to be at work.
Re: 90% = shit (Score:2, Insightful)
Yes. There are no distinguishable brain characteristics for this kind of thing, every person is different, there are no physiological comonalities that denote a decision people will each feel very differently about. More sci-fi bullshit from futurist nincompoops tested on a small, hardly representative group of people one time, if not an outright simulation. Spare us. Very little in real life is that linear.
Re: (Score:3)
Re: (Score:1)
Re: (Score:1)
As someone who almost never talks to anyone but has attempted suicide, such asking is very effective. Best if the person asking is a stranger because then guilt from making the person feel bad will be far less than if a relative was asking.
Re: (Score:2)
With 17 samples? That is on a level with saying the kids next door quite often play football, so they might win the Premier league one day. It is true: they might - but they might all commit suicide before they win the Premier League.
Its not training, its bullshitting!
Re: (Score:2)
They could try simply asking if the person is suicidal.
They did, evidently that's how they got 17 volunteers who reported suicidal ideation. Then they compared against people who actually tried to commit suicide, and got a pretty damned good match with admittedly a small dataset.
It's not clear to me why we want to do any of this, but it maybe it has application in more useful fields. I'm not sure why we want to stop people from committing suicide, I think we'd want to help them do it quickly and painlessl
Re: (Score:2)
Suicide tends to not be that 'rational choice for rational reasons' that you seem to believe it is. It tends to be an impulsive act, involving a significant amount of high-grade magical thinking--and if you object to religion, you certainly should be objecting to that!
Of the remaining group--well, it might be better to aim more for offering them the ability to do something of value with their death, a satisfying death instead of merely a quick and painless one. There are things out there which are most li
eSnoop (Score:1)
They've invented political affiliation detection, education level detection, eGaydar, and now suicide detection. What's next, fingerprints that indicate one jacks off twice a day...I mean once a day?
Re: eSnoop (Score:1)
Re: (Score:2)
How about just asking me?
Getting scary (Score:5, Insightful)
An abstract of the nature.com article is here [nature.com].
I hope this algorithm will help prevent suicides.
However, the increasing ability of machines to read minds is getting a little scary. Some day we'll be able to read someone's emotions without hooking the person up to a machine - just point a reader at their head.
Government employee: What do you think of our dear leader?
The person's brain shows the emotion of revulsion.
Government employee: Off to a re-education camp, for you and your family!
Re: (Score:2)
George: Jerry, just remember. It's not a lie... if you believe it.
Re: (Score:2)
Once this is easy and cheap, governments around the world will be using this en masse.
Full paper (Score:3, Informative)
The full paper (or at least some version of it) is available online here:
https://nocklab.fas.harvard.edu/files/nocklab/files/just_2017_machlearn_suicide_emotion_youth.pdf [harvard.edu]
It's probably a preprint without the last-minute changes, but it should be good enough to understand the research.
Re: (Score:2)
I hope this algorithm will help prevent suicides.
Why?
Re: (Score:2)
I hope this algorithm will help prevent suicides.
Why?
I know you're just trying to be provocative here, but a majority of people that survive a serious suicide attempt will recover and go on to live better lives. Maybe you're advocating for personal freedom or something - but in most cases, a suicide attempt is best understood as a symptom of profound mental illness, so it is a good thing to treat that illness and save a life to the same extent that it is a good thing to cure a person's cancer.
Re: (Score:2)
No, I am not. I am chronically depressed, have attempted suicide in the past and genuinely think that preventing somebody from committing suicide is interfering with their personal rights. Helping them climb out of the depression pit is a different thing.
Re: (Score:2)
Oh... I'm sorry you have to deal with that kind of struggle. I'm sure it isn't easy coping with that kind of thing and I appreciate you being candid about it, obviously that changes the discussion.
I do have to ask - if you had a suicidal friend, wouldn't you still want to help them reconsider, or at least wait it out a bit? I do agree that we should generally give people the ability to govern their lives as they see fit, and I think in cases like that of Terry Pratchett, he should have had the right to end
Re: (Score:2)
Oh... I'm sorry you have to deal with that kind of struggle.
Don't be. I don't see it as a struggle, rather "life is shit and I live in it" - kind of like taking each moment at a time and wait for the big finale of a rather poorly played show.
I'm sure it isn't easy coping with that kind of thing and I appreciate you being candid about it, obviously that changes the discussion.
I don't mind talking about it, over the years I've become detached, I look at myself like watching TV: there's some dude there with a miserable life and I am watching the show. The only difference being there's only one channel and you can't turn the TV off, so you're stuck with the 24/7 show. It has become... not hard. Just...
Re: (Score:2)
And then there's the middle aged dude who lost his entire family in a plane crash, bodies maimed beyond recognition and all - that's beyond help, that man would never ever function normally again
This is simply untrue. He might never forget, any more than he would forget the death of anyone he loved, but he is not "beyond help". Many people have had to cope with awful tragedies and gone on to live good lives.
Re: (Score:2)
Good for them. Others haven't.
Re: (Score:2)
I hope it won't be (ab)used till it's a bit more accurate.
Assume one person in 1000 is suicidal. 90% accuracy means that out of 1 million people (1000 suicidal), it'll correctly identify 900 suicidal people as suicidal, and misidentify 99,900 non-suicidal people as suicidal.
This is not particularly useful at this time. Perhaps when the fraction of the population that is suicidal at any given time is up to 30%+....
Minority Reportish (Score:2)
Completely scary that mind reading is about to become mainstream. Lie detector tests are going to seem quaint very soon.
--
"A lie that is half-truth is the darkest of all lies.” -- Google, Aphroism's about lies
Re: (Score:1)
Meh, I wouldn't worry about antifa. They're all just a bunch of pimply pasty little pussies that act tough when they're in big numbers, but if they find themselves facing a real threat (the police aren't, btw, and antifa knows it,) they run and hide. Videos of this happening are all over youtube -- antifa keeps shouting about how cops are evil, they then block the wrong person from walking to work who beats the fuck out of one of them, and all of a sudden they start cowering and asking why the police aren't
Prevention (Score:1)
I discovered a way to detect someone's intent to scan my brain.
Unlikely it would work, in practice... (Score:3)
While the most common reason for suicide is linked to a depressive state, it is not always the case that depression is linked to any ongoing psychological abnormality that could be diagnosed clinically, or that anyone would have a reason to want to do a brain scan on you in the first place over.
In fact, there are some life experiences that, if you didn't experience any kind of depression, and could always completely detach yourself from any emotional investment, then *THAT* would be an indication of something being wrong with you. Losing a beloved family member, sudden unwanted changes in living circumstances, being wrongfully accused of a crime... all of these things and more can be legitimate reasons for a depressive state that can turn suicidal.
But such depression is entirely circumstantial, and not indicative of a larger scale psychological dysfunction... and because of its ephemeral nature, would not generally be caught on anything like a brain scan.
While the theory for this might seem wonderful, I have serious doubts it would actually ever save anyone's life.
Welcome Pre Crime (Score:2)
Re: (Score:2)
Mod parent down, spam links to some bullshit Indonesian betting games.
Seems (Score:2)
the one's trying to prevent someone to end his/her life are just afraid of their own final destiny in this plane of existence with their body. Won't help them though, it will happen to them. .... .
Who owns your life? You or well, once upon a time, it was the king, but who is it now - Mr Trump?
Seems a legal way is to get a doctor's opinion that it's OK to end it because of
There is so much messing around with one's basic rights - about your body, what you do with whom in your bedroom, what you can speak, what
Databrokers already know (Score:1)
You don't need brain scans for this. Databrokers already try to figure this by datamining your data.
Re: (Score:2)
Databrokers
Google, Facebook and Twitter all need to commit Sudoku.
Re: (Score:2)
They can already do it by face recognition - with at least 50% success rate!
Re: (Score:2, Insightful)
Re: (Score:2)
Honestly it sounds more like a pilot study. fMRIs are fantastic for trying to narrow down what parts of the brain seem to be involved in what, but they're seriously expensive things to run.
Since I expect most people here are more in the engineering side of things--pilot studies are a form of providing proof of concept. It's the cheap(er) version you slapped together in order to demonstrate that your idea actually has a snowball's chance in hell of working. They are not always going to be terribly open ab
Re: (Score:1)
False-positive for metal fans (Score:3)
Suicidal, or 'ideators'? (Score:3)
Especially given that depression tends to ruin your focus, motivation, and ability to execute a plan (even one you fully agree would be in your interests); people who would merely prefer to be dead are a larger group, likely by a fair margin, than ones who do something about it. If all your fancy brain scan can do is provide the same 'are you thinking about suicide?' data that a few minutes of sympathetic questioning by a vaguely competent counselor or psychologist can, it is a scientific curiosity. What would actually be interesting is telling us which of the suicidal ideators are justing fantasizing; and which ones are preparing.
(Now, if you really wanted to get futuristic, you could look into having a reasonably efficacious treatment option available for those you identify...)
Where do suicidal thoughts come from? (Score:2)
It is easy to explain the suicidal nature of bees and ants. They are genetically identical to their queen, who is their sister. So any child of the queen is the child of the worker, so she (all workers are females) is willing to die for her queen.
In some sense an insect colon
Re: (Score:2)
From the evolution point of view, suicidal thoughts would be selected against.
Not necessarily. A tribe member which is effectively a burden on the population can be purged to benefit the group as a whole. Abortion (both natural and intentional) as well as infanticide have been present across a wide swath of human history. Pro-social behaviors, including altruism, are present as well, even in lower species.
Many tribes have extended familial relationships, which could encourage selection for tribal altruism. Once selected, the trait would take quite some time to die out even if our com
Re: (Score:2)
fMRI studies = Kluger Hans (Score:2)
Might as well point a dowsing rod at people, or have a panel of self-described experts look for "tells" not unlike a gaggle of highschoolers.
Stay out of the 'control group' !!! (Score:2)
It would have to be a group of people who participated in a suicide study and died under mysterious conditions made to look like suicide.
the usual suspects (Score:2)
Artificial Intelligence Can Now Predict Suicide With Remarkable Accuracy [slashdot.org] — 16 June 2017
All we need now is an algorithm capable of picking out Winona Ryder (as goth-girl Lydia) from a Beetlejuice police line-up, and that about wraps it up for suicide prediction.
How pattern classification actually works (Score:2)
Pattern classification, as used in this study, is not neural mind-reading. Here's a good article by a neuroscientist that explains how it actually works -- and its limitations -- in plain English.
Pattern Classification Explained [lisafeldmanbarrett.com]
I invented a cheaper machine. (Score:2)
I invented a cheaper machine. It doesn't need to scan your brain, it just needs to know your age. If you're under thirty, it will guess yes and be right most of the time.
Cost versus reward (Score:1)
Look, MRI scans are expensive. Interpreting them is even more expensive.
Even the cheapest headset MRI scans run 3-4 figures.
So, in terms of detection, lack of availability, and even having such scans be useful, this is just not going to change anything.