SoylentNews is people
SoylentNews
Most of the time, people don't actively track the way one thought flows into the next. But in psychiatry, much attention is paid to such intricacies of thinking. For instance, disorganized thought, evidenced by disjointed patterns in speech, is considered a hallmark characteristic of schizophrenia. Several studies of at-risk youths have found that doctors are able to guess with impressive accuracy—the best predictive models hover around 79 percent—whether a person will develop psychosis based on tracking that person's speech patterns in interviews.
A computer, it seems, can do better.
That's according to a study published Wednesday by researchers at Columbia University, the New York State Psychiatric Institute, and the IBM T. J. Watson Research Center in the Nature Publishing Group journal Schizophrenia. They used an automated speech-analysis program to correctly differentiate—with 100-percent accuracy—between at-risk young people who developed psychosis over a two-and-a-half year period and those who did not. The computer model also outperformed other advanced screening technologies, like biomarkers from neuroimaging and EEG recordings of brain activity.
The article does not elaborate on how the transcripts are produced.
Automated analysis of free speech predicts psychosis onset in high-risk youths
Original Submission
(Score: 4, Informative) by Aichon on Friday August 28 2015, @03:07PM
Given that psychosis is not the norm, couldn't you just say, "Nope, not psychotic", 100% of the time and come out ahead of a model that is only accurate 79% of the time? I'd love to hear what their false positive and false negative rates were.
And then 100% accuracy for the computer? Was their sample size ridiculously small or their sample set biased? The article has this buried in the body..
[...] their work poses several outstanding questions. For one thing, their sample size of 34 patients was tiny.
And the original paper [nature.com] had this:
Thirty-four CHR youths (11 females) had baseline interviews and were assessed quarterly for up to 2.5 years; five transitioned to psychosis.
CHR stands for "clinical high-risk".
So, their sample size was ridiculously small, but they also picked a sampling from among some of the most difficult patients to correctly categorize. It'll be interesting to see if they can apply it more broadly or without the various exclusions that they applied when selecting their sample set (e.g. no dangerous patients, English speaking for basically entire life, IQ >= 70, etc.).
(Score: 3, Interesting) by Runaway1956 on Friday August 28 2015, @03:20PM
Maybe I should continue to trust the old women in my life. When two or more old women declare someone to be "not right in the head", it seems they are always right. Both keywords are important here: "old" and "women". Young women aren't any brighter than us guys are. Hell, young women MARRY the psychos of the world! (Or, at least date them.)
(Score: 2) by Phoenix666 on Friday August 28 2015, @03:24PM
Hmm, there may be something in what you say. Older women also seem to have an uncanny knack for detecting fevers by kissing foreheads--no thermometers required.
Washington DC delenda est.
(Score: 2) by c0lo on Friday August 28 2015, @09:32PM
Of course, this is how they build their expertise.
Besides, no other fertile male (but psychos) is available when they want to marry.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 1) by Francis on Friday August 28 2015, @04:33PM
It's not a matter of psychotic versus not psychotic, it's a matter of Schizophrenia versus Bipolar versus other possible diagnoses. Somebody that's psychotic is usually pretty easy to identify. It's putting them into the correct diagnostic category and providing appropriate treatment that can be an issue. And if you don't get it right, it means that you're trying to treat the wrong part of the brain with any medications you want to use.
(Score: 2) by Phoenix666 on Friday August 28 2015, @07:29PM
The distinctions between schizophrenia, bipolar, and psychotic seem a bit fuzzy to me. When I read the article, the verbal symptoms the system was filtering for were quite present with a person I know who has the most severe form of bipolar disorder it's possible to have. He also did hallucinate.
Washington DC delenda est.
(Score: 1) by Francis on Friday August 28 2015, @09:19PM
That's largely because they are rather fuzzy. Mental illnesses occur along a continuum and there's no particularly good way of drawing the boundary lines. Additionally, the behaviors are driving by the brain and the availability of tools to look at that is still quite limited. Not to mention that there isn't much understanding of how to apply what's seen in the brain to a treatment program.
(Score: 2) by Phoenix666 on Friday August 28 2015, @09:34PM
I can say that dealing with that bipolar person in my life was incredibly stressful. It freaked me out.
But I still don't want such fuzzy definitions and slapdash science to become the basis for public policy that government and corporations (because, how can you separate the two anymore?) use to deny people their basic rights.
Washington DC delenda est.
(Score: 1) by Francis on Friday August 28 2015, @11:55PM
In the real world things are fuzzy. You can't create clear categories or distinctions for things that don't have them. You're left with things like blood quantum, and they bring their own set of problems. The human brain just isn't understood adequately to make for clear distinctions that are reliably diagnosable. Most of the time the goal is to get close enough that treatment is possible. Asking more than that is probably not realistic at this time.
(Score: 2) by frojack on Friday August 28 2015, @07:06PM
but they also picked a sampling from among some of the most difficult patients to correctly categorize.
It looks to me that they picked sampling from "clinical high-risk" patients most likely to develop psychosis (as predicted by some unspecified means).
They then dogged these for 2.5 years of longitudinal study. (Shrinks stalking you for 2.5 years might have some effect, No?).
How would it work on the general population of some random Junior High school?
No, you are mistaken. I've always had this sig.
(Score: 0) by Anonymous Coward on Friday August 28 2015, @07:36PM
> They then dogged these for 2.5 years of longitudinal study. (Shrinks stalking you for 2.5 years might have some effect, No?).
What does it say about you that patients getting regular treatment is something you consider 'stalking?'