[HN Gopher] Autism, through the eyes of a computer (2019)
___________________________________________________________________
Autism, through the eyes of a computer (2019)
Author : wjb3
Score : 31 points
Date : 2024-02-04 20:49 UTC (2 hours ago)
(HTM) web link (www.spectrumnews.org)
(TXT) w3m dump (www.spectrumnews.org)
| wjb3 wrote:
| By using machine learning on data from wearable biosensors,
| researchers can predict imminent aggressive behavior in autistic
| children and teenagers in psychiatric inpatient settings.
| temp0826 wrote:
| This is the most boring version of a Minority Report-esque
| future I could have imagined.
| alvarezbjm-hn wrote:
| Rather cathastrophic.
|
| "Some experts caution that these tools can never substitute
| for the judgment clinicians hone through years of experience
| -- but should instead augment it"...
|
| Good luck with that. Tools can't have responsibility. (Bad
| for customers, good for organizations)
|
| Elaborating: What % of autist people kills someone out of
| aggressive behavior? I am quite sure we don't know, because,
| I suspect, it is low. Compare that to the general population.
|
| The only reason they want to put tracking devices on autists
| is because. unlike healthy people, autists can't fight back.
|
| Once the technology is mature, you try to enforce it on (sell
| it to) exconvicts, immigrants and jaywalkers.
| mistermann wrote:
| [delayed]
| bitwize wrote:
| More like Blade Runner, with the Voight-Kampf test and that.
| WWLink wrote:
| Ehh... I can tell you reasons why that won't always work.....
|
| First off, sensitivity to stimuli. I see someone wearing a
| wristband/watch in one of the pictures in the article. Some
| autistic people are super sensitive to wearing things like
| watches, necklaces, bracelets, anklets, headbands, etc. Oh
| wait, sensor gear is almost always one of those things...
|
| The other is that as much as people talk about repetitiveness
| and rituals, I can tell you that sometimes there is no pattern!
| My brother who is most strongly on the spectrum, is older than
| I am... so I've known him my entire life. I have a pretty good
| idea what he likes, doesn't like, and what stresses him. There
| are still occasional days where he has a meltdown and I sit
| there and go through my 'notes' and can't figure it out.
|
| One time I was trying to recount this experience to his doctor
| and she said it sounded like I was looking for signals in a
| field of noise. She was right. Sometimes it's just too random
| to really know.
|
| I imagine it's easier if you have a willing subject who can
| communicate with you why they got upset after the fact.
| Unfortunately for my brother, "why" is a question he doesn't
| like answering - and when he (rarely) does, the answers are
| usually not very credible like "Why were you so angry?"
| "because you haven't taken me out shopping on Saturday yet!"
| (and it's Tuesday and we just went Saturday)
|
| (The funnier part is he'll often ask me "why" I do things lol
| :D)
|
| Anyway, back to the article. I don't like it. Inpatient
| psychiatric settings are already a scary sorta place. You get
| locked in a building you might not be familiar with, might have
| to wear different clothing (not usually but sometimes), be
| exposed to different noises and scents and temperature... and
| be around other people, also with other issues.. Being forced
| to wear sensors sounds like it'd add to the awfulness. It's
| also not a great environment to observe someone because they
| might just "shut down" from their normal behavioral patterns.
| AND IF THEY DON'T "SHUT DOWN" they might be encouraged to do so
| by getting a strong dosage of medicine.
|
| That's not fun. I get where they're coming from - psychiatrists
| (that take insurance, especially state medicaid plans which
| most patients would probably have) are usually overworked,
| underpaid, and it's a stressful job where everyone's hoping for
| an easy fix but it's not easy.
|
| Still, I appreciate the sacrifice the patients in those
| research projects are making, and the work the researchers are
| putting in. Especially for adults on the spectrum. There are
| vast vast amounts of resources expended on helping kids with
| autism but the moment you turn 18 it's basically "good luck
| lol"
| o11c wrote:
| > I can tell you that sometimes there is no pattern!
|
| As an autist: you don't see the patterns that you've
| disrupted. We don't talk about most of our problems.
|
| Often it can be something like: "3 days ago, you forgot to
| mention ahead of time that you were going to the store. As a
| result, everything has been stressful and I responded badly
| to things that normally would've been below the alarm
| threshold."
|
| Generally I find that 3 days is about how long oneshot
| disruptions cause problems, though this is tapering. 1 day is
| usually enough for my coping strategies to kick in and
| mitigate the _worst_ at least. Of course, ongoing problems or
| unrelated but coincident problems can stretch it longer.
| WWLink wrote:
| I think 3 days is a reasonable look-back period for
| analyzing. I want my brothers (both are on the spectrum,
| though one is affected more than the other) to be happy and
| less stressed. Speaking selfishly, it... makes my life a
| lot easier. Honestly when I have those "what happened?"
| moments I can go back as far as a week or two sometimes
| looking for small shifts.
|
| One that really sucks for all parties is the unpredictable.
| Something like, the doctor being late to their appointment
| or our assistant we hire to help not being able to come
| that day (or coming late)... Or THE WEATHER MESSING PLANS
| UP. Especially the anticipation that weather might screw
| the plans up. Plans that are "we've done this for the past
| 26 saturdays so why can't we do it this saturday? Can't you
| just make it not rain or something?" :D (I am exaggerating
| for extra silliness)
|
| Those are the worst x_X
| crtified wrote:
| Presumably, some combination of measurable signals (e.g. heart
| rate, blood pressure, perspiration, certain movements, muscle
| tenseness, ....) could do the same thing, regardless of whether
| autism were present or not.
|
| Patient cooperation may be a factor - or, " _those durned
| prisoners keep taking their sensors off_ ". And then when we
| take the next logical step - surgical implants - it does indeed
| start to feel a tad Minority Reporty.
| gamepsys wrote:
| Just wait until optical sensors can collect enough data to
| make the same predictions.
| mistermann wrote:
| [delayed]
| toddmorey wrote:
| There is related autism & self-regulation research out of Harvard
| & Boston Children's[1] that uses video games & biofeedback
| sensors in a very ingenious way:
|
| The children participate in common games on a tablet where the
| difficultly (and the stress of playing) slowly increases. Signs
| of frustration / stress from the bio monitors like a rapid
| heartbeat make clouds slowly start to cover the screen a bit like
| a screensaver. That provides the children a visual clue to do
| some self-soothing. Take deep breaths, your heart rate lowers,
| the screen becomes clear again so you can keep playing.
|
| The idea is to practice emotional regulation and emotional
| awareness a few hours a day to build competency.
|
| [1] https://www.mightier.com
| 54543vjjj wrote:
| Autism will end up being torn into independent diagnosis of
| correlated traits. There's no underlying nexus between the list
| of symptoms beyond statistical correlation.
|
| High IQ? ADHD? Bipolar? Eye sensitivity? Loneliness? It seems
| little is in common beyond correlations.
| StopTheTechies wrote:
| > There's no underlying nexus between the list of symptoms
| beyond statistical correlation
|
| Correlation is correlated with meaning, so what are you saying
| coldtea wrote:
| This description is not even wrong. It's random
| misunderstandings combined.
|
| There's a list of autism symptoms used diagnostically. The list
| above is nowhere even close. Neither High IQ is used as a
| diagnostic of autism (and you can have autism and low IQ), nor
| ADHD (hell, until recently it was the opposite: doctors used to
| believe you can't have both ADHD and autism), nor bipolar
| disorder (at best a comorbidity).
|
| I'll give you "eye sensitivity" (sensory sensitivity), which is
| one of the diagnostic symptoms.
|
| As for loneliness, that is a second order effect under certain
| social circumstances, not a psysiological result of autism.
| Might as well consider bullying a symptom.
| sureglymop wrote:
| Often there is a diagnosis of autism _and_ of the comorbidities
| you mentioned.
|
| And there is not always only a correlation and not a causation.
| People with autism, adhd etc. have a higher risk for depression
| and social anxiety, often induced by the experience of living
| with these conditions.
| dang wrote:
| Discussed (a bit) at the time:
|
| _Autism Through the Eyes of a Computer_ -
| https://news.ycombinator.com/item?id=20876988 - Sept 2019 (6
| comments)
| gamepsys wrote:
| I do like the attempt at aiding diagnosis with 'digital
| phenotypes.' Currently, the number one input into getting a
| mental health diagnosis is self reported (or
| guardian/caretaker/teacher reported for minors) subjective
| experiences. This generates a huge problem with screening or
| diagnosing. The current state of the art for autism diagnosis
| involves a panel of experts and multiple days of
| testing/observation/interviews. We currently know there are
| physical differences in the body and statistical differences in
| behavior for autism. I don't think this watch is the answer --
| but I do appreciate the recognition this article makes that
| diagnosis is currently inefficient and inaccurate.
|
| The diagnosis problem exists for other mental health conditions
| too, but Autism is perhaps the worst example in the industry.
| shutupnerd0000 wrote:
| This is the exact problem with most mental health diagnoses -
| patient answers a set of 10-30 questions, nurse practitioner
| nods solemnly and writes prescription for controlled substance.
| We are in the Stone Age with mental health.
___________________________________________________________________
(page generated 2024-02-04 23:00 UTC)