Schizophrenic Computer Points to New Theory of Disease | Hacker News Reader

Schizophrenic Computer Points to New Theory of Disease

spectrum.ieee.org

69 pointscaustic18 comments

Back in the 90s, Neal Stephenson wrote a great short story about the implications of such technology: http://www.vanemden.com/books/neals/jipi.html

I couldn't help but think of a drug user's eagerness to attach deeper meaning to seemingly unimportant events while under the effects of heightened dopamine levels.

And drug use is thought to trigger schizophrenia in certain people: http://en.wikipedia.org/wiki/Schizophrenia#Substance_abuse

Only by a small portion of the psychiatric community. The vast majority believe that the causation runs the opposite direction, that it's self-medication. This has been discussed for years for just about every serious mental illness, in fact. There's no real consensus either way, but most mental health professionals I know agree with this.

(I grew up in a house of mental health professionals, so I've had no dearth of opinions on the matter, to say the least.)

Yet another problem that can be reduced to the chicken V egg dilemma.

I think this ties in closely with the dopamine hypothesis of Schizophrenia and Amphetamine being the chemical model for said disease.

This is truly fascinating, both from a computer perspective as well as a psychological perspective. I have a person very close to me who may be schizo-affective (diagnosing these things can be a bitch!), so I'm all for anything that helps us understand the disorder.

You're bang on when you say diagnosing these things can be a bitch. Especially so with schizo-affective disorder where symptoms that are akin to schizophrenia and symptoms that are akin to bipolar are both present. I misdiagnosing twice has happened before landing on schizo-affective.

I had a relative who had this happen. Their symptoms seemed so much like schizophrenia, with bizarre delusions and voices. On the other hand, bouts of depression and bouts of hypomania were common. Before the mood component was fully realised, he was slapped with a Paranoid Schizophrenia diagnosis - which can particularly life shattering. A year or so later, a diagnosis of Bipolar was slapped on him, which is a bit of improvement because I believe the prognosis for Bipolar is markedly better than Schizophrenia.

Eventually, it was realised Schizo-affective Disorder fit the bill. After a bit of research, it turns out that such a diagnosis fit particularly well. The particular type of delusions commonly experienced by those with Schizo-affective disorder were present in him, amongst other things.

You can probably tell from what I've said above that it is indeed a bitch. It's almost life destroying and he was reduced to a lab rat in the process (finding the right medication is really a case of an educated guess and experimentation). I really am all for this sort of research because the more we advance with it, the less people will have to go through what my relative went through (and relatively speaking, he got off lightly).

I can't help but think: no wonder the computer is experiencing schizophrenic symptoms; they never let it sleep!

How does assigning greater importance to learning lead to mixing up identities?

There doesn't seem to be a link to the actual paper (if there is one yet) - the DISCERN link is to a 1998 paper. The idea of a "story parser" is interesting...

http://nn.cs.utexas.edu/?grasemann:cogsci11

I believe this is the paper, from the abstract it seems like hyperlearning is related to greater consolidation of memory, which might imply creation of connections between things that are not really connected

The "connection between things that are not really connected" part match my own experience of having schizophrenia very well.

It's not uncommon for me to have thoughts that very much makes sense and is meaningful to me while others simply can't make any sense of it what-so-ever.

This means I'll have to be cautious of my own thoughts and I often doubt them.

Thoughts as well as meanings change over time too. Sometimes I write down ideas I think is brilliant to save for later. Revisiting those writings can be troubling because I might not be able to make a similar sense of what I wrote earlier.

Thanks. I can see how this would lead to "derailments", particularly if the story line was modeled as arcs (given it's a "connectionist model"), but the "delusions" seem less obvious... unless things, including identities, were also modeled as arcs between different ways of referencing the same thing (e.g. my address, where I live, my home). Sort of using the same model for instances (events in a story) and schema (actors in a story) - which makes sense, and may well be how it works in the brain. Reuse, don't reinvent, is what nature does when it can.

I have doubts about Science based on making up computer models and seeing what happens (although it is a lot of fun to do). OTOH, they clearly casting this as merely a hypotheses - which is pretty reasonable.

DISCERN is old - this researcher has apparently been working on it since 1990.

According to popular mechanics, the article was published in the online version of Biological Psychiatry. This seems to be the issue in question: http://bit.ly/n66BeO

Not sure I got the "mixing up identities" comment part right but just wanted to point out schizophrenia has nothing to do with multiple personality disorders.

9th para:

  DISCERN was just as eager to adopt the life of a crime boss named Vito
  or to thrust its assigned identity onto another character.

Oh, right.

Reading that at first I thought the writer had completely mixed up schizophrenia and multiple personality disorders. Something that is not uncommon at all in the media and pop culture.

I also have a schizophrenic computer, but that points to the fact that I shouldn't have upgraded to OS X Lion.

Reply on news.ycombinator.com