This reminds me a lot of what I read about phantom limb pain in V.S. Ramachandran's book Phantoms in the Brain[0] and the kind of treatment he helped pioneer for that.
In the book he describes an innovative but quite lo-tech approach using a cardboard box and cleverly arranged mirrors that enabled patients to visualize the missing limb and effectively, as I understood it, rewire their brains to reduce phantom limb pain.
That example does make me wonder how critical the digital component of the treatment is. The article notes:
> Patients were then encouraged to engage in a dialogue with the avatar, who was controlled by a therapist.
I wonder if this would work with similar effectiveness using puppets or actors or perhaps even mirrors.
This is indeed exactly the same principle, where you fool the brain into connecting a buggy representation with a concrete object that can be gradually modified into something that better fits reality.
Re. digital vs. puppet, that's a good question. I suppose that the closer you can get to the hallucinations the better the treatment will work. So it may work with low tech gear, but it will probably be less effective (because it would be harder to reach the suspension of disbelief that's probably required for the treatment to work).
It is not really a buggy representation though. Autocorrecting your behaviour based on other opinions is a huge part of being human. Its a loss function that we seek to minimise. Whats buggy about it is that you start hearing the process that results in a loop. Listening to the outside world normally has a higher priority. Thats how you autocorrect you inner model. But now you are autocorrecting your inner model based on your inner model. As a result your theory of mind skills deteriorate.
tldr;
Was smoking weed for over 10 years, later on started clubbing and mixing weed with stimulants like Ecstasy and Amphetamines, got a drug induced psychosis within 5 months. Symptoms: Auditory Hallucinations, Paranoia, Persecutory delusions - was searching for hidden cameras around me every time I had an episode.
My two cents on this study - it works because the voices you hear are what you think others would think of you in this particular situation. Once you start having a conversation with your voice[s], your expectations (your inner modelling of others) change.
A close friend of the family had hallucinatory voices. He thought he could read minds. He decided to play Texas Holdem poker in the casino to take advantage of his mind reading inner voices. After a painful period of not having any money, he figured out that the voices weren't real nor could they read what the other players were holding in their hands. He is totally normal now.
Absolutely helps. Separation from a screen for 20~ hours can cause an episode for me. Left my PC at home to read a book and ended up having a panic attack and hallucinating a skull out of a blanket. Panicked and left my grandmothers house at the time. Only about a month and a half ago.
I want to add two things that aren't directly related to the article, but which might contribute to further discussion.
First, there's an anthropologist who's found that the stereotypical, combative voices are only universal (among a small sample) to the American/western participants in her study.[1] In cultures that offer alternate explanations for them ("spirits" or such), the voices are much more frequently friendly or even helpful.
Second, but less rigourous, that "new approach" mentioned in the Stanford article sounds remarkably like how someone one of my high school teachers knew handled his own voices: he did have trouble with them for a while and wound up needing to take a break from the handyman outfit he (if I recall correctly) ran with his brother. Eventually, though, his therapist recommended he stop trying to push the voices away completely, and the next time my teacher saw him, a few years later, he was back working again, and just telling people "if you hear me talking as I fix your porch, it's not to you."
It's not my own depiction, but I like the idea that part of why the voices here get so aggressive is that we're dead set on denying their reality -- if we try to ignore them, they just shout louder until we can't ignore them. And, as this article shows, the way we cast someone's voices as controling or, at the very least, setting the course of their life doesn't do anybody any favors when that leads so easily to feeling like they're subject to the voices' whims.
Another win for talk therapy and important implications for therapeutic applications of ai. A great business opportunity where a lot of good can be done. The pharmaceutical market is huge.
You probably meant health care, this will probably shrink the pharmaceutical market by a non-trivial amount. Hopefully the effects will be long lasting and therapy won't have to be repeated (or if it does it will be with months if not years of remissions without treatment).
The people who will benefit from that treatment will probably need to take less or no anti-psychotics daily.
A great win since these drugs have ugly side effects (obesity in the short term, and neurological problems after decades).
My implication was that a lot of money is spent on antipsychotics and if this non-pharmaceutical approach can be automated it might be able to capture some of the antipsychotic medication market for very little cost.
This reminds me once again how much of our reality is in the brain. So, if you "think" you hear voices - you do, if you think the thing you hear are just your thoughts running wild - you do that and act accordingly. I guess in the first case you would be terrified (after all you hear voices) and in the second case you could be amazed (what a cool thing your brain is since it can imagine voices, "their" thoughts, etc.) How much of the bad things in this world and in life come from wrong perspective... :(
I don't think an abundance of negative, however vivid, intrusive thoughts could leave anyone in a positive state of amazement. Maybe if they were positive?
In the book he describes an innovative but quite lo-tech approach using a cardboard box and cleverly arranged mirrors that enabled patients to visualize the missing limb and effectively, as I understood it, rewire their brains to reduce phantom limb pain.
That example does make me wonder how critical the digital component of the treatment is. The article notes:
> Patients were then encouraged to engage in a dialogue with the avatar, who was controlled by a therapist.
I wonder if this would work with similar effectiveness using puppets or actors or perhaps even mirrors.
[0] https://en.wikipedia.org/wiki/Phantoms_in_the_Brain