AI predicts effective depression treatment based on brainwave patterns

My understanding is that the diagnoses are eventually right, but they take a lot of trial and error. You would not be modelling the flawed diagnostic if you take the final result of the trial and error process.

On the contrary, any measurable biomarker for depression is very good news.

In psychology almost all diagnoses are just checklist of symptoms either reported or observed. If the patient checks enough of them, they have that. Accurate knowledge of the underlying causes and conditions or clear tests that verify them don't usually exist.

Depression is extreme version of this, so called diagnostic trash bin. Symptoms of depression are present in many/most other psychological and physical diseases. If the patient shows symptoms of depression first time, it's relatively common for doctor to order full blood panel and other tests to rule out all other causes.

In the end, if nothing else is found and symptoms continue, all that is left is the proverbial "trash bin". If you don't have anything else, but have depression symptoms you are diagnosed with depression. It's not a surprise that medicating depression is just trying different drugs semi-randomly to find out what works.

If we can directly measure for the effectiveness of treatment, it may help to understand how they work, what depression is, or how many different conditions there are behind depression diagnosis.

Lots of thoughts.

First, there's a sort of framing issue with "subjective answers to questions" per the press release. These subjective experiences are the very nature of the thing of interest: feeling desperate, crying for help, loss of energy, and so forth. To dismiss them is, at some level, is akin to a chemist dismissing oxygen gas as an appropriate target of study because it's all subatomic particles at some level. These types of press releases often start with this notion, in my mind, because it's necessary to create some sense that the study in question is revolutionary.

I do think they're something of a bootstrapping problem with supervised approaches, which is why in psychopathology and psychiatry research there's a trend toward modeling symptom/behavioral presentations using "unsupervised" approaches. This is partially the impetus behind NIH RDoC, for example, and when you do this, you end up with similar but not identical constructs as DSM diagnoses. The recent DSM (DSM-5) was supposed to address this in part but got mired in politics.

If you look at the actual results, by far the most important features in predicting future symptom state are previous values of those symptoms. So follow-up symptom A is best predicted by baseline symptom A. This autocorrelation importance is a sort of law of behavioral individual differences. The EEG variables are adding in the prediction, but relatively weakly, which further speaks to the central importance of the subjective variables, which is informationally mediating whatever is going on under the skull.

This is really further underscored if you start thinking about things like this: are the EEG signals adding most to prediction of "physical" symptom improvement? There's also a literature about severity of symptoms; it's possible to make predictions of symptom response based on other criteria from previous studies and this paper does little to address that in a head-to-head comparison. K-fold cross validation is also good, but isn't the same as replicating something on entirely different sets of people with different sorts of heterogeneity due to unknown sources.

It's an interesting paper but people need to be cautions about the hype that goes on in these areas (biological psychiatry and AI).

For extremely long-range correlations you need a lot of optimization and approximations because the problem space is so large. To navigate enormous computational spaces in a reasonable time you need intelligence.

Because the "first order" "data correlation" tools that we use (typical statistical measures like correlation coefficients, means, stddev, etc) are inadequate for capturing the critical correlations in the kind of large, real world data sets that ML is shining in right now.

One example, if you have a thousand dimensional data set, and 3+ of dimensions are correlated, but only for certain value ranges, to find something like that with classical statistics, you would need an intuition and some digging and even then you might not see it.

If you suitably prepare your data and throw it into an ML it's trivial. Think of ML as automation for statistical inference. Because all neural nets really do is learn complex, super multivariate probability statistics - down to individual pixel-pixel relationships, for image data.

Edit: I'll add that one other reason ML works so well is that parameters in nets can learn (and represent) complex functions which are impractical (if not impossible) with mathematical notation. Which is what most complex real life distributions probably actually look like. It's like the bridge between analog and digital math, a sort of topological compression, if that makes any sense

I've been depressed for pretty much my entire adult life (I'm 51). Ten years of Zoloft did nothing for me. Should have quit after two. Eventually, I too, found the magic of psychedelics, combined with meditation and extensive body work. Therapeutic uses of MDMA, psilocybin and ayahuasca have completed transformed my outlook and helped me unwind the trauma stored in my body. I am very grateful. This research sounds very promising.

>>"We know that depression is very heterogenous, and that there are at least 1,000 unique combinations of symptoms that can be diagnosed as depression," said Williams, who is the director of Stanford's Center for Precision Mental Health and Wellness. "We've found that brainwave measurements can be used to help identify which particular symptoms change with antidepressant treatment and which do not."

They're not claiming to detect depression, I don't think there is "sorts" of depression. They're measuring a change in symptoms to evaluate a medicines usefulness.

The headline is misleading but I do feel their work has merit. Years ago I had an isolated Major Depressive Episode. It's clear to me that during that period of time there was something different about my brain.

As the saying goes, "huge if true". All of these are very under-determined, and depression is frequently misdiagnosed as or instead of conditions like ADHD, bipolar, and autism.

Well there are lof or other diseases that employers could require test for and that can harm productivity. But fortunately this is forbidden by law in most countries. So I don't think it's more an issue that it already is.

I was depressed for 10 years. Changing diet, environment, and exercise did not help. Psychiatric medication and therapy helped.

I don't disagree, but when you link a paper, I think you should disclose that you're the sole author.

> Depression is not something to be dominated and rushed away.

Agreed. That is how you end up with constantly escalating self-medication in my experience. When I realized all psychiatry had to offer me was drugs I was highly skeptical of (and yes, I did try a few), I decided to just switch to drugs I actually like. It does the trick, but it could be hurting my long term health more than it's worth. I certainly don't seem to be making any progress in my productivity or the quality of my choices, both of which could use a lot of improvement.

My experience with depression, an isolated major depressive episode that lasted most of a year, suggests that antidepressants were highly effective. I didn't feel like any clues were to be found. It felt much more like a chemical switch in my brain was in the wrong position.

I agree that tech driven solutions sometimes neglect the human aspects of the human experience.

In this case it might be a tool for better understanding and application of the right human interaction as well. It seems the AI is used to find the right type of treatment and even catching circumstances that clinicians would have missed.

A lot of depression cases are purely chemical in origin. At least 20% of patients are 95% recovered once they find the right medication. Keep in mind that the psychiatric definition of “depression” is quite different from the layperson’s definition. It’s not a sadness or melancholy, it’s basically a shutdown of almost all higher-order mental functions.

Look at societal rates of depression across different cultures. We’ve built this individualistic society and shown everyone the “great things” individuals can do. I think there are plenty of reasons american culture specifically allows for depression. I’m happy to hear other hypotheses - I realize my lived experience is not objective truth.

That's a bit dangerous simplification since depression can result in not having someone to talk to. (Or not willing to) Making it a circular: we'd be less depressed if we were less depressed. Maybe not false, but not super helpful.

bottomless joy by one of the modern mystics and his kriya yoga practices

please share detail

I really would like to understand what is the thought process behind downvoting this quiet, personal and respectful comment. Sometimes this community really baffles me.

I can see some on here taking issue with this statement. My current issue with existing modern medicine is that we are very limited by what we can measure but there is a body of scientists that take the existing literature as gospel.

Meanwhile we have materials which have in some shape or form been passed down throughout human history which we willing discard as their models no longer match reality.

This is sometimes the right thing to do, it is also sometimes the wrong thing to do, after all - all models are wrong, some are useful.

On the flipside, a broken clock is still right twice a day.

Empiricism is the best we have at the moment, but we must not overlook the shortcomings of only working with things we can measure.

EDIT we should not kid ourselves with research not being shaped by that which we think to measure or it being shaped by what we can currently measure today. Before we had microscopes, it was very hard to consider germs being a thing after all.

The body is exquisitely complex and we barely understand the mind-body connection. Observations from as far as distressed myofascial tissue to gut flora can affect the mind so it's perfectly reasonable for me that a proportion of patients who experience depression find help via "alternative" means.

I guess another way of thinking about it is that "depression" is actually far too vague a classification for an issue, especially when considering the myriad causes