With respect, your idealised assessment still has the problem with how that translates to depression, or (after) lack there of.

Definitely agree that our understanding is only really beginning in these kinds of areas though ultimately you are often going to be treating a disorder of subjective experience with these drugs. That makes it very hard for us not to be wishy-washy these days

I feel like the points you two are stating work best in concert. We do need to start asking more empirical questions about "what is depression in this person" and obviously can't do that without substantial amounts of data to drive our investigation, and this is where I agree with the original post that we're very lacking. To me at least, the first step in handling something subjective is trying to define it, and for that we need to better understand it.

Unfortunately our perceptions can't be boiled down to chemicals in the brain. You can be depressed and have normal chemical levels. Just because increasing some chemical will make you happier, doesn't necessarily mean that's the only way.

Until we're able to analytically evaluate perception (which we are no where close to understanding), what you're describing isn't possible

> You can be depressed and have normal chemical levels

Maybe if you only consider the chemicals that are normally considered.

There is no reason that chemicals are the only actively defining part of our mind. It may very well be that chemicals are everything we need to messure, it may also turn out that we need to messure a lot of other things. Point is we do not know, and have a growing problem to figure it out.

What exists in our biology but chemicals?

Structure. An integrated circuit is just doped silicon, but understanding its chemical composition tells you practically nothing about how it actually works.

This "chemical composition" is like reading packets off a broadcast network. It's not the full state, but it's good enough to reverse-engineer what's going on at a high level.

Blood work, including blood neurotransmitters and hormones, can tell you a lot about someone's state of mind specifically because of this.

Present-day science is driven by placebo-controlled research.

Your "need to be able to say" can have no effective equivalent placebo, and any conclusion you drew from such a case study would be unpublishable and invalid.

This is not to say that I think modern science is correct in this, but it does already recognises the flaws in what you suggest - recording accurate data is not necessarily an existing major problem of science. You can easily record all that data yourself (if you have the resources/cash/time to arrange it), which will probably work out great for you, but it won't prove anything more generally.

This is how I train athletes - very scientific, algorithmic, and so forth. We measure literally everything we can with force plates, EMG sensors, high-speed biomechanical video analysis, etc. Yet this technique is still mostly looked down upon by the usual suspects. Fortunately, the top tier of the athletics world is moving in this direction.

Who are the usual suspects? I'm not aware of the current situation in sports training.

The professional world, mostly. The people that have the most to gain are the least likely to look into stuff like this.

The Moneyball revolution in sports has taken over front offices, scouting departments, and stats geeks. But not the holiest of holies, the world of player development. Keeping a player on the court for a few extra minutes per game could result in a huge gain for the team. Keeping Lebron James or Troy Tulowitzki healthy could literally be worth $20 million or more to the team. Yet they pay nothing on that side of the game and do not look for new avenues or approaches.

It is insanity. But it is also professional sports, so, it is what it is. Eventually, there will be a revolution. I only hope it's within the next 20 years.

Totally agree. Science is the art of measuring things and we don't even approach what's easily possible today in scenarios like yours.

Some day we'll look back on our methods of prescribing and follow-up with the same horrified face we make when we look at Civil War amputations.

On the other hand, continuing scientific advancement requires an exponential amount of resources and I think we may see a high watermark where we have to come to terms with the unsustainability of objective progress in terms of perpetual growth.

We already look back at the early 20th century futurism's optimism as comical, but if we manage to hold our civilization together for another century or more I suspect we'll look back at today's faith in the practical power of data science as equally naive.

Take a stroll through the clinical descriptions of some common medicines and see how often you see this phrase:

"The exact mechanism of X is unknown, however..."

It's kind of frightening.

(two off the top of my head: http://www.drugs.com/pro/klonopin.html & http://www.drugs.com/pro/prozac-capsules.html)

> Doctor asks, "Well, how do you feel?"

This is why I am working on a line of medication whose only effect is to make people tell doctors that they feel much better now. It's showing great promise in early trials.

Shock therapy had some interesting results in regards to this.

http://en.wikipedia.org/wiki/Electroconvulsive_therapy

Why would you choose endpoints that may (or may not) be correlated with the desired outcome, rather than the outcome you are interested in?

To rule out root causes. If the data showed increases and serotonin, decreases in stress hormones and improved sleep quality and the patient still replies "bad" to the question "How do you feel?" Then obviously a serotonin imbalance is not the problem (which is all SSRIs are designed to do).

If the patient still feels bad and the data shows no serotonin change, then the doctor can conclude that maybe serotonin is still the problem, but Zoloft isn't the right drug to fix it (or the dosage was wrong).

In other words, more data = better ability to get a real answer.