Bang or BAM? On respecting complex problems

There are simple problems that can be solved with a single bang. The task of understanding the (human) brain is not a simple problem. On the contrary, the classic quote

“The brain, the masterpiece of creation, is almost unknown to us.”

attributed to Nicolaus Steno in 1669 is – by any large – still very much true today. This is owed to the fact that brains – let alone human brains – are essentially unparalleled in terms of their complexity, both in terms of their structure, as well as their function (the activity patterns they are able to produce).

If anything, the past 150 years or so of “modern” research on the brain give us an appreciation for the magnificent scale of the complexity. A single synapse is awe-inspiringly complex. Each single neuron typically has thousands of such synapses (in addition to quite a few other functional parts). Each local circuit contains a plethora of many different varieties of neurons. In the interest of brevity, I will skip a few levels of organization here, but a whole (human) brain contains on the order of 100 billion (or slightly less, but regardless of the precise number about 10 for every single person on the planet or in the neighborhood of the number of stars in the galaxy) neurons (and about an order of magnitude more glia cells, which might play functional roles as well), organized in intricate functional structures, constantly producing dynamic electrochemical activity patterns that in turn change the structure of the synapses and neural connectivity that produced these patterns.

From this concise outline, it should be obvious that understanding the human brain (or any brain for that matter) is not a simple problem.

A great deal has already been written about the Brain Activity Map project (BAM), so I will make this short.

As welcome as the money will be to researchers who have gotten used to funding rate in the single digits in the past decade, it is important to be realistic what 3 billion dollars can buy you.

Depending on the number you use (total program cost, procurement cost, unit cost or flyaway cost), a single Northrop Grumman B-2 spirit bomber costs on the order of 1-2 billion dollars. For the purposes of this argument, I think it not unreasonable to argue that 3 billion will buy you two operational ones. This is a conservative estimate. What matters for the purposes of this calculation is the ultimate cost to the taxpayer which is – if anything – higher.

The US made 21 of them (20 are left, after one of them crashed in Guam in 2008. After originally ordering over 130 at the end of the cold war, but cutting back to 21 after the fall of the Soviet Union). Ever since they became available, they participated in most major campaigns, e.g. pacifying Yugoslavia in the late 1990s.

BANG or BAM? Can we understand the human brain for the cost of these?

As awe-inspiring as these marvels of advanced technology might be, as unparalleled their ability to rain down death and destruction with impunity, they are designed to solve relatively simple (as in tractable and straightforward) problems.

To meaningfully understand structure and function of the human brain, I think that it will take money on the order of magnitude that would buy a fleet of B2 spirits that will blacken the sky. And that is just the funding aspect of it. Money will be necessary, but it won’t necessarily be sufficient. Could the Manhattan project be pulled off without decades of “basic” (there is nothing basic about basic research) research and some fortuitous insights by Einstein himself? Probably not, no matter how much money one threw at the problem. There is a place for a “targeted science” approach in neuroscience, but I will focus on this in another piece.

It is not a bad thing to have challenging goals. It is also not a bad thing to spend money on research (effectively spending money on understanding the world around us). But it is important to be realistic about the magnitude of the challenge and the magnitude of funds devoted to it. There needs to be a balance. If not, money is either wasted, or one sets oneself up for failure, or both.

Human societies can accomplish a great deal if they put their mind to it (in close analogy to reliable computation with unreliable components). We got Buzz Aldrin to the moon and back, safely. But we could not protect him from being ravaged by depression for decades afterwards. Understanding the brain matters, and we do not understand it yet. Our ability to mitigate suffering originating from the human brain is at the very beginning. For instance, it has been suggested that current therapies for clinical depression (most based on manipulating levels of monoamine neurotransmitters) are beneficial for only one in ten patients.

We did build the B2 bomber. We do not yet understand the human brain. I think the latter can be done, but we need to get our priorities straight and have the scale of our challenges match the scale of our resources. This is a political discussion – the US annually spends hundreds of billions of dollars on defense. It is not a foregone conclusion that money spent on keeping world peace is necessarily misspent, but there are many competing policy goals. In times of scarce resources, nothing should be sacrosanct, all options should be on the table.

I’m all for getting serious about understanding the human brain. But in order to do so, we need to actually get serious about understanding the human brain.

What do we want? A bang? A big bang? A BAM? Or rather, a really, really BIG BAM?

If one wants to begin to start understanding the human brain or – with apologies to Churchill – end the beginning of understanding the human brain, one could argue that we need a really big BAM, or more than one.

From what we already learned about the brain, it is clear that the gains in understanding are worth the tremendous cost and effort. However, it is equally clear that it won’t be possible to gain substantial further understanding on the cheap or quickly.

Adequate relative scaling matters. For prospects of success.

Does the brain deserve some respect?

Regardless: How much would an actionable understanding of brain function be worth to you? How much would you be willing to pay, as a personal share?

Update: It now seems that the numbers are in. Way to politicize neuroscience for $100 million. For comparison, people in the US paid 320 times that in overdraft fees in 2012 alone. It is not unreasonable to think in dimensions that scale with the magnitude of the problem, even if the absolute numbers are starting to get a little high. The total projected life-cycle cost for the F35 program is just over 1.5 trillion dollars. The US built over 12,000 B17 bombers to win WW2. They also built a large number of B24s, B25s, B26s and B29s. It is all about how serious one wants to be about addressing an issue. An estimated billion people suffer from some kind of brain disorder right now. To say nothing of the untold billions who will suffer from one in the future. So even a multi-trillion dollar investment might not get a bad return on investment in terms of reduction of suffering if it yields anything tangible. Can we afford it? Can we afford not to do this? Some people wonder how people in the dark ages got by, without the benefit of understanding the source of all their troubles with regard to bacteria and viruses. I wonder if people in the future will ask the same question (mutatis mutandis) about us.

That said, it is undoubtedly true that we are in urgent need of new tools to probe the brain, as our available methods are woefully inadequate for the task. And the new acronym – BRAIN – is also much improved. Hedged excitement as the most suitable way to move ahead?

Regardless, one needs to be cautious of overselling. The history of AI research provides a cautionary tale, showing that in the long term, overly hyping something without being able to delivery inevitably leads to backlash, which leads to funding cuts (“winters“).