Opinion – Suzana Herculano-Houzel: Pavlov’s dog’s dopamine

I am frankly amused by some colleagues’ contortions to differentiate humans from other animals, or “smart” animals from others.

A few years ago, at a meeting of animal behaviorists (which I certainly am not), I discovered that they distinguished between “learning” and “conditioning.” Not that they had a very precise definition for either term, but it was clear that for them what could be learned by “conditioning” was not learning.

Why does it matter? Because, for these colleagues, what is learned at school, or in life, is not “purely conditioned”, but learned at the cost of conscious rumination.

Oh how I wanted to see the faces of these colleagues reading the most recent issue of the journal Science, where a group of researchers led by Vijay Namboodiri, at the University of California at San Francisco, proposes nothing less than a revolution in the way of thinking about learning.

The problem is that the simple “dopamine-gives-pleasure” or “dopamine-predicts-reward” version of the story is just too simple. Here at (University) Vanderbilt, we have our own dopamine expert, Erin Calipari, who tells anyone who will listen the various holes in the story. Her beautiful work shows that dopamine does something much more interesting—which I summarize for her as signaling what’s “worth the effort.”

Namboodiri and his team show something similar, and go further: they propose that dopamine signals not only what WILL be good and prospectively worth the effort, but also what was good, and retrospectively already was worth the effort. Once this is done, the animal combines dopamine with the memory of what it has just done and that’s it: there is an association formed between cause and effect.

The difference, which explains a lot, is that it doesn’t matter how bad the predictor is; what matters is the result. This solves, for example, the mystery that roulette and other gambling games are so engaging and even addictive, despite the fact that the chance of profit from your action —the bet— is minimal.

Prospectively, the chance of “your” number coming up when you bet on roulette is slim. But in hindsight, every time your number actually comes out, it’s because… you played the roulette wheel (or it wouldn’t be “your” number). From where: playing roulette (the action) is potentially profitable (the result).

Of course, putting one thing together requires having a memory that you’ve just played roulette – but that’s what the brain does almost by definition: the brain creates representations of events, and if memory is simply the reactivation of those representations, then who it has a brain it has a memory, oh my.

Which is to say that Pavlov’s dog, which was “conditioned” to salivate whenever it heard the bell announcing the impending meal, genuinely learned to associate food with the memory of the bell.

As my father always reminds me: no matter how small the probability, the only people who win the lottery are those who play the lottery, and even if he hasn’t won yet, the fact that someone does is proof enough that the effort can be worth it. the pity.