That Neuroscience Guy - Neuroscience Bites-Bayesian Thinking

Episode Date: November 29, 2023

In today's Neuroscience Bite, we discuss what Bayesian Thinking is, and what it means for your brain. 

Hi, my name is Olof Kregolsen and I'm a neuroscientist at the University of Victoria. And in my spare time, I'm that neuroscience guy. Welcome to another Neuroscience Byte. Okay, I'm not going to lie. This one's going to be a little bit technical and a little bit nerdy. I'm going to be talking briefly about Bayesian statistics and why they matter. The reason I'm actually telling you this is that the idea of Bayesian statistics and Bayesian thinking is creeping up everywhere. It's become a catchphrase or a buzzword.

I do some work with our medical program at the University of Victoria. And even within the medical program, they're talking about doctors making decisions using Bayesian thinking patterns or Bayesian logic. So what does that actually mean? Now, I didn't want to do a full episode on this because basically it's not a stats podcast. It's a neuroscience podcast. And while there's some crossover in my mind, it would be a bit podcast, it's a neuroscience podcast. And while there's some crossover in my mind, it would be a bit much, but I wanted to explain the basic logic and what people are talking about. It's actually not too hard to understand the core idea when someone says that they're using a Bayesian decision-making process. The idea is this. Let's say that you have two choices to make, choice A or choice B.

You might buy pizza or you might not buy pizza. You might buy pizza or you might buy sushi. Well, in a Bayesian world, you have a prior belief. You have prior knowledge about these choices. have a prior belief, you have prior knowledge about these choices. It's basically, in a sense, your memory of what pizza is like and your memory of what sushi is like. All right, that's your prior belief. And when you hear about Bayesian statistics, they talk about priors. Now, when you do statistical models, there's a lot of complex stuff that happens behind the scenes. But at the end of the day, they're talking about this prior belief. It's your knowledge before something happens. Then what happens is you get data, right? You go out into the world and you get data. In this case, it means you either get pizza or you get sushi. Now, what does data allow you to do? It allows you to update your prior belief.

And that's what you do. You basically use the data to change your prior belief. And that's what you do. You basically use the data to change your prior belief, and that becomes what's called in the Bayesian world your posterior belief or the posterior. And it's really that simple. You have a prior belief, you get data, and then you update your prior belief based on the data to have a posterior belief. And then if you were going to make another decision, that posterior belief that you updated, well, that becomes the

prior belief for the next decision. So when people talk about Bayesian decision-making and Bayesian thinking, that's all they're really talking about it. Now the math behind it is fiendishly complex. It really is. But the idea is straightforward. So I thought I'd use that as a little bite because I've heard this phrase, Bayesian decision-making and Bayesian thinking a lot. And I just thought I'd share a little bit of knowledge in case it comes out in your world as well. Anyway, that's all I've got on Bayesian thinking. Remember the website, thatneuroscienceguy.com. Please send us ideas on X or threads at thatneuroscienceguy or email us at thatneuroscienceguy at gmail.com.

We're planning 2024 and we really need to know what you want to know about the neuroscience of daily life. And of course, the podcast. Thank you so much for listening and please subscribe. My name is Olof Kergolsen and I'm that neuroscience guy. I'll see you soon for another full episode of the podcast. Thank you so much for listening and please subscribe. My name is Olof Kergolsen and I'm that neuroscience guy. I'll see you soon for another full episode of the podcast.