What happens in our brains when we’re trying to be funny

Jean Mary Zarate: 00:05

Hello and welcome to Tales from the Synapse, a podcast brought to you by Nature’s careers section in partnership with Nature Neuroscience. I’m Jean Mary Zarate, a senior editor at the journal Nature Neuroscience, and in this series we speak to brain scientists all over the world about their life, their research, their collaborations, and the impact of their work.

In episode three, we chat to a researcher and amateur stand up comedian who explores what’s happening in our neural networks when we’re trying to be funny.

Ori Amir: 00:41

Hey, I’m Ori Amir, I am, I guess. a professor at Pomona College. That is a college in the US in California, not far from Los Angeles.

My research was all over the place. I studied everything from how visual processing happens in the brain, to, you know, moral processing, moral cognition. And finally, I also studied creativity and specifically how humour is perceived and created in the brain.

So how the brain essentially understands humour, and how it generates, like, new jokes.

Ori Amir: 01:33

So I was born in a small village in Israel called Tel Aviv. And I grew up there. I was miserable.

I had some good times, but mostly miserable. But I started in Israel, I did my bachelor’s there. I thought I wanted to become a clinical psychologist, to treat myself mostly.

But then I realized after one year that it’s not my thing, but I really enjoyed the science of psychology, and I got more and more into neuroscience and artificial intelligence.

That was Israel, and then I realized that, you know, maybe I can switch scenery a little bit. Maybe I can go and do my advanced degrees in the US. So I went and studied at the University of Santa Cruz in Los Angeles. I did my PhD in neuroscience essentially.

I did a lot of brain imaging. I had my brain scanned way too many times during this process.

And at that time, I also realized that there isn’t a possibility, you can actually do stand up comedy, like if you want. You can just go to an open mic, or like a comedy club some, some nights and just try it out, to just do it. So I decided to try it. And after the first moment, I got hooked.

Ori Amir: 03:02 (clip from stand-up show)

So I wasn’t always a foreigner. I was always creepy. I grew up in Israel, where I was a creepy local.

I love being a foreigner, I can say the craziest things. If you go “Well, I guess it must be normal in his country.

“You say creepy, I say tomato. It’s culture differences. If you don’t like it, go back to where I came from.”

Ori Amir: 03.15

And I ended up doing that as a serious hobby throughout my time in the US.

And at some point, I realized, well, I live in Los Angeles, I know a lot of famous comedians. And I have an MRI, access to an MRI machine where I can scan the brain and see what’s going on in the brain when you are in the process of coming up with funny ideas.

And at that time, there was no research on what goes on in the brain when you are being comedically creative, when you’re actually coming up with a funny idea.

There have been some studies, you know, maybe 20 studies about what’s going on in the brain when you are processing comedy, enjoying comedy, but nothing about the creative process.

And I figured, okay, I know these famous or sem-famous comedians and I have an MRI machine, not far from Hollywood. I can just bring these people to the MRI machine, scan their brains, see what’s going on. And so that’s 99% of my life story. Nothing much else has happened.

Ori Amir: 04:55

So I wanted to see what goes on in the brain of comedians when they are in the process of coming up with a funny idea.

Now the MRI is quite limited because it’s a very noisy signal. And there is a lot, you know, in the brain itself. Whenever you, whenever you do something as complex, like a lot of it is working.

And it’s not necessarily parts of the brain that are critical for writing comedy exactly. It’s, you know, you’re looking at a picture, you will see visual cortex activated. You know, you are thinking about what to say. You have language areas activated.

So you want to be able to have multiple events of coming up with a funny idea within a certain window of time, so that you can average out this activation and see some kind of reliable signal from the noise.

And you have to, so you have to give them a task that pretty much forces them to do that.

So the New Yorker captioning task, essentially, you have a cartoon of a mouse pointing a revolver at a cat. And you have to come up with a caption. Now the winning caption for this particular image was “Six rounds, nine lives. Do the math.”

So that was, but that’s the task, the task is to come up with something funny that one of those characters would say in the situation, right. So that’s, that’s the task.

So you see multiple, multiple such images, you have to come up with multiple captions. And you need a control condition that pretty much has all of the elements of the experimental condition, but leaves out that part of having to come up with a funny idea, right?

So you basically want to show them the same kind of cartoons, and have them come up with a caption, but this time, the caption should not be funny. Just, you know, what you would normally hear in the situation.

So this way, in both the control condition and the experimental condition, you have the part of the brain that does language, the part of the brain that visually processes the image, but what separates the condition, if you contrast them, if you subtract the activation of the control from the experimental condition, then hopefully what you’re getting is specifically those areas of the brain that are uniquely involved in generating funny ideas.

So we had 13 professional comedians, nine amateurs, and I believe it was 19 controls. So the idea is to see the difference between, you know, professional comedians and controls. But also to see if there is some kind of, you know, continuum there.

So what we saw was that the professional comedians have more activation in temporal cortices, certain areas of the temporal cortex in the front of it, where it is sort of like a high level, semantic area, the kind of area where associations from different parts of the brain sort of converge.

So if you have some remote concepts that usually don’t go together, but you want to find a way to link them in a meaningful way, you would go to that cortex, that part of the cortex, yeah.

So you see more activation there, the more experience you have doing comedy.

So there’s a lot of sort of, sort of associative sort of brainstorming going on. And you see less activity in the prefrontal cortex, which is essentially the area in the front of the brain, where it’s one of those areas that developed late in evolution.

So humans have a particularly large prefrontal cortex. But what we think it does in this context, and why we think that the more experience you have doing comedy you see less activity there, is because it has to do with control over the creative process.

So that’s the area that would tell you the goal of what you’re supposed to be doing. So “Oh, I’m supposed to come up with something funny,” and you’re supposed to do this, supposed to do that.

That’s like the conductor of the orchestra, essentially. And I guess you can say that you need less of that, the more experience you have doing comedy.

Maybe it’s even in the way of coming up with, with a particularly novel, and original idea, because you’re sort of like, sort of hands-on directing the process. If you let it have a little bit of freedom, a little less control, you might actually come up with more novel, original ideas.

Another thing we saw was greater activation in the striatum, the part of the striatum that is involved in reward processing before the professional comedians were coming up with the funnier ideas.

So the funnier the joke would be, then the more activation you see in the striatum before they come up with it.

And so that’s somewhat somewhat tricky to interpret, because it could mean one of two things.

It could mean that you need to, sort of, like set up the environment. So like the general feeling in your brain, so that you will be more inclined to come up with funny ideas.

Or it could mean that they are just really good at predicting, “Oh, now I’m going to come up with a funny idea.” This has a lot of potential.

So that prediction is rewarding. I guess that might be less of a clear cut kind of story than the previous one. So you may or may not want to cut it out.

But if you want to translate it to advice that you can give a comedian is like try to have more fun and you will be funnier, right.

So if you’re having more fun, you activate your reward regions more. And that might give sort of like a mental context. Sort of a sense in the brain where humour is more likely to spontaneously emerge.

The other advice, the more like immediate advice, is pretty much a confirmation of the advice that comedy coaches, especially improv comedy coaches, have been giving for a long time, which is, “Get out of your head.”

And by “get out of your head,” if you think the actual meaning of it is to let your mind flow naturally. You know, don’t don’t try to, you know, force the direction of where your associations go.

But in terms of what it means, in terms of the neurosciences, you know, don’t over activate your prefrontal cortex, let it rest a little bit, and let the associations in the anterior temporal cortex do their thing.

I do enjoy jokes that are logic-based and are absurdist, and have some commentary about how silly is, how people behave or think. Just some commentary about that. So essentially, parodying, like the acceptable narratives. I would say that’s my favorite. My favourite stuff I’m trying to do, and that’s what I enjoy the most listening to.

Ori Amir: 13:48 (from stand-up)

If y’all want to have a PhD like me, here’s what you’ve got to do. It’s going to take seven years, the first five and a half years to work very hard on developing a silly accent.

Then you do some original research and it all culminates in a dissertation defence in which you present your work in front of five important neuroscientists. And if you fail, they eat your brains.

Ori Amir: 13:55

I noticed that when I started doing comedy, and I wasn’t as good, people were telling me “Well, you know, you will have a hard time making it because of your accent. And then once I became good, people will tell me, “Well of course you’re funny. It’s because of your accent.”

Ori Amir: 14:05 (from stand-up)

Yeah, my dream is to become a professional comedian and an amateur neurosurgeon. This way I can just cut brains for fun.

Ori Amir: 14:15

You know, the one place where my comedy fails horribly, like without any saving grace, is whenever I try to go back to Israel and do comedy there.

So when you’re talking about the appreciation of humour, and you can, there are two types, two major types of studies. One that looks into humour versus non-humour, and see like what areas of the brain are activated. So you actually find, similarly, that temporal, these areas in the temporal regions appear to be activated to a different, to a different extent, or with a different timeline.

So, when you’re getting a joke then you have a quick spike in areas like in the temporal areas. Whereas when you are constructing a joke when you’re being creative, then you see a gradual increase in activity there.

So that’s like one of the differences you might see. Otherwise, you would see reward regions activated, as you see in human creation.

But again, the activation has a different timeline. It happens after getting the joke as opposed to when you create a joke. Apparently, it comes before you fully formulate the joke.

And the other type of study looks into the differences between different types of humour. So, and there, it’s actually not particularly surprising.

So if you have visual gags, you would see more activation in visual areas. Or if you have language-based humour, you see more activation in the language area.

If you have humour that relies on understanding other people’s mind, viewpoint or whatever, you have activation in regions involving in theory of mind, which is a fancy word for just understanding other people’s perspective.

And then there are studies that try to sort of break down to different parts of humour processing. And that comes from different theoretical perspectives.

So there’s like a researcher from Taiwan, I believe, that recently just published a lot of studies about humour in the brain, and they were looking at the different stages of humour processing.

So you start from realizing that there is some kind of incongruity in the narrative, right? So the setup and the punchline result in some kind of incongruity based on your original interpretation of the setup.

And then you have to sort of change your perspective. You have to revise your understanding of the setup so as to reach a resolution.

And this process, and then you find it funny. It’s like one theory that I feel does not cover all types of humour but it’s definitely a prominent theory of what is humour.

So that’s incongruity resolution hypothesis. And they show, okay, see that area of the brain is involved in the detection of incongruity, and that area is more active during the insight or the resolution part.

So I am working on some projects that attempt to use artificial intelligence to generate, like, clever humour.

So there have been a lot of work on you know, that uses artificial intelligence to make essentially patterns or well like structured jokes, like your mama’s jokes or something like that, or “I like my man how I like my kind of jokes.”

So these jokes have been possible to make quite successfully using artificial intelligence for 20 years or so.

The humans usually have to select the ones that makes sense out of the multiple things in the output, but I’m working on a couple of projects that I can’t maybe completely reveal that might help people in marketing for example, use artificial intelligence to generate jokes that are appropriate to what they’re trying to make, right?

So basically, okay, I have a commercial about Coca-Cola and a polar bear and I need some link between those two concepts that have comedic potential. So it’s supposed to help with that

What I did see in terms of what’s already out there is a pretty scarily good algorithm that explains jokes.

So it’s based on, you know, some general language model of Google that is not completely, as far as I understand it, available to the general population yet.

But it’s basically, you give it a joke and in a German fashion, just to explain, “Well, the reason why this, you know, it is funny is because the dog has died. And because War and Peace is a very big book, but the manual of the company is also very big.”

And so it’s referencing, it explains the jokes, but it does a very good job of, like, as good as it gets. And so, that’s scary.

Like, I want to see if the inverse is possible, like if you can just write a serious bunch of serious statements, and it translated into humour, which is like one of the projects we’re trying but that’s, you know, that that’s probably, that’s a very difficult problem, for sure.

That would be the killer app of humour. Of humour creating AI would be basically like a Google Translate, but instead of translating from one language to another, you give it a serious statement, and it makes a joke that essentially says the same thing.

So I think humour is probably one of the hardest problems for AI. It’s what we call AI-complete, meaning, you have to pretty much be able to do everything that humans can do in order to do humour, right?

So in order to do humour, you would have to be able to master all other human cognitive functions.

Whether it will happen and when it will happen, I used to be more certain about it. I used to be in the camp that was like, “Okay, AI is definitely coming, it’s definitely coming soon. It’s like, you know, less than 10 years away, and it’s going to replace us and going to exponentially, you know, improve itself.

After that point, assuming nothing will go wrong and explode or whatever, which is very possible, if we, if you do achieve any kind of level of artificial intelligence, that has self control.

And now, you know, I give room for the possibility that it might not be possible, more so than I did in the past.

But put it this way, I think if artificial intelligence, human level artificial intelligence, that sort of self improvement is possible, we’ll find this out in the next 10 years.

I’m afraid that if I make any jokes about artificial intelligence, I will get in trouble in the future. Artificial intelligence would cancel me. So I’m refraining from making any such jokes.

Jean Mary Zarate: 23:39

Now, that’s it for this episode of Tales from the Synapse. I’m Jean Mary Zarate, a senior editor at Nature Neuroscience. The producer was Don Byrne. Thanks again to Ori Amir. And thank you for listening.