How hallucinations happen

Have you ever had a hallucination? These are where you hear, feel or see something that isn't really there. We had thought that these experiences were just made up from scratch by the brain. But neuroscientist Paul Fletcher has found that instead the brain uses our past experiences and superimposes these like a filter or a stencil onto what we're doing right now, creating a hallucination. People who are more prone to hallucinations like individuals with certain mental illnesses seem to be better than average at doing this. Connie Orbach went to see why...

Paul - Well, we took some stimuli that are very, very difficult to perceive clearly. They just look like black and white blobs when you first see them, and we asked people to try and identify those stimuli. We then gave them a strong clue by showing them the picture from which those stimuli had been created. When they were back to the original black and white blobby images, they were then able in some cases, to see what those represented. Now we reasoned that if people with hallucinations are more likely to superimpose their prior expectations or their prior knowledge on noisy images or perceptions then they might actually be better at this task. They might show more of an advantage.

Connie - What did you find between the two groups?

Paul - Well, the key difference was that if you go from before having seen the clear template to after, everybody gets better because they've all been given some prior information that allows them to perceive the image more clearly. But we found that the degree to which there was an improvement was significantly greater in people who are prone to hallucinations.

Connie - What we're talking about here is an image. Is it just visual perception or is this with other forms of prior sensory information?

Paul - So, it's certainly the case that across the different sensory domains - touch, taste, hearing, smell, et cetera - that we think that people use what they already know in order to perceive it more clearly. So, I've got an example of an auditory version of this sort of task. It's not identical but it's the same sort of thing. So, I'll play you this stimulus...

Connie - I think it said zoo at the end, but that's about it.

Paul - Yeah, that's very good actually. Most people hear that as a sort of meaningless clangy mechanical bird song type sound. But if you then play the clear version from which that speech was taken...

Recording - The camels kept in a cage at the zoo.

Paul - Now you have that prior experience, we can go back to the original and then it should be much, much clearer because you have that prior knowledge.

Connie - That was with the auditory. Could you show me some of the visual images?

Paul - Sure. So, I'll show you this image here.

Connie - Okay, so it's hard to explain but in the same way that the auditory one was, you could hear it but you can't make it out. It's just a black and white image but it looks like blobs as you explained.

Paul - Can you see anything in it?

Connie - No. I'm really trying to impose an image on it, but actually, no. I can't see anything at all.

Paul - Okay. I'm now going to show you the clear version from which that picture was taken.

Connie - It's almost completely clear. So the picture was of a woman with a cowboy hat on kissing the nose of a horse. And then when you go back, you can clearly see the outline of her face, her hat, and the nose of the horse within the black and white image, but there was absolutely no way I could see that before.

Paul - Right. So, these images were especially designed for that so that people mostly couldn't pick up very much at all. And then we were able to give the clear image and then test them the second time. Just as you did, a lot of people found that they could now see them. But people who were prone to hallucinations and such like, actually showed a greater advantage there. But what's key here and what we wanted to show is, we can understand the mechanism perhaps by which hallucinations arise through appealing to the normal perceptual mechanism. Rather than just speculating on some gross derangement of function in the brain or some lesion or something, we can actually begin to think about the mechanics of how the experience arises. What we're suggesting is, the hallucination just like normal perception is to an extent a creative process based on what you already know.

Connie - It's interesting that perception is an active form and it's not passive. It maybe has been thought.

Paul - Yeah. So, a lot of people in the past have speculated on perception as a very accurate picture of the world that we get simply by listening to what's coming in. Actually, what's coming in very ambiguous and it's also very noisy. We can sample the world in terms of heat, light, force, chemical composition. So, we have these signals that are very useful coming in. But somehow, we have to assemble that into our reality. In order to do that, we really need to bring something of what we already know. Otherwise I think we would view the world like babies. It would be a mass of colours and shapes, and noises that we really couldn't understand.