Smells boost long-term memory

A key question in neuroscience is how are memories “locked in”, or consolidated, so that they last in the long term? Speaking with Chris Smith, Laura Shanahan, at Northwestern University, has been asking volunteers inside a brain scanner to memorise things that she pairs with different smells...

Laura - We wanted to understand this process of memory consolidation, and how it can be influenced by sensory stimuli like odours and sounds during sleep. I think the best way to explain this process is to talk about the first reactivation experiment that happened about 10 years ago. So, in this experiment, human participants learned a visual spatial memory task while they were smelling an odour. And in this case it was a rose odour. The idea being that, if the scientists delivered the same rose odour later on while they were sleeping, it might help to improve their memory. When they woke up and they did find this was the case actually. So when participants woke up from sleep they were better able to recall those memories that were linked to that rose odour.

Chris - So you show people something you want them to remember; you present that thing alongside the smell of a rose. And then when they're off sleeping you represent the rose odour and that seems to reinforce the memory recall - their ability to get the memory back - somehow?

Laura - Right. That's exactly right.

Chris - So what was the difference between what was done then and this set of experiments you're publishing now?

Laura - Well, in our experiment we are really focused on the brain mechanism. So how do these stimuli, like odours, influence consolidation during sleep. We used MRI, which is a technique to image brain activity. And the first thing that participants did was go into the MRI scanner and they saw pictures from four categories appear on a four by four grid: animals buildings faces and tools. And the idea here was they were supposed to try to remember where those pictures appeared on the grid.

Chris - So they might see, for example, a picture of a cow and it might be in the bottom left?

Laura - Right. Exactly. So this part of the experiment served two purposes: first, the participants started to learn these object locations. And, second, we could use the MRI data to look at the patterns of brain activity for each subject in response to the different category images. So your brain activity looks different depending on whether you're looking at animals, buildings, faces or tools.

Chris - And then what happened?

Laura - So, next, participants came out of the scanner and they learned to associate objects from those categories with different odours. So, for instance, we might pair a banana odour with animal images; and maybe a cedar odour could be paired with building images, for instance. Next we did a variety of memory tests to see how well subjects remembered those original object locations. So those grid spaces that they had learned previously in the scanner. And then finally we fitted subjects with an EEG cap so an EEG cap is another way to measure brain activity and in sleep research it's often used to tell how deeply someone is sleeping. So we're able to wait until they fell asleep and then we could deliver two of the four category specific odours right to their noses, while they were sleeping in the scanner.

Chris - And the critical point about by using two of the four, you then had the other two as a sort of control, didn't you; the two you hadn't presented versus the two you had you could compare the relative performance of the recall of one so the categories against the other sort of categories?

Laura - That's right. So our hypothesis was that participants would perform better when they woke up for those objects that had been cued. That is if we delivered the banana odour during sleep then they would be better able to remember those paired animal images when they woke up.

Chris - And is that what you found?

Laura - So yeah, in fact what we did see is that participants perform better for those cued object locations compared to the non-queued object locations. And what we're even more interested in was to see what was happening in the brain during this time; and specifically we were looking for those same patterns of brain activity associated with the four picture categories to reemerge in response to those odours. So, for instance, if we presented a banana odour to a sleeping participant, we wanted to see whether the animal pattern of brain activity would reemerge at that time.

Chris - And is that what happened?

Laura - We did see a link between the reemergence of these category patterns and behaviour, specifically in a brain region called the ventromedial prefrontal cortex; and this part of the brain is near the front and it's important for many things; among those is recalling remote memories. So what we saw, essentially, is that the more these category patterns were reactivated in this part of the brain, called the ventromedial prefrontal cortex, the better that participants later remembered the associated material.

Chris - So what do you think is actually underpinning this effect, because obviously we haven't evolved to learn by people presenting cues to us when we sleep at the right time in order to reinforce our memories. So this must be pointing at some other fundamental process that underpins how memory works. So what do you think's going on? What is this telling you?

Laura - Yeah. So I think that the fundamental idea here is that memory replay - the idea that the same neurons or brain areas are active during learning are active again later on during sleep in order to reinforce our memories - this is happening all the time. And when we do this specific reactivation technique, trying to use these sensory cues in order to influence memory consolidation, we're kind of hacking into that natural replay phenomenon that already happens; we're taking advantage of that and using it to direct memory consolidation in a way that we choose in our experiments.

Chris - So normally your brain would naturally be replaying experiences to consolidate the ones it wants to keep and reinforce them. But by pairing the thing with the smell first, what you basically do is make the brain present that replay phenomenon more, so you represent the thing that the smells are associated with more often so it gets consolidated better and possibly faster?

Laura - Exactly. Our research supports that theory by showing a specific area of the brain where this phenomenon of memory reactivation is correlated with behaviour.

Chris - Now why is sleep so critical? Because, obviously, we can remember stuff during the day when we want to. Is it just that the brain's having a noisy a time when we're awake, so it's harder to get the result you're looking for?

Laura - Yeah that's definitely one theory. Sleep has been shown for decades to be a particularly important time window for memory consolidation; and specifically deeper sleep is a time period that's really important. And so in these reactivation experiments we specifically target slow wave sleep delivering odours while subjects are very deeply asleep. And this seems to have the particular impact on memory. In our experiment we never delivered odours during wake; but in previous experiments of this kind it's been shown that delivering these sensory cues during wake don't really cause the same memory effects. So we think there's something unique about the brain and its environment during sleep that allows us to do this...