In 1998, Nicola Clayton and Anthony Dickinson published a paper in Nature reporting results of experiments that showed that scrub jays had “episodic-like” memory. This was the first behavioural evidence for this kind of memory from a non-human animal. Eighteen years after the paper was published, I spoke to Nicola Clayton about the making of this study and what we have learnt since about episodic memory in animals.
Citation: Clayton, N. S., & Dickinson, A. (1998). Episodic-like memory during cache recovery by scrub jays. Nature, 395(6699), 272-274.
Date of interview: 14th July 2016
Hari Sridhar: What was the specific motivation behind doing this study?
Nicky Clayton: It came out of an interest in food-caching behaviour.
HS: And where did that interest come from?
NC: Well, I have always been a bird watcher and have observed and wondered about this behaviour numerous times.
HS: How did the collaboration between you and Anthony Dickinson start?
NC: Tony and I first met at a conference in 1996. In fact, at that conference, when I first suggested to him that jays might have memory he was completely dismissive. He said: “I can’t think of a single reason why they would need it”. And I said “Well, what about the fact that they hide perishable food? Wouldn’t it be important for them to be able to remember when or how long ago their cached their food, in order to know when they needed to recover their caches of food?” He didn’t know anything about the caching of perishable foods and so he said: “Oh gosh, I hadn’t thought about that”. So then we went away and spent much of the rest of the conference talking and discussing. I told him what I knew about the behaviour of the birds in the wild, from my observations of them, and it was at that point that he said: “Oh well, I just made an assumption, I didn’t know about this”. And I said “Well, this is something that we could test empirically”. And he said: “Absolutely, yes”. So then we got into a collaboration designing experiments to test whether or not the birds could remember the “what, where and when” of past caching episodes.
Around the same time, Endel Tulving was on sabbatical at the University of California, Davis, where I was based. I talked to Endel a lot about what, from a human psychological point of view, he would find convincing as the evidence potentially for episodic memory in animals. Obviously we called it “episodic-like” because we have no way of knowing whether or not they have the associated phenomenology that would accompany the conscious experience of episodically recording the past in humans. But in the absence of agreed behavioural markers in animals there is no way of knowing. Endel agreed that episodic-like memory was a fair term to use since we were just talking about the behavioural criteria. He also said that if the birds were to do this it would be a serious challenge for the work of cognitive neuroscientists. While he was not sure that he would go as far as saying that that showed that animals had episodic-like memory, it would certainly cause him to tighten up his definition of what episodic memory in humans is.
So we did the experiment and to our amazement we were able to show that jays do remember the “what, where and when” of specific past caching episodes. We, obviously, controlled for relative familiarity, which is an important factor to be controlled for in the human research. Receiver operating characteristics suggest that familiarity and recall are distinct and separate processes, both psychologically and in terms of areas of the brain. So we controlled for that and showed that, yes, they did appear to be remember specific “what, where and when” of caching episodes.
Then, having talked to Endel about it, we thought that the work was of sufficient salience to be submitted to Nature. We did that, they just sent it out to review and then, a couple of months later, I heard that the paper had been accepted, almost with no revisions actually. I think the reviewers were John O’Keefe, who won a Nobel Prize recently for his work on place cells, and Endel Tulving. So that is sort of the brief history of this paper.
HS: In the last line of the paper you say “This is the first conclusive behavioural evidence of episodic-like memory in animals” Since then has it been discovered in other animals?
NC: It has been shown in other animals, but that was the first demonstration. Some of the other studies haven’t really controlled for things properly. I mean we didn’t do everything in that paper but, obviously, in subsequent papers we were able to show that they can also remember how long ago things ripened rather than perish, and that they can keep track of multiple different types of perishable food. So it’s not just that they remember after a short delay and forget after a long delay. We have been able to show that they actually remember the different types of food items in the integrated episodes. But since then, yes, it’s been shown in chimpanzees and even rats.
HS: Did you do all the work related to the experiments – training of birds, care of birds, etc. – yourself?
NC: Well, I had a team of people. So for the first experiment I did pretty much all of it myself, because I was an assistant professor setting up my lab. But then, obviously, as my lab grew I had PhD students and postdocs to help. Kara Shirley Yu was the first research assistant I had, and she was a co-author on a number of the initial papers. She helped me collect the data, but was not involved in helping design the experiments, analysing the data and writing it up. Most of that I did on my own or in conjunction with Tony. Whilst I was in Davis and he was in Cambridge, we used to talk on the phone once a week, and then I was invited to apply for a position in Cambridge, which I got. Obviously, since then life’s been a lot easier because Tony and I have been in the same place.
HS: How long did the writing take?
NC: I am a fairly quick writer, so I don’t think it took more than a month or so to write it up.
HS: During the writing, did you communicate with Anthony Dickinson mostly on the phone?
NC: Yes, Tony and I spoke mostly on the phone during that period, because I was in California and he was in Cambridge.
HS: Did he visit California anytime during when the experiments were being conducted?
NC: Oh yes, he came to visit and spent about three weeks, I think, in California.
HS: Did this paper have a big impact on your career and your future research?
NC: Yes, definitely. I am a Fellow of the Royal Society now and I would say that the Nature paper was one of the big influences. To be elected, the key question is what contribution you have made to science, but also, specifically, they want to know if you have changed the field. Is the field different as a result of the work that you have done? I think the work presented in this paper was one of three key pieces of evidence in favour of my case. The second was the work on the social side of caching, particularly the research I conducted with my husband Nathan Emery, showing that it is only experienced thieves who re-hide the caches when others have watched them cache. That was also published in Nature. Thirdly, there was another paper in Nature on the jay’s ability to plan for the future, which is a necessary follow-on from the episodic-like memory. In addition to the empirical findings, another seminal piece of work really was a review paper I published with Nathan Emery in Science suggesting that jays and other members of the crow family are as clever as the great apes. So when it comes to intelligence you can think of them as ‘feathered apes’ – that was a term Nathan coined. So I would say that these were probably the key papers. And the power comes from converging evidence. If you have just got a one-off study, well, that’s interesting, but that’s all it is. It is only when you have a body of evidence all leading in the same direction that things become really convincing.
HS: This paper has been cited over 900 times. At the time when it was published did you anticipate that it would have such a big impact?
NC: No, I had no idea.
HS: Do you know what it mostly gets cited for?
NC: There’s probably two main strands, I think. One is memory research in general – mental time travel, episodic memory, future planning etc. And the other is within comparative cognition. For example, our work showing that these birds are as clever as the great apes. I think it’s been cited quite a bit in comparative cognition and even in developmental cognition as well.
HS: The material you used in these experiments – peanuts, worms, lego blocks, ice trays – do you continue to use these in your experiments today?
NC: Oh yes, I still use ice cube trays full of sand and corn kibble and a few Lego Duplo bricks.
HS: I read somewhere that when you moved from California to Cambridge you took the jays with you? Do you still use some of these birds for your experiments?
NC: Yes, I made that conditional on my job offer. I said that I would be delighted to accept the position but only if I could bring my jays with me. So I requested 25 avian lectureships! And I’ve still got five of those jays left. The oldest one is 21 now. But I also use Eurasian jays these days, which are easier to get hold of in the UK, and delightful to work with. It’s a real honour to be able to work with corvids. They are so intelligent.
HS: Do you know what has happened to the experimental infrastructure you used when you were in UC Davis?
NC: Well, it was used by other people the last time I visited Davis, but I don’t know what’s going on now as I haven’t been back in years. I have so many fond memories.
HS: If you were to repeat this experiment today, would you do anything differently?
NC: I don’t think I would, actually. I think, as a starting point, once you have made the initial discovery, there are, obviously, other things to control for. But I think I would still do that initial experiment in pretty much the same way, because it was guided by the bird’s natural behaviour. That’s where the inspiration came from.
HS: Were the main findings of this experimental study substantiated by evidence from field studies later on?
Not of this experimental study specifically, because field studies aren’t amenable to comparing birds with and without experience of degraded worms, but the studies on the social cognition of Corvids have been replicated in both field and lab studies, and other laboratory studies have found evidence of this specific experimental study on memory (namely episodic-like memory) in other Corvids and also other species of birds and mammals.
HS: Were the findings controversial when they were published? Did they attract a lot of attention?
NC: It attracted a lot of attention and I was very lucky to have people like Endel Tulving supporting me. A lot of people in human memory research probably wouldn’t have looked at my research on birds. It wasn’t be a species of bird that most people in this field hadn’t even heard of at that time. But Endel Tulving told everyone about it. So it immediately got the attention of other influential people in the human memory work, who have become good friends and colleagues since. People like Dan Schacter, Morris Moskovitch and others.
I know at that time I was worried that people would develop a rat model and the jay work would just be forgotten. The rat brain would be so much more amenable to doing neuroscience studies, because we know so much more about the rat brain. Also, the rat brain is, obviously, a better model if you are interested in mammalian brains. I was also not willing to do invasive work with my scrub jays. With rats you can just order them and specify the number of males and females and there they are. With my scrub jays, I have to find the nests and hand-raise the chicks and all that sort of thing, so they are too valuable a resource for me to do invasive work. So given all this, I was worried that people would just turn to rats and the jays would get ignored, but that wasn’t the case. I was very lucky I think.
HS: Have you ever read the paper after it was published?
NC: Yes, but only in terms of refreshing my mind about what we said at the time.
HS: What would you say to a student about to read this paper today? What should he or she take away from it?
NC: The same as what it was then really. I think it still remains the same. When Endel Tulving coined the term episodic memory he talked about spatio-temporal relations. I would then point out how his definition has changed over time to include integrated memories of multiple features, which we showed in a subsequent experiment. I would point out the issue of distinguishing between recall and familiarity, which was really developed by the time of his 1983 book. And then I would point them to the idea that, since then, people have made this link between episodic memory and future planning in the form of mental time travel. The influential work of Dan Schacter and others show that these memories are primarily for the future – The Prospective Brain hypothesis. But I don’t think I would have wanted to make that initial paper more complex by having future planning in it. Nature papers are very short, so you need quite a simple design. And I think the trick to Nature papers, from what I have seen, is, really, to have a couple of really solid experiments that provide the firm footing in other journals. Then if you do a really exciting experiment and it works, then you can publish a short paper in Nature with this body of other evidence in other journals to substantiate your claim. I think a lot of people have done that. That’s the technique that, for example, Richard Morris used in all his rat spatial memory water maze work. I would say it’s also the technique that John O’Keefe used with his seminal work on place cells.
HS: Among the papers you have published, is this one of your favourites?
NC: Yes, it’s one of my favourites because it the first of my papers to reveal the remarkable cognitive abilities of these wonderful birds.