Food-Caching And Spatial Cognition: Harsher Winters Favor More Caching And Better Cognitive Abilities

One of the large questions in the study of the evolution of animal cognition is why there is large variation both among and within species in various cognitive abilities as well as in brain size. Much research has focused on whether differences in brain size or size of various brain regions are indeed associated with differences in particular cognitive abilities, but it is also important to understand how such differences have evolved in the first place.

One of the excellent models to investigate the relationship between the brain and cognition, as well as the evolution of cognition, is with food-caching animals. Food-caching birds such as nutcrackers, jays, and chickadees, in particular, have been at the center of such investigations because they cache tens and even hundreds of thousands of individual food items, either seeds or invertebrates, during late summer-fall and then rely on these caches for survival during the winter. These birds use spatial memory associated with an enlarged hippocampus compared to non-caching species. Variation among species in spatial memory ability and the hippocampus has been linked to differences in food-caching propensity, but the data have not been consistent.

In my laboratory, we used two species of food-caching chickadees: the black-capped chickadee and the mountain chickadee. We have shown that (1) there is larger variation within species in spatial memory ability and hippocampus morphology and (2) such variation is associated with differences in winter climate, both on a large geographic scale along longitudinal and latitudinal gradients (from regions with milder winters to regions to harsher winters) and on a small geographical scale along an elevation gradient in the mountains as winters are harsher and longer at higher elevations. So black-capped chickadees from Alaska, Maine, and Minnesota have larger hippocampi with more hippocampal neurons compared to chickadees from Kansas and Washington state, and mountain chickadees from higher elevations in the Sierra Nevada in northern California have better spatial memory associated with larger hippocampus containing more neurons compared to mountain chickadees from lower elevations.

Contrary to the popular notion prevalent in social media, there is no good evidence that the chickadee hippocampus expands and shrinks yearly. Instead, most studies show that the number of neurons in the hippocampus is very stable and is similar between wild birds and birds reared and maintained in captivity, but there are differences among populations, with chickadees living in harsher winter environment having better spatial memory and larger hippocampus. We favor the explanation that these differences have evolved via natural selection as in harsher environments caches are more critical for survival and so the ability to cache more food and to successfully find previously stored food should result in higher survival.

It is impossible to investigate the role of natural selection in shaping differences in spatial cognition associated with differences in a winter climate and with reliance on food caches in laboratory conditions, so we endeavored to study this question in the wild using mountain chickadees at low and high elevations in Sierra Nevada in northern California. We have developed a novel method using RFID technology that allows testing spatial learning and memory in many wild birds that have been previously banded with unique PIT-tags. We maintain “spatial arrays” based on 8 RFID-equipped feeders mounted equidistantly on a square frame and suspended among trees so that rodents and bears could not reach it. Each feeder has an antenna embedded into the perch, a programmable RFID board and a door controlling access to food and operated by the RFID-controlled motor. All the bird sees is a feeder with a perch and an opening to food with the door while all electronics are hidden inside the feeder. When a chickadee lands on a perch, its PIT-tag ID and time are automatically recorded.

Credit: Vladimir V. Pravosudov

First, we determine the IDs of all birds coming to our arrays and then we assign each bird to only one feeder that would open the door and provide food to that bird. If the bird lands on a “wrong” feeder, its ID and time of visit are recorded, but the door stays closed. Only when the bird lands on a “correct” feeder matching its ID would the door open so the bird can get food – sunflower seeds.  The spatial cognitive task is for each bird to learn and remember the location of the one correct assigned feeder and to stop visiting all other feeders. Since we know the time of visits to each feeder within the array, we can easily reconstruct the path to the rewarding feeder and estimate the number of errors (visits to unrewarded feeders) prior to visiting the correct feeder. As birds learn, the number of errors goes down with each trial (a visit to any feeder in the array) and so we can measure how fast each bird learns spatial location of one feeder.

Following initial learning stage, we proceed with a reversal learning task in which we change the location of the rewarding feeder and each bird has to learn that the previously rewarding feeder does not provide food anymore and that a feeder in a new spatial location is now providing food. This task measures memory flexibility or how fast the bird can learn that the switch has occurred and start visiting the new feeder while avoiding visiting the old feeder. Using this method, we have been able to test hundreds of birds and to show that there is indeed large individual variation and that birds at high elevation seem to do better at the initial learning task, but worse in the reversal learning task compared to chickadees from low elevation. This was an intriguing and interesting result suggesting that larger memories associated with more food caching at high elevation might interfere with new memories (e.g. proactive interference).

The question now is whether reversal learning ability is under selection. One way to do that would be to show that individuals with better memory performance have higher survival. Another way would be to compare learning performance of different age groups – if birds that have survived for at least one year (e.g. adults) show better performance than the young birds in their first winter (e.g. juveniles), it would suggest that memory is important for survival. The key thing here is to know if adults’ better performance is simply because of their older age and more experience (in which case we may see improvements in the same individuals based on age) or because they have better memory ability.

In the current study, we used the latter approach and compared the spatial cognitive performance of young and adult chickadees. We found that chickadees from high elevation did better at learning the location of the first rewarding feeder, which is in line with our lab study findings. On the other hand, chickadees at high elevations did worse in a reversal learning task when we changed the location of the rewarding feeder.

The key finding here was that worse reversal learning performance at high elevation was driven by the young chickadees. At both elevations, adult survivors showed much better reversal learning performance compared to the first-year young birds. Most importantly, worse performance by young birds was associated with errors made to the previously rewarding feeder. What it means is that young birds kept going to the feeder that had stopped providing food much more frequently compared to the adults, which prevented them to learn the location of the new rewarding feeder as well. At the same time, we did not see differences between young and adult chickadees when they were learning the location of the first feeder.

These results suggest that cognitive flexibility manifested in the ability to learn changing spatial locations may be indeed under natural selection. At the same time, we did not see evidence that survivors have improved their performance compared to the last year and so it does not appear that the differences we detected between adults and young chickadees are based on age and experience. Our continuing research should provide more answers.

These findings are described in the article entitled Memory in wild mountain chickadees from different elevations: comparing first-year birds with older survivors, recently published in the journal Animal Behaviour. This work was conducted by Maria C. Tello-Ramos, Carrie L. Branch, Angela M. Pitera, Dovid Y. Kozlovsky, and Vladimir V. Pravosudov from the University of Nevada of Reno, and Eli S. Bridge from the University of Oklahoma.