Spatial Memory Flashcards
- What is Spatial Cognition?
1. 1 The importance of spatial memory for animals.
Animals learn and remember places:
animals in general need to remember locations for a number of reasons, for example,
A warbler can remember the locations of friends and enemies (Godard 1990). He does an experiment in which he wanted to find out how these birds kept track of the location of enemies and friends. He played a song of any rare bird and tested whether males will approach the speaker. if they do approach it this means the bird is angry and they do not know where this neighbour bird is. in the breeding season, if they played the neighbour bird in the incorrect location, the birds tended to approach the speakers. if they play the neighbour in the correct location, they do not approach the speaker. what is also interesting is that they retested the birds after eight months when they came back from migration. they still remember the location of where they expect to hear the neighbours.
Birds and other animals also remember the locations of food and shelter (e.g. Balda & Kamil, 1992). A Clark nutcracker bird can remember the location of the food that they store in the fall when they go back for it during the winter. nutcrackers are expected to retrieve their food after 11 days all the way to 285. they do much better than what is expected by chance.
So how do animals remember these locations?
Long distance - Navigating across unfamiliar terrain e.g. migration
Short-range - Navigating through the familiar area e.g. during foraging
How do we study it? - Psychological vs. biological methods
- What strategies animals use to navigate long distances?
The astonishing ability of animals from all taxa to find their ways over hundreds or thousands of kilometers is a subject in itself (see Alerstam 2006; Holland, Wikelski, and Wilcove 2006). It has been given a huge boost in recent years by sophisticated satellite tracking systems for recording not only the position but the activities, temperatures, and so forth, of migrating animals. The sensory and neural mechanisms required can also be studied in some of the same species (Frost and Mouritsen 2006). Notwithstanding their vastly different scales, however, long-distance and short-distance travels are largely analyzed with the same basic conceptual framework (Bingman and Cheng 2005). Distance, direction, and position information are important however far one is going, and the degree to which it is maplike is an issue whatever its scale.
Learning & long-distance navigation:
Perdeck (1958) did an experiment in which he showed that some aspects of navigation through migration are learned. while the birds were making their way back from the breeding area to wintering area. he translocated them, so will they continue going in the same direction as they will be going before they were translocated or whether they will compensate and go to the wintering area, which would be where they were expected to go.
moved migrating starlings 1000km off course.
what he found was adults will compensate and fly to the wintering area, whereas the juveniles who were migrating for the first time did not; Adults corrected for the translocation – juveniles didn’t. this suggests that there is an innate component to this navigation skill. but over the course of time, birds learned and can use this information more flexibly.
Thorup et al. (2007) replicated Perdeck’s study across the USA. they translocated what crowned sparrows from Washington state to Princeton (New Jersey) to see whether the birds will go to where they were taken from, if they will go in the same town where they were travelling, or if they keep going in the same direction. so what we see here is that adults compensate, but juveniles keep flying in the same direction (this is true for over 100 kilometers. adults head to the west, while the juveniles keep heading south. Adults “know” where to go!
But how are birds doing this?
Long-distance - Magnetism:
one possibility is that animals are using magnetism. the primary directional cue for nocturnal migrants are the Earth’s magnetic field and, on clear nights, the stars, but the pattern of stars varies with geographic location, time of night, and season, and it changes over geologic time.
Kishkinev et al. (2015) – compared Eurasian reed warblers moved 1000km east to those “virtually translocated” by changing magnetic field to that 1000km east. he applied the same methodology as the previous experiment. he translocated warblers who were flying from Rybachy (R) to an area near Finland. At this spatial location, he moved him to Zvenigorod (Z) to see whether the birds will continue to go in the same direction or whether they will compensate.
he found that the birds who were released in R , they went the normal way, those release in Z compensated.
In his 2nd experiment, he took them to a different location and put a magnetic caite around the closure to replicate the magnetic field at R and Z. he found that birds at virtual R traveled where they would have travelled in real R, and birds in virtual Z travelled where they would have travelled in real Z.
this means that one way in which birds knew where they were, was based on the properties of the magnetic field.
Long-distance - Stars:
Insight into how birds nevertheless use the stars to tell
direction comes from classic experiments by Emlen (1970) with indigo buntings (Passerina cyanea). He raised three groups of birds indoors out of sight of the sky, but late in their first summer two of those groups were exposed to the ‘‘night sky’’ in a planetarium. For one, the stars rotated normally, around the North star, whereas for the other the center of rotation was the bright star Betelgeuse.
When all the birds then spent autumn nights in the planetarium under stationary star patterns typical for the time of year, the birds with no experience of the sky were not well oriented, but those exposed to the normal sky oriented Southward, indicating that they had somehow learned to use the stationary star patterns during earlier exposure to the normal night sky.
The third group treated Betelgeuse as the North star, flying ‘‘south’’ with respect to it, indicating that the star or star pattern near the center of rotation of the night sky is used to give direction. Magnetic information interacts with this information during normal development (see Able and Bingman 1987; Able and Able 1990; Weindler, Wiltschko, and Wiltschko 1996).
Long-distance - Smell:
another possibility is through smell. Papi (1974) did a series of experiments in the 70’s with pigeons in which he combined smell and wind direction in order to test that birds could use this to find their way back home. There were two different treatments, an oil treatment and the dependent turpentine treatment. within these conditions (treatments) birds will have the wind coming from different directions and a combination of odors (oil or turpentine). what they found was that birds in groups A and B learned to associate where they live with wind direction and smell, this was also true with groups C and D in the turpentine treatment.
For flying honeybees, distance is measured by optic flow, the angular motion of images past the eyes. Evidence comes from experiments such as;
bees flew down a tunnel decorated with vertical black and white stripes to find sugar water (Srinivasan et al. 1996). With the food at a fixed location, bees learn where to expect it as evidenced by their circling around over the usual place of food in unrewarded tests. When image motion was eliminated by replacing the vertical stripes with horizontal ones, the bees searched equally at all distances. When the tunnel was wider or narrower than usual, the bees searched at a greater or lesser distances respectively.
To understand why the effect of the tunnel’s width, that is, the distance of images from the eyes, means that angular image motion is important, think of how nearby objects cross your visual field faster than those farther away when you are in a moving car. Changing the
density of the pattern inside the tunnel also changes the rate of image motion, and accordingly, in natural landscapes the bees’ subjective estimates of distance as revealed in their dances is greater when they have flown over a richly patterned landscape than when they have flown the same distance over water (Tautz et al. 2004).
Long-distance - The Sun:
The sun is useless as a landmark because it moves continuously relative to the Earth, but many diurnal animals use it for directional information, that is, they have a sun compass. For example, the desert ants in use both the sun and patterns of polarized light it creates in the sky for directional information when computing their paths home from food (Wehner and Mu¨ller 2006).
Reading direction from the sun regardless of the time of day requires both a stored representation of how the sun moves across the sky at the current location and season (an ephemeris function) and an internal circadian clock. when planning where to go animals need to compensate for the current time of day when using sun compass. they have to take into account if they plan to go in the morning or the afternoon because then the sun would be in different positions and therefore the cues will work in a different manner for the to orientate in a particular location.
this is called time compensation, we know that animals do this because if you change what time of the day they think it is, they adapt to it. To show definitively that an animal is using a sun compass it is necessary to shift its internal clock and test whether orientation shifts accordingly. when this occurs, birds that are clock shifted will think it is the morning. Shifting an animal’s internal clock results in predictable direction errors.
there are experiments that actually tested the use of the sun as a cue. Manx shearwaters forage over thousands of miles.
Padget et. al (2017) clock-shifted birds (4hrs earlier/later) in their burrows and found deflections in their flights home. when the birds were foraging and it was time to go home, the birds who were shifted -4hrs tend to skew one way (north). those shifted +4hrs tended to skew another way (south). birds using time compensate to find their way back home.
- What strategies animals use to navigate short distances?
Learning & short-range navigation
animals have to navigate around familiar areas everyday and we know that this navigation system is learned. B1 and B2 are based on radar recordings and the dark dots represent flowers. what we see is that
bumblebees form repeated foraging routes “traplines” over time. traplines refer to efficient routes between resources that animals use. what is important is that these routs are not totally fixed and animals can flexibly change these routes as they learn about the properties of the flowers for example.
for example, in phase one there route B in particular that a bee will follow when visiting their flowers, so we can follow the direction of the arrows to see that this is the particular route that the bee will follow. however, when the quality of one of the flowers is increased (fo example the nector), the bees are going to prioritise the visit to this flower and change the sequence of the visits accordingly.
the question is how do they do this?
Short-range - Routes:
homing pigeons do something very similar to what the bees do. the experiment involved releasing the pigeons several times from the same location and see how they find their way back home. Homing pigeons in Oxford form individual, repeatable routes when released multiple times from the same location. in all of the cases there are no straight lines to get back home. so critically, if you release the bird the way from the route back home they still don’t go directly home, instead they make their way back to where the original route was and then they follow the route back home. so when released close to route – re-join the learned route rather than fly straight home.
Desert ants also learn individual routes to and from a feeder. there are three representations showing this which show different patterns for each ant. One ant can have multiple specific routes (to get food and to get home)! over the course of the different visits they always seem to go the same way . they have learned series of views at landmarks to find a way to come back from the food and critically sometimes an individual ant can have multiple preferred routes. so what this indicates is that they have multiple memories or different memories for multiple paths. Accordingly they estimate distance using about the only cue available, the number of steps they have taken
The similarities between bees and ants imply that they compute distances using essentially the same implicit countinglike process but on qualitatively different inputs. We know very little about whether and how any mammals, for example nocturnal rodents, sense distance traveled as such. Most laboratory studies of path integration in rats or hamsters test primarily its directional component: in a confined space, animals can choose which way to head but have little choice in how far to go.
How do ants find their way back to the nest?
Short-range: Path integration
this means keeping track of distance and direction – know the vector home. if you know how far to travel in terms of distance and and direction, every step you take you can draw an arrow home. desert ants can do this. Information used differs across species. the distance to home/food in ants is retrieved from the steps they make, in hamsters from vestibular cues.
we know this for ants as when it was time to go home, they headed off in the wrong direction.
A foraging desert ant (Cataglyphis fortis) takes a long and tortuous path for food, but as soon as it finds a prey item it heads straight back to its nest over a hundred meters away. These ants return to the vicinity of the nest using dead reckoning, an internal sense of the direction and distance of the nest from their current position. That they know both distance and direction can be shown by catching an ant in a matchbox just before it starts its homeward journey and releasing it several hundred meters away.
It does not head for the nest but takes a path parallel to that which it would have taken from the point of capture. When it has gone about the right distance, the ant circles around as if looking for the nest in the place where it should be. This behavior shows that the ant must be performing path integration on the outward journey. This behavior increases the chances that the nest is found, which is vital because the hot sand surface can be lethal to ants that do not escape underground quickly enough.
Humans also use pattern integration, but we don’t use steps. we use vestibular information (our ears) to tell us roughly how far we have been travelling. the actual cues that species use differ. dead reckoning in humans was noted by Darwin (1873), its role in spatial learning
described in gerbils (Meriones unguiculatus). they studied mother gerbils and their pups who had a nest at the edge of a large circular arena. If the pups were taken from the nest and placed in a cup somewhere in the arena, the mother soon began to search for them.
The effect of rotation speed (rotating the cup slowly or fastly) reflects the fact that in mammals information about changes in angular orientation is processed by the vestibular system, which senses accelerations and decelerations above a certain threshold (McNaughton, Knierim, and Wilson 1995; Wallace et al. 2002).
Short-range - Landmarks:
When features of a goal are not immediately perceptible from a distance, other objects in fixed locations, that is, landmarks, can guide the animal to it. A classic demonstration of landmark use is Tinbergen’s (1932/1972) study of homing in the digger wasp (Philanthus triangulum). These wasps lay their eggs in a number of burrows, which they provision with bees. Each bee that a wasp collects requires a separate foraging trip, so the female wasp has to learn the location of each of her burrows.
This learning takes place during a brief orientation flight. When leaving the nest for the first time, the wasp turns and faces the nest entrance and flies around in
ever-increasing loops, apparently inspecting the entrance and the objects around it. If the objects surrounding an established nest are altered while the wasp is inside, a new orientation flight will be elicited the next time she departs (T. Collett and Lehrer 1993; Lehrer 1993).
landmarks are not simple, we can either have ‘global’ or ‘distal’ cues which are far away from us or ‘local’ or ‘proximal’ cues which are closer to us to find our way around.
an experimental done with Columbian Ground squirrels found in North America, they tested if they could find a way to escape burrows. the experimenters presented a predators call from an eagle and what the squirrels tend to do when they hear these calls is to escape and hide in the burrows. the experimenters tested whether the quireels could use global cues or local cues. they placed the squirrels into an arena, and in some cases the walls were low.
there was a tarp covering some of the cues meant that they could not use local landmarks. in other cases, the walls were very high meaning that they could not use the global landmarks. sometimes the wall was kind of medium high and the squirrels could use some of the global landmarks. finally, there was a condition in which the squirrels could not use any landmark. what the experimenters wanted to see is how long it will take the squirrels to find escape. what they found was the ground squirrels take the longest when they blocked out the global cues. however, other species might benefit from using local cues.
so how do animal species benefit just from using local cues?
- animals can use landmarks such as beacons - head towards landmark
- the landmark such as Vectors - to search for the distance and direction from each landmark (gerbils use this)
- Relative (environmental) geometry - an abstract (for example, searching in between two landmarks) relationship between landmarks. Young chickens, pigeons, black-capped chickadees, two species of fish, and monkeys encode the locations of goals relative to the geometry of an enclosure, even in the presence of features like corner panels or a colored wall that disambiguate the geometry (Cheng and Newcombe 2005).
- How do we study spatial memory in animals.
Short-range: traditional approaches
- Ethology/Behavioural Ecology (a naturalistic approach to the study of animal behaviour, so studying animals in a natural context)
- Comparative psychology & neuroscience (understanding cognitive mechanisms underlying the particular behaviour, for example spatial cognition in animals)
———————————————————————————
Ethology/Behavioural Ecology
Tinbergen - classical experiment on spatial cognition or navigation of wasps. he placed a ring of pine cones around the wasps nest. and when the wasp left the nest, Tinbergen moved the pine cones to see if the wasp could use the pine cones as landmarks. he found that they did even though he hadn’t moved the pine cones, they will still fly into the ring to find the nest.
Comparative psychology & neuroscience
Tolmone - came up with the concept of cognitive maps. he developed this type of maze where the rats were trained to run in. controlled experiments in the lab.
experimenters tend to test spatial memory by using well-defined paradigms, particular set of tasks, that can be manipulated pretty precisely to figure out what cognitive mechanisms animals are using to find food.
T-maze = animals in this type of maze are trained to find food in another part of the maze. what tends to happen is that they play in a different location in the maze. two possible responses; at the top of the maze they go to the same location they were trained to find the food (place response) or they go down the maze and turn the same way they turned during the training (response response).
Radial maze - food can be found in some of the arms, test animals are asked to remember in which location.
Water maze - rat/rodent put into a type of swimming pool with a pake milky water. has to find the location of a platform which is just above the water. clear learning performance to find the platform.
however for a long time, there was little interaction between these two disciplines, even though they were interested in similar phenomenon. they took different approaches to study this phenomenon.
What processes are involved in spatial memory?
How do animals compare with humans?
in comparative psychology it is important to identify whether there are any cognitive skills that are uniquely human. a study compares pigeons and humans in the touch screen task. both of them were trained to peck/press a correct location to get a reward. they were trained/presented with this situation. there are a few landmarks involved (star, rectangle etc.), what they have to do is find the location contained by a square (square isn’t presented initially). if they peck/press outwith the square area they will get rewarded.
in a different test the landmarks are moved apart, they are still supposed to search within the square.
The differences between the two different approaches:
Comparative psychology & neuroscience
- the number of species used to test this ability
- interested in the mechanisms behind a particular cognitive skill
- interested in what animals can do when facing a new situation
- the kind of cognitive representations animals can make
- interested in how humans compare with them (with other animals), are these uniquely human.
Ethology/Behavioural Ecology
- interested in what animals are learning
- interest in how they use the information
- and how this is shaped by evolution and ecology
——————————————————————————-
What can psychology and biology teach us about spatial memory?
•Spatial learning likely associative.
•Animals can learn abstract relationships.
•Involves processes like working memory and attention.
•Learning important across a range of scales.
•Important role for familiar routes and landmarks.
•Animals can use wide range of cues.
Today there is much more overlap e.g. psychological tools to ask ecological questions
another example
long distance - beacons
Beacons are sometimes referred to in the psychological literature as proximal cues, that is, cues close to the goal, as distinct from distal cues, the landmarks to be discussed in the next section. (Local vs. global cues is much the same distinction.) Often animals can use either proximal or distal cues, depending on which are available. A now-classic demonstration was devised by Morris (1981). A rat is placed in a circular pool of water in which it swims until it finds a small dry platform, a plexiglas cylinder standing somewhere in the pool.
For some rats, the cylinder is black and visible above the water. Thus the platform can function as a beacon, and because rats would rather be dry than swim, they soon learn to approach it wherever it is in the pool. For other rats, the water is made opaque by the addition of milk, and the platform is transparent and slightly below the water surface. These rats must use distal cues, objects in the room surrounding the pool, to find the
platform, and they also quickly learn to approach it, provided it stays in the same place from trial to trial.
When the platform is removed on test trials, these rats still head directly to the correct location and swim around it as if searching for the platform. This behavior has typically been taken as evidence for learning the specific place where the platform is, but it may often reflect instead learning what direction to head relative to distal cues (Hamilton et al. 2008). Information from beacons is not inherently spatial because it is not vector information but rather information about value.
8.3 Acquiring spatial knowledge: The conditions for learning
- exploration
The tendency to explore novel objects and environments is one of the best examples of spatial behaviors that expose animals to the conditions for
learning. The rat sniffing a novel object, the young pigeon flying in circles over its loft, or the bee performing an orientation flight (Wei, Rafalko, and Dyer 2002) are actively exposing themselves to objects and spatial relationships that they need to learn about. - Learning about redundant cues: Competition or parallel processing?
O’Keefe and Nadel (1978) suggested that exploring novel items in a familiar space allows an animal to update its cognitive map in the same way as a cartographer adds a new farmhouse or removes a hedge from a printed map. Incorporating all available
cues into a cognitive map would ensure redundancy when primary cues fail, which could be important for tasks like getting home. Indeed, an example of backup
mechanisms is illustrated in Figure 8.13. As another example, experienced homing pigeons tested on sunny days use a sun compass, but birds tested under thick cloud cover can home just as well, relying on landmark memory, olfaction, magnetic information and/or infrasound (Keeton 1974).
Do animals have a cognitive map?
what is a cognitive map?
the representation embodied in a cognitive map is typically assumed to encode distances and directions and to enable mental operations on them.
Distances and directions are the metric properties of space. Blueprints, city plans, road maps, and globes are useful because they represent distances and directions
accurately. But plenty of useful maps do not preserve such vector information.
network map - for example, a subway route map is useful for planning a trip on the subway because it shows which station is on which route and what order they can be reached in. this kind of map can be used without its representing distances between stations or angles between connecting routes. Indeed, because these may not be represented accurately, a tourist wanting to explore the city on foot would be foolish to use it as a guide.
a vector map - a map that preserves distance and direction information that allows the planning of novel routes to unseen goals. How useful it is depends on the density of identifiable locations represented. For example, a tourist starting from an obscure side street
armed only with a vector map of the city landmarks has to wander around until finding a place marked on the map (a potential limitation).
a cognitive map - in the sense of a global representation of space equivalent to an overhead view that preserves distances and directions among an infinity of locations. Whether cognitive map always means the same thing is a problem too, as we see by surveying some of the landmarks in its history.
Do people have cognitive maps?
Research on spatial cognition in human adults and children is a large area in its own right and can be given only a brief mention here (for an introduction see Newcombe and Huttenlocher 2000). As indicated by the scattered mentions of findings with people, much contemporary work in this area is closely integrated with that on other species, especially in looking at spatial behavior in terms of a number of distinct
subprocesses and in failing to find evidence for overall cognitive maps.
Nowhere is this more evident than in a prominent opinion piece titled ‘‘Human spatial representation: Insights from animals’’ (Wang and Spelke 2002). Wang and Spelke proposed that rather than depending on an enduring allocentric map, much human spatial
behavior depends on momentary egocentric representations, specifically dead reckoning, orienting by the geometry of surrounding space, and viewpoint-dependent matching of remembered to current views of the environment.
Evidence for each of these processes comes from animal data like that reviewed throughout this chapter
and from analogous experiments with people. In one key example, people viewed a room with a few objects in it and were then blindfolded, disoriented, and asked to point to the objects and the corners of the room. Errors in pointing indicated that the objects had not been integrated either into a map of the room as a whole or into a single configuration (Wang and Spelke 2000).
Evidence that recognition of a familiar scene takes longer from a novel viewpoint supports the suggestion that encoding is viewpoint-dependent. However, more recent research (Burgess 2006) indicates that
human spatial representation has both egocentric (self-centred) and allocentric (having one’s interest and attention centered on other persons) components, which exist in parallel. In experiments like those just summarized, greater experience, a larger environment, and other factors make allocentric representations more evident.
This approach is clearly much in the spirit of other research emphasized in the present chapter in dissecting spatial cognition into distinct parallel but interacting mechanisms and eschewing discussion of overall maps. Whether two systems defined in terms
of function, egocentric and allocentric, will provide a useful way forward remains to be seen.
conclusion
The study of spatial orientation is a very active area using a wide variety of species and approaches from fieldwork to neuroscience (Box 8.4). Among areas of research in comparative cognition it is exemplary, perhaps unique, in the way in which data and
theorizing have been integrated across species and approaches as for example in the book edited by Jeffery (2003). The richest bodies of data come from three very different groups of animals: small nocturnal rodents (rats and hamsters), diurnal, central-place foraging insects (bees, wasps, and ants), and birds that orient over tens to hundreds of kilometers (homing pigeons and migratory species).
The ways in which these animals perceive the world (consider for instance the very different visual
systems of rats, pigeons, and bees) and the cues relevant for orientation in their natural environments differ enormously, yet some orientation mechanisms such as landmark learning or path integration and their interactions have been analyzed in a way that cuts across phyla. To some extent, this integrative approach has resulted in a theoretical orientation based on ideas from human psychology being replaced by one
rooted in data from nonhuman animals.
This chapter began with descriptions of the wide range of mechanisms animals use for getting around. By itself, each of them has advantages and disadvantages. Dead
reckoning is most useful for short journeys back and forth to a central place, especially in an environment with relatively few landmarks, as in the dark or on the
desert. Other ways of getting back and forth to a starting place include route learning both in the sense of a memorized sequence of motor patterns (response learning) and in the sense of a sequence of responses to landmarks.
Dead reckoning (process of calculating current position of some moving object by using a previously determined position, or fix, and then incorporating estimations of speed, heading direction, and course over elapsed time.) and route learning in either sense leave the animal lost if it is displaced too far off its usual route. However, stimulus generalization between familiar and unfamiliar views of the environment gives route learning some flexibility. The varieties of spatial information—from landmarks, beacons, dead reckoning, environmental shape—are processed in different cognitive modules which take different kinds of input and output decisions about what distance and/or direction to move relative to different kinds of cues.
This raises the question of how the outputs of
different spatial modules are combined during the acquisition and use of spatial information. Are different kinds of information processed in parallel, do they compete for learning as in conditioning, or are they integrated in some other way? When are modules used in a hierarchical manner, and why? When spatial cues have acquired their significance, do they compete for control or are their outputs averaged? When does each kind of combination rule operate?
For instance, does the system that has been more reliable during evolution or individual experience or that evolved earlier take precedence? A great deal of attention has been devoted to the question of whether any animal integrates different sources of spatial information into a unified allocentric representation of distances and directions, a cognitive map. This question turns out to be difficult to answer, partly because cognitive maps can mean different
things to different people. Focusing on the specific cues available to animals and how they are used in specific situations provides better understanding of how animals get around than attempting to prove or disprove use of a cognitive map.
