Contemporary Animal Cognition
• Test cognition with particular predicted behavioural outcomes
• Animal cognition uses theory and models
• Explain behaviour that cannot be explained through simple learning mechanisms
Potential Lines of Inquiry in Contemporary Animal Cognition
1. How is information stored?
2. How is it retrieved and utilized?
3. How do animals navigate through the world?
4. How do animals judge the passage of time?
5. How are relevant stimuli grouped and sorted into categories?
What is Memory?
• The ability to recall or respond to information presented earlier
• Inferred when current behaviour influenced by previous experience
• What’s the difference between studies of learning and studies of memory? ( you first do learning and then do memory studies)
• Different emphasis on different aspects of the procedures:
1. Acquisition (if interested in learning study)
2. Retention Interval (if interested in memory study)
3. Retrieval (memory)
Comparison of learning and memory experiments:
Different types of memories:
• Working – short-term memory, held long enough to complete the task at hand
• Reference – long-term memories for associations (Honig, 1978)
– Procedural – memory for well-learned tasks and skills (HM: could get better at a task)
– Declarative – memory about the world ( HM: didn’t remember anyone)
• Semantic – store of information necessary for functioning within our culture
• Episodic – memory for particular episodes
Working vs. Reference Memory
• Hunter (1913) tested memory in animals ( Rats, dogs, & raccoons were subjects)
• Procedure: Had to choose one of three doors that was signaled to have food
• Signal was on briefly
• Duration between the signal offset and choice was varied.
• Procedure: Light comes on one of the doorsà stay for (10s, 20s,30s, etc)àwait for sometimeànow pick one of three door where
you saw the lightreinforced if correct.
Hunter (1913) Results
• Main results: Longer the delay, the lower the accuracy
• Species differed in their working memory capacity.
– Rats’maximum delay = 10 sec
– Raccoons’maximum delay = 25 sec
– Dogs’maximum delay > 5 min ( have a huge working memory)
Memory Paradigms I: Delayed Matching to Sample (DMTS)
Training an Animal on DMTS
• Start with B until bird masters task – simultaneous matching to sample – no working memory is needed in this task (because
you are to just look at them and tell which ones look similar)
• Then, after a delay, you move to C – DMTS – requires working memory to remember what the lines looked like in B.
Factors effecting the DMTS:
1. Type of the sample stimulus
– Some stimuli are more effective than others
2. Duration of sample
– Shorter the duration, the lower the accuracy since it is harder to memorize now.
3. Delay after sample before choice – Longer the delay, the lower the accuracy
Classic DMTS Experiment Varying Sample Duration and Delay
Procedure: DMTS (same as above)
• As you increase sample duration (X)àpercent correct increases ( see the numbers
on the sides of each line which refers to how long they saw it)
• As you delay moreà the percent correct decrease (see the overall pattern)
Trace Decay Hypothesis of Memory
• Stimulus presentation creates a change in the nervous system
• These changes decay over time
• Trace related to the characteristics of the stimulus
• Thus, the above data could be explain by the fact that For 1 min= small neural
change vs. the 14 min which produced a lot of neural changes and thus that's why you remember them. But overtime, these
changes do go down (THE RATE IS ALWAYS THE SAME)
Test of the Trace Decay Hypothesis
• 0 seconds normal decay as we say in above
• 6 s no decay until 6 seconds delay in testingthese decay starts.
• Trace decay is not supported
• The previous hypothesis does not seem to work.
• Varied the different delayà there is decay but does depend on for
how long you got the of training.
General vs. Specific Rule Learning in DMTS
• Animals may:
1. Focus on correct 2) inhibit incorrect or 3) use both strategies.
• How do animals learn which stimulus is correct?
• 1) Matching rule choose the stimuli that look same as what was shown.
• 2) Specific stimulus response relationships which include the specific rules or the general rules
• Which rules that the participant use Depends on the size of the stimulus set:
– Small set – specific rules (select red after exposure to red. You do not choose red if green
– Large set – general rules (if you see green just choose green
o trails-unique procedure: different stimuli serve as a sample on different stimuli only
once. Accurate performance only possible if you use general rule.
Memory Paradigms II: Spatial Memory in t he Radial Maze (8 arms with food at each end of arm)
prococedurePut a rat on the maze the rat learns it very quickly and the correct responses increase (i.e. do
not go back to the same area of the maze where you just went)
Results/graph:: with continued practice mean number of correct responses consistently are above 7
( very good). This is because they are evolutionary primed to do this behavior.
How do rats do this?
• Do they use odor cues? NO
• Do they use a fixed response pattern? (i.e. LLRR pattern. Even if you rotate the maze, thus the
pattern if exist could not be used. Found that The rat still got it right NO
• So what do they use to guide their behavior in the radial maze? If you take away the visionàreally
affect it. Thus seems to be the landmarks that are surrounding each of the ends of the arm.
Landmark Use in the Radial Maze
• Rotate landmark group: Rotate the mazeà you rotate (while the external cues are the same) the
maze then you put the rat back again.
• Transpose the landmark: change the environmental cue (i.e. A is no longer at spot a)à really
affected the percent of the response they go right.
• MP: they were using the landmarks to get to the right place. Effect of Delay in the Radial Maze
• Procedure:After making 4 choiceà take them outà delay ranges from 0-24 hr. as shown in graph then put them back into the
maze and observe
• After 4 hr. appears to dropà but we can see that rats still do really well up to 4hr, which is a huge number.As times increase
really low but still better then chance level (which is the dash line)
• MP: working memory exist for up to 4 hrs. and usually is more then >7 words (since the rats had 5-8 choices to make in the
Radial Maze Response strategies
• Diagram=different ways of solving the maze.
• Most of them did B
• They used the central area as the starting pointà this is the way it work in the wild.
Stages of Memory (depends on all three)
Acquisition and Stimulus Coding
• Memories/experiences are not stored as replicas of what actually happened
• Audio tape of birdsong – sounds just like the bird, but the song is coded in the magnetic patterns on the tapes How then are
memories coded in these birds?
Types of Stimulus Coding
• Cognitive maps and navigational codes
– How an animal encodes and represents the world
• Retrospective and prospective coding
– Does an animal remember where it has been (retrospective), or where it has yet to go? (Prospective)
Navigational Coding Strategies
1. Beacon following (following a landmark or using it as signal. i.e. go towards
that big tower)
2. Landmark use: distinction stimuli not at the goal but fixed to the goal
3. Relational landmark use (btw star buck and this turn their)
Coding in a Natural Spatial Task (get to the food below rock)
1. Beacon following: look at the rock and go to it.Another example is that
honeybees fly directly to the food source. The food source acts as a beacon.
2. Landmark use: NE of the bush
3. Relational: it is between the bush and the tree.
Honey bees fly directly to the food. Beacon following
1) pigeon also use landmarks use strategy.
• In a small box they also go the centre
• BUT When you expandà it looks like the they didn’t just went to the center as the small boxes show so they NOT going to the
• Rather the big box shows that they are moving to a place 45 degree from the corner.
• These corners are fixed locations (thus the landmark usages)
Retro- and Prospective Coding
• How are memories for events or places visited encoded?
• Two types of coding:
1. Retrospective (been to)
2. Prospective (will go)
• Animals can use both types of encoding processes
When Might Retro- or Prospective Coding Be Used?
• Prospectiveà initially high memory load at the end very less load. • Retrospectiveà initially no memory but at the end a lot of memory load
• The best thing: Animals should start off using retro ( low memory)à then half way through we should switch to prospective
Flexibility in Coding in a Radial Maze Task
• How much their performance deviate from the control (which is perfect) the above
graph shows us that
• What they did was change the places where they would put.
• Results: Animals do switch the type of memory
Coding Strategies and Task Demands
Mean number of error rates were measured between rats (left) and human ( right) on a spatial
task requiring identification of new place after a delay. The delay was imposedAFTER both
participants had visited different location (thus you use retro right now. You have to remember
where you visited and you memory load should begin to increase)
Graphs and results: both humans had rats same results. Y=mean error rates and X= number of
items visited before the test. Shows that both switch their memories. The third graph.
Retention and Rehearsal
• After acquisition, retention phase
• Potential issue is rehearsal
– Animals keep information “active” in working memory through some form of