PSYC 358: Comparative Cognition (Exam 1)

Morgan's Canon

Do not attribute a behaviour to higher-order cognition if it is sufficiently explained by lower-order cognition

Tinbergen's Four Questions

Proximate questions: causation, development Ultimate questions: function, evolution

Modularity/modular cognition

How well animals complete particular tasks in particular realms (e.g. numerosity, caching)

Examples of modular cognition: Caching/Colour

• Nutcrackers have excellent spatial ability and can cache huge amounts of food and find it again; Mexican jays are very bad at caching

• Mexican jays are very good at delayed match-to-sample tasks involving colour choice; Nutcrackers are very bad at these tasks

Adaptive specialization/domain specificity

Refers to the fact that some associations are learned more readily than others due to evolutionary history (e.g. Garcia effect)

Sensory bias

When a preexisting sensory bias becomes used for something other than its original purpose (e.g. mate selection); aka arbitrary bias or runaway selection

Example of sensory bias: Tungara frogs

Male tungara frogs make a "whine and chuck" sound; each component of the sound stimulates one of the regions of the female frog's ear, stimulating the female's ear to the maximum and causing her to approach. These areas evolved for predator detection and foraging, and so these represents sensory bias as the males have coopted those regions in the ear

Three methods of studying perception in animals:

1. Physiology/neuroscience

2. Changes in natural behaviour

3. Studying learned behaviour through psychophysics tasks

Stimulus filtering

Ability to filter out unwanted/unimportant information

Forms of recognition

1. Habituation
2. Perceptual learning
3. Imprinting

Example of perceptual learning: rats

• Experimental group were in a cage with circles and triangles hung on the wall, control group were in a cage with no hangings
• Both groups asked to discriminate between triangles and circles
• Experimental group improved more rapidly

Serial reversal

Training procedure in which one stimulus is initially rewarded for and one is not, and then the reward contingency switches

Learning set

Procedure used by Harlow involving multiple blocks, each containing 6-8 trials, with the reward contingency shifting each trial; the 2nd trial is measured to look for the optimal strategy of win-stay, lose-shift

Types of categories

Perceptual, functional, relational

Perceptual categories

Categorization based on similar appearance

Functional categories

Categorization based on shared meaning; difficult to study in animals as we don't know what functional categories they have

Functional equivalence

Ability to understand different forms of an object that mean the same thing

Stimulus equivalence

The ability to respond to distinct stimuli as if they were the same due to prior associations between those stimuli and a common outcome

Relational categories

Categorization based on relationships between items in a category not based in perceptual features of function

Example of relational categories: chimps

• Chimps and researchers were on opposite sides of a table with cups that may have food under them

• Trial 1: physical match

• Trial 2: relational match (e.g. smallest cup)

• Trial 3: relational match with distractor

• Trial 4: relational match with distractor and novel items

• Trial 5: relational match with distractor and dissimilar transfer

• Chimps were 75-80% accurate

Working memory

Memory for an event on a specific trial

Reference memory

Memory for the general principles and skills of a task, not about any particular event/trial

Example of habituation: rats

• Rats have time to explore two different rooms (contexts) with two junk (meaningless) objects

• Rats are put in a test room that matches one of the two rooms, but with one novel object for that context

• If rats spend more time exploring the novel object, that demonstrates habituation/memory

• As delay increases, time spent exploring the two objects evens out

Radial arm maze

A central location with (often) 8 arms, used to test memory in rats through the following procedure:

• The rat is allowed to freely explore and get to know the maze
• The rat is placed into the maze with 4 arms open and baited, allowed to get the 4 rewards, and removed
• The rat is placed back into the maze with all arms open, and accuracy of going to the remaining 4 doors is measured

Advantages of the radial arm maze

• Task is more natural and thus has richer memory cues

• Movement is a natural response

• Avoids recency bias/confusion

Example of chunking: rats

• In a 12-armed radial arm maze with 4 items each of 3 different food types, rats are better at finding the items if the initial 4 doors opened are all the same type

• Rats find the food in order of the food type (i.e. all 4 of one, then all 4 of another)

Episodic-like memory

Memory in animals for the "what, when, and where" of events in the animal's life

Example of timing and memory: hummingbirds

• 8-armed radial arm maze with 4 arms baited every 10 minutes, and 4 arms baited every 20 minutes
• Hummingbirds always returned to the 10-minute arms roughly every 10 minutes and the 20-minute arms roughly every 20 minutes, demonstrating memory for when they were fed where
• Hummingbirds arrived at roughly 10 and 20 minute intervals, showing a tense of interval timing

Example of proactive interference: birds

• Birds studied a "list" of locations (open holes in the floor with bait)
• Test phase involved clusters of open holes, with one hole in the cluster being the baited hole from List 1
• Birds studied a second list
• Test phase was the same, with some holes open from List 1 to act as interference
• Accuracy decreased from List 1 recognition to List 2 recognition

Example of metamemory: monkey

• Monkeys were either forced to complete a match-to-sample task (33% of the time), or given a choice whether to complete the task or forfeit and gain a less-values reward
• For one monkey in the study, as the delay increased and his accuracy on forced tasks decreased, he declined the task more and more, demonstrating metamemory

Getting there vs knowing where

Getting there: ability to move in the environment Knowing where: ability to know locations in the environment

Dispersal

Movement of animals away from their place of origin to new locations

Landmarks

Stationary objects that serve as distal cues for a goal location

Beacons

The goal location; can move in the environment

Example of landmarks: beewolf wasps

• Pinecones were set up around the mound (home)
• When wasps left, they moved the pincones
• Wasps returned to the pinecones, rather than the mount

Example of template matching landmarks: bees

• Changing the size of the landmark affected the bees' estimations of distance
• If the size of the landmark increased, bees didn't travel as far
• If the size of the landmark decreased, bees travelled further

Example of cue redundancy: rats (Morris water maze)

• One group learned the maze with landmarks and beacons, one group initially had landmarks blocked with a curtain (and then had the curtain removed)

• The group that had landmarks blocked initially could not use landmarks to navigate in testing, as they had only learned to use the beacon

Example of exploration and navigation: rats (Maier 3 table task)

• Three tables with a y-shaped apparatus connecting them

• Rats are allowed to freely explore 1, 2, or all 3 tables in learning phase

• In testing, rats are placed on 2 baited tables, allowed to eat some (not all) of the food, and then placed on the third empty table; researchers measure the accuracy of returning to the two baited tables

• Performance is poor when rats are only allowed to explore 1 table, it improves when rats are allowed to explore 2 tables, and it improves further when rats are allowed to explore all 3 tables

• Conclusion: exploration aids in navigation

Example of shortcutting: rats

• Central box with three goal boxes with paths between them that are closed in testing
• Middle goal box has chocolate cereal (preferred reward), one of the outer two has cereal
• In learning phase, the paths between goal boxes are closed
• In test, these paths open to see if rats take the shortcut (evidence of cognitive maps)
• Most rats take the shortcut

Numerosity discrimination

The ability to discriminate between different quantities, which may function based on a magnitude system or an object tracking system

Object tracking system

Form of numerosity involved in small number discrimination that involves directly tracking each object

Example of numerosity discrimination: lions

• Researchers played a varying number of roars to groups of lions that varied in size

• Lions were more likely to approach if their group outnumbered the number of roars played (showing an understanding of difference in quantity)

Example of magnitude (timing/quantity): rats

• Rats trained to press the right lever for 2 pulses lasting 2 seconds, or the left lever for 8 pulses lasting 8 seconds
• In the quantity test, the pulses lasted a total of 4 seconds, but those 4 seconds contained either 2 (right lever) or 8 (left lever) pulses
• In the time test, there were 4 pulses, but those pulses lasted either 2 (right lever) or 8 (left lever) seconds
• Regardless of whether time or quantity was tested, rats pulled the right lever for a magnitude of 2 and the left lever for a magnitude of 8

Example of relative numerosity: pigeons

• Pigeons are trained to peck at arrays of Os, and not peck at arrays of Xs
• Pigeons are given arrays of Os and Xs mixed together
• The proportion of pecks pigeons make matches the proportion of of Os to Xs

Example of spontaneous numerosity: monkeys

• Monkeys were tested with a match-to-sample task, where the possible items matched the sample with either number (correct choice) or a distractor similarity (e.g. shape, colour, surface area)

• Monkeys were tested with no prior training (i.e. they were not trained to choose based on number)

• Monkeys selected the answer based on number, demonstrating spontaneous numerosity

Example of object tracking system: monkeys

• Monkeys watched researchers add apples to two boxes, then got to choose a box

• Monkeys were good at this until the number of items exceeded about 9

• Task performance depended more on number of items than ratio (e.g. monkeys could discriminate between 1 and 2, but not 4 and 8), suggesting they were individually tracking each object

Example of numerical operations: chimps

• Chimps are familiarized with boxes of 2 or 3 lemons; in testing, they spend equal time staring at both boxes

• Chimps watch a box start with 1 lemon, a screen covers the box while 1 lemon is added, and the screen comes up to reveal 2 or 3 lemons

• In testing, chimps spend more time staring at the 3 lemons than 2 lemons, demonstrating understanding of an impossible event

Example of ordinal comparison: rhesus monkeys

• Rhesus monkeys were taught to order quantities of 1-4

• Rhesus monkeys were tested on quantities of 5-9 without prior training

• Monkeys could successfully order quantities of 5-9, suggesting they understand the order of magnitude and are not just memorizing a sequence

Acquisition

Learning to engage in simple discriminations

Three mechanisms underlying discrimination:

1. Simple stimulus response

2. Stimulus control

3. Concepts

Untrained responses

Built-in discriminations that we do not have to learn (e.g. Tinbergen’s sticklebacks)

Example of stimulus equivalence in animals: rats

• Step 1: train responses to 4 stimuli, associating pairs with a common response (e.g. if a tone or flashing light is presented, press a lever; if a click or solid light is presented, do a nose poke)

• Step 2: reassign ONE stimulus from each response pair to a new response (e.g. if a tone is played, run on the wheel; if a click is played, pull a chain)

• Step 3: test if the other stimulus from the response pair elicits the new response (e.g. if the flashing light is shown, does the rat run on the wheel; if the solid light is shown, does the rat pull the chain)

• If the rat has stimulus equivalence, it should perform the new response in step 3

Testing abstract concept of sameness: pigeons, capuchins, rhesus monkeys

• Animals were given two abstract photos and had to indicate if it was the same or different (using discrimination + serial reversal methods)

• In the baseline condition, the same photos were used repeatedly; in the transfer condition, new photos were used

• All three species were successful; performance was higher in the baseline condition but improved with the more sets that were shown in the transfer condition

Three methods of concept formation:

1. Elemental

2. Exemplar

3. Prototype

Elemental concept formation

Forming concepts based on basic (perceptual) features

Exemplar concept formation

Storing representations of items in a group, and comparing to-be-categorized items in the environment to those representations; requires storing a lot of information

Prototype concept formation

Taking members of a concept and creating an average, then comparing items to that average (prototype) and categorizing them based on how well they match

Example of non-declarative reference memory: cyclid fish

• Intruders entered the tank of cyclid fish, which are territorial

• Fish experienced either the same or different intruders, entering from either the same or different locations

• Location did not matter for displays of aggression; displays of aggression for lowered for the same intruder

• Shows 1) memory for that intruder and 2) that the memory is related to the individual fish, not the location in which they appear