Spatial Learning Flashcards
Place cells
Cell fires when individual is in a specific location within the environment.
Cell firing is highest when the animal is within the place field.
Place cells next to each other do not necessarily have adjacent receptive fields.
Supports the cognitive map hypothesis.
Grid cells
Cells in the EC, with receptive fields arranged in hexagons formed of equilateral triangles. The cell will fire when the individual is located within the receptive field.
This means that the cell will become repeatedly activated and then deactivated when the organism moves across the environment.
Allows interpretation of distance travelled.
Supports path integration hypothesis.
Tolman
Supervised and unsupervised learning in rats. knew that animals move towards objects, e.g. food.
Placed rats in mazes with either food (sup.) or no food (unsup.). Rats not given the food were able to find the food faster when presented on day 11 > food given all the time.
Unsupervised learning - learned bout the environment without feedback, so when given the food, showed insight to move towards the food using the most efficient route available. (not simply stimulus-response behaviour).
Olton
Radial arm maze (rats). Rats learn which of the arms are baited. Use reference points in environment to learn locations. They then must use this insight to make judgement of where to move to (only towards food - most efficient strategy).
Measure speed, direction, behaviour etc.
Reference memory
Olton
Used between trials. e.g. remember which arm(s) are baited.
Working memory
Olton
Used within trials, e.g. remember which arm(s) have already visited.
Egocentric cues
Ideolithic cues that are internally generated.
e.g. vestibular input - head direction cells.
Only available in environment that is known well.
Can accumulate errors through path integration.
Allocentric landmarks
External landmarks within the environment. Can be used to form cognitive map within area known well.
Supervised learning
Monitored by feedback system.
e.g. animal learns that the CS predicts the US. Therefore will do CR in anticipation.
Supervised by a monitoring system - calculated deviation from the expected outcome. (Expected-Actual).
In response to deviation, synaptic weights are altered to reach expected outcome.
Unsupervised learning
No feedback provided.
e.g. spatial learning.
no immediate reward/reinforcement for learning.
MWM
Morris Water Maze (Richard Morris).
Cued learning in MWM
Does not require navigation.
Can be performed with hippocampus lesions.
Involves stimulus-response pairing.
Spatial learning in MWM
Does require navigation.
Rat needs to learn the relationship between the external cues provided - form cognitive map of where the platform is located.
Animals keep track of location integrating ____ and ____ motion.
Linear, Angular.
Cognitive Map hypothesis
O’keefe and Nadel.
Generate an internal representation of the environment based on the external cues provided.
Located within the hippocampus.
Place cells provide good neurological basis of cognitive map.
Does not require internally generated egocentric cues (e.g. vestibular).