lecture 22: animal cognition i (memory) learning, psychology 5310 spring, 2015 professor delamater

Lecture 22: Animal Cognition I (Memory)

Learning, Psychology 5310

Spring, 2015

Professor Delamater

Historical Issues of Comparative Cognition

1. Grew out of evolutionary theory and search for evolution of mind2. However, strongly experimentally focused. It developed as a reaction against

anthropomorphic approaches that emphasized consciousness and intentionality3. Also, it has been influenced by general-process approaches to the study of learning

where in addition to searching for general cognitive processes, there isrecognition of the importance of species-specific processes.

Learning versus Memory

1. These are really intertwined processes. The focus of learning and memoryresearchers is usually on different aspects of a learning experiment.

• Whereas learning experiments focus on acquisition processes, memory experimentsFocus on retention and retrieval processes.

Types of Memory

1. Procedural vs declarative (episodic) – learning to do something (S-R association?)vs learning that such and such will occur (S-S association?). Episodic memoryis usually thought of as a memory that binds together what, when, and whereinformation.

2. Working vs reference – Working memory is retention of information long enoughfor use in a particular task (keeping track of a phone number, etc).Reference has to do with long-term storage of acquired information.

• How do we study these different types of memory?1. S-S vs S-R studies in Pavlovian and instrumental learning are instructive…2. Lots of different paradigms employed to study working/reference memory…

Working & Reference Memory Paradigms

1. Delayed Matching to Sample task (DMTS)2. Spatial Memory tasks

Morris Water Maze, Radial Arm Maze

• Three popular paradigms that have been used with lots of different animal species.

Working & Reference Memory Paradigms1. Delayed Matching to Sample task (DMTS)

Working and Reference Memory

• Birds trained with a 0-s delay show decreased accuracy when tested with longer delays.• But notice that birds trained with longer delays show less of an effect of delay in testing.• This suggests that the normal delay gradient does not simply reflect a trace decay process• Working memory appears to be more active than that.• Reference memory in this task refers to animals learning to always choose the matching

stimulus

Sargisson & White (2001) study

Working & Reference Memory Paradigms1. Delayed Matching to Sample task (DMTS)

Do animals learn a General Rule or Specific S-R associations?

• This question can be studied by varying the number ofstimulus sets one uses to train the animals.

• If only one set is used (Red, Green), then animals seem tolearn specific rules. This is known because when theanimals are transferred to a new stimulus set, they donot spontaneously match to sample. They have to learnto do that. Had they acquired a general “matching” rulethen they should have spontaneously transferred this ruleto new stimulus sets.

• However, if you train the animal with more stimulus setsthen their transfer increases, suggesting that they may beusing a general matching rule under those circumstances.

• The trials-unique procedure is another one that can be used to assess learning ofgeneral rules.

Working & Reference Memory Paradigms2. Spatial Memory – Morris Water maze

• Animals learn to find the hidden platform.• They generally do so by relying on external

spatial cues (window, door, etc).

Working & Reference Memory Paradigms2. Spatial Memory – Morris Water maze

• Animals learn to find the hidden platform.• They generally do so by relying on external

spatial cues (window, door, etc).• However, if a beacon is placed above the platform,

then they learn to approach those cues and someresearch suggests that this interferes with themlearning about other spatial cues.

• They may also learn about the general geometry ofthe pool (e.g., circular, or rectangular, etc).

• Thus, they may “code” the stimuli in various ways:specific stimuli, spatial landmarks, geometry relations

• Some have suggested that the animals develop aninternal representation of the spatial arrangement ofthe pool, a “cognitive map,” and base their swimmingbehavior on that.

Working & Reference Memory Paradigms2. Spatial Memory – Radial Arm Maze

• Animals learn retrieve all food rewards at the ends ofeach arm without revisiting arms.

• This task obviously involves “working memory” sincethe animal must keep track of where its been orwhere it needs to go to find food within each trial.




• The effects of memory delays can be studied:4 Arms – Delay – 4 Arms Procedure

Delay = 1 or 25 hoursTrain with either 24 or 48 hour ITI

Crystal and Babb (2008) study

• Memory declines with the delay (as expected), butalso memory is slightly better with longer ITI in training.




• The effects of memory delays can be studied:4 Arms – Delay – 4 Arms Procedure

Delay = 1 or 25 hoursTrain with either 24 or 48 hour ITI

• Memory declines with the delay (as expected), butalso memory is slightly better with longer ITI in training.

• In addition, Reference Memory can also be studied:4/4 Baited-Unbaited task: In this case, the same 4 arms

are always baited with food and the other 4 never baited.The rats learn which ones are baited (reference memory) and they also learn not to

revisit arms they’ve previously visited (working memory).

Memory Processes: Stimulus Coding1. Stimulus Coding

a. spatial codes (beacons – landmark-goal relations, geometry)b. Prospective vs Retrospective codes

Chickadee study: Gp Expt – Eat Sunflower Seeds, 30 min later mealwormsGp Control – Eat Sunflower Seeds (no mealworms)

• The birds consume fewer sunflower seeds,presumably because they expect to get the morepreferred mealworms later (prospective code).



12-arm Radial maze study (Cook, Brown, & Riley, 1985)

Prospective vs Retrospective CodesProspective code (memory for where you have NOT been)

Errors should be greater early, and less later in the task.Retrospective code (memory for where you HAVE been)

Errors should be less early, and greater late in the task.

• Different groups were allowed to retrieve food from 2, 4, 6, 8, or 10 arms beforebeing taken out of the maze for a 15-min delay interval. After this, they wereplaced back in the maze and allowed to retrieve the remaining food items.The number of errors (revisits) made was assessed in all the groups.



12-arm Radial maze study (Cook, Brown, & Riley, 1985)

Prospective vs Retrospective CodesProspective code (memory for where you have NOT been)

Errors should be greater early, and less later in the task.Retrospective code (memory for where you HAVE been)

Errors should be less early, and greater late in the task.

• Errors increased with number of choices,but then decreased thereafter.• This suggests that rats use a retrospectivecode early and then switch to aprospective code later in the task.

Memory Processes: Retention1. Stimulus Coding2. Retention (Rehearsal) Processes

Directed Forgetting/Remembering Task

• On some trials a cue occurs which signalsthat there will be no comparisonchoice.

• These are called “forget cue” trials.• At issue is whether on a probe choice

that occurs on such a trial, performacewill be low. If it is, this suggests thatthere are active memory processesoccurring during the retention interval.

Memory Processes: Retention1. Stimulus Coding2. Retention (Rehearsal) Processes

Directed Forgetting/Remembering Task (DMTS)

• Performance on Forget Cue probe trials is worse than on Remember Cue trials• This suggests that memory is not merely a passive process, but is active

Memory Processes: Retrieval1. Stimulus Coding2. Retention (Rehearsal) Processes3. Retrieval Processes

Reminder (i.e., retrieval) cues are thought to reactivate items into working memory.This can make them more available for use in a task.

• Baby can move the mobile by kicking. They learn to increase their kicking responses

• This is an example of simple instrumental conditioning in the human baby.

Rovee-Collier Mobile Task





• This study shows that 6-month old babiesremember better when they are tested in thecrib with the Same liner as was used during training.

• This shows that retrieval cues were effective.






• This study shows that 6-month old babiesremember better when they are tested in thecrib with the Same liner as was used during training.

• This shows that retrieval cues were effective.• In this study, babies normally forget when tested

one week after training. But if they are exposedto the retrieval cue one day before testing(Same odor condition) they remember better.


Memory Processes: Forgetting1. Is the memory lost or is it non-retrievable?

• Rats learned a one-way shuttle avoidance task• Placed in the white compartment and shocked

once entering the black compartment.• Latency to reenter the black compartment is

the measure of learning.• This latency is long in conditioned animals, but

it is shorter in animals given extinction.• However, a hypothermia treatment administered

immediately or 30 min later, but not 60 min,disrupted memory of extinction.

• Hypothermia caused amnesia for extinction.

Briggs and Riccio (2007) study

Memory Processes: Forgetting1. Is the memory lost or is it non-retrievable?

• Rats learned a one-way shuttle avoidance task• Placed in the white compartment and shocked

once entering the black compartment.• Latency to reenter the black compartment is

the measure of learning.• This latency is long in conditioned animals, but

it is shorter in animals given extinction.• However, a hypothermia treatment administered

immediately or 30 min later, but not 60 min,disrupted memory of extinction.

• Hypothermia caused amnesia for extinction.• But, reexposure to the hypothermia treatment before testing brought back the extinction

memory, suggesting that the amnesia was a retrieval deficit and NOT a permanentdisruption of the consolidated memory.

Briggs and Riccio (2007) study

Memory Processes: Consolidation, Reconsolidation, Memory Updating

1. Traditional View:Rehearsal within STM leads to consolidation in LTM

2. Modern View:Memories are either active or inactiveActive memories can be consolidated into an

inactive state.Previously stored (inactive) memories can be

made active again through a retrievalprocess.

But this places the memory in a vulnerableactive state where it can be altered.

Memory Processes: Consolidation, Reconsolidation, Memory Updating

• Rats underwent Pavlovian fear conditioning (Tone – Shock Pairings)

• Then, one group was exposed to Tone alone,but this was followed by an infusion ofa protein synthesis inhibitor into theamygdala.

• A control group received infusion of a controlsolution.

• Both groups were then tested the next day forfear to the Tone CS.

• The protein synthesis inhibitor group showedvery little fear CRs in this test.

• The result means that reactivating the Tone-shock memory induced

reconsolidation ofthe memory, but this also placed it in avulnerable state of rewriting.

• The protein synthesis manipulation may have updated the memory by wiping it out.

Nader, Schafe, and LeDoux (2000) study

lecture 22: animal cognition i (memory) learning, psychology 5310 spring, 2015 professor delamater

Documents

memory learning

memory experiments

episodic memory

memory researchers

vs reference working

reference memory birds

workingreference memory

different types of memory