Forgetting & Remembering

Question 1

DELAYED MATCHING TO SAMPLE (DMTS)

Answer 1

• used w/pigeons/monkeys/dolphins
• modern Skinner-box technique for studying delayed reaction
• doesn’t necessarily require recognition of sample identity/comparison stimuli
• many variations:
  ODDITY FROM SAMPLE
  SYMBOLIC MATCHING TO SAMPLE
  MULTIPLE SAMPLES

Question 2

ODDITY FROM SAMPLE

Answer 2

• choose the comparison that doesn’t match

Question 3

SYMBOLIC MATCHING TO SAMPLE

Answer 3

• comparison stimuli aren’t the same as sample; subject must learn “code” connecting them

Question 4

MULTIPLE SAMPLES

Answer 4

• design becomes test of recognition memory/list recognition task

Question 5

PROACTIVE INTERFERENCE

Answer 5

GRANT & ROBERTS (1973)
- ordinary DMTS on some trials
- other trials 2 samples presented (10/0s apart)
- animal must respond on last sample basis
- result = 0s performance (sample immediately followed by other) is worse
- good evidence for proactive interference in DMTS

Question 6

RETROACTIVE INTERFERENCE

Answer 6

GRANT (1988)
- ordinary DMTS
- brightness illuminating chambers increased in some trials during sample/comparison delay
- result = ^ illumination condition performance is worse
- NOT good evidence for retroactive interference in DMTS

Question 7

RADICAL ARM MAZE

Answer 7

• experiments w/rats
• means of presenting item list to animal
• animals allowed to run freely chosen arms in early exps; removed for delay interval; returned on test
• later experiments control the arms experienced on study phase

Question 8

RADICAL ARM MAZE: STRUCTURE

Answer 8

• central platform = trial start
• doors open to arms off central platform
• animal can run down arm to recessed food/baited sucrose pellet
• aim = animal shouldn’t revisit empty rooms aka. must avoid already visited arms subsequently

Question 9

RADICAL ARM MAZE: EARLY WORK

Answer 9

OLTON & SAMUELSON (1976)
- most striking result aspect: performance = v good BUT artefact of free choice procedure
- free choice procedures suffer from possibility that this isn’t so much due to memory as stereotyped response patterns on animal part (ie. always turning left after exiting arm)

Question 10

RADICAL ARM MAZE: 2AFC PROCEDURE

Answer 10

• after forcing animal to visit randomly chosen arm set; animal given choice between 2 arms (visited VS novel); required to choose novel
• aka. must learn visited avoidance
• remarkably good at all these trials

Question 11

RADICAL ARM MAZE: PROACTIVE INTERFERENCE

Answer 11

HOFFMAN & MAKI (1986)
- 8 arms partitioned into 2 sets (A/B); 4 arms each
- initial exposure phase to B arms; followed 2h later by study phase where rats forced to A arms
- next 2h -> free choice test phase among 8 arms; B set = rewarded
- control didn’t get initial interference phase
- exps = worse > controls; indicated PI effect of initial B arm exposure

Question 12

RADICAL ARM MAZE: RETROACTIVE INTERFERENCE (1981)

Answer 12

• 8 arms partitioned into 2 sets (A/B); 4 arms each
• initial exposure phase to B arms; 3 other similar mazes in dif rooms ran
• rat returned to test maze; performance in choosing A arms = worse > controls (experienced same delay between initial phase/test BUT not intervening mazes)

Question 13

RADICAL ARM MAZE: DECAY

Answer 13

ROBERT & SMYTHE (1979)
- basic 2AFC procedure; rats forced to 1/3/5/7 arms; given choice between visited/unvisited arms
- 3/5/7 = backward serial position plotting (1 = last arm visited, 2 = second last, etc.)
- aka. 3 condition: 3 = first arm visited
- rats run at roughly constant rate -> if elapsed time = only factor -> backward curves should superimpose
- within statistical error limits, they apparently did
- BUT issue in decay explanation; could also be driven by arm number visited post arm test (interference account)
- currently no way of telling

Question 14

ROBERT & SMYTHE (1979): EXPLANATIONS

Answer 14

DECAY-BASED
- arm visited already long time ago
INTERFERENCE-BASED
- rat visited lots of other arms afterwards
SOLUTION
- each rat ran 1 arm; hold in middle for time it took to run other 6 arms; test
- interference based = good performance
- decay based = equally poor performance (ANS)
- so… it IS decay

Question 15

RADICAL ARM MAZE: CODING

Answer 15

• memory code = retrospective or prospective?
• retrospective = remembering arm just visited
• prospective = remembering arms yet to visit
• animals definitely use retrospective encoding BUT can they employ a prospective code? can they switch between the 2 flexibly?

Question 16

DMTS: ROITBLAT (1980)

Answer 16

• animals make more mistakes when blue/orange = sample
• aka. give correct ans for orange to blue sample (vice versa)
• less confusions on red sample trials (though orange confusion = expected)
• suggests prospective code
• retrospective coding = remembering colour; should mix up similar colours BUT mix up line orientations (12.5 degrees VS 90 degrees) more aka. prospective code

Question 17

RADICAL ARM MAZE: COOK, BROWN & RILEY (1985)

Answer 17

• showed that rats could switch flexibly between retrospective/prospective codes
• 12 arm maze; forced to 2/6/10 arms
• taken out for 15m
• tested on visited VS unvisited arm

Question 18

RADICAL ARM MAZE: MCLAREN (1990)

Answer 18

• 8 arm maze w/15m retention interval
• more errors made in 4 arm condition > 6 arm
• Cook et al (1985) reported similar findings w/serial position effect in conditions showing highest error rate where arms entered early in trial = ^ likely to be incorrectly revisited post delay
• argument of indicating switch from retrospective -> prospective code at appropriate list length