Memory: Retrieving Memories from Long Term Storage



I. Capacity

A. generally estimated to be virtually unlimited

B. Landauer (1986)

1. previous quantified estimates
a. number of memories = number of synapses (1013)

b. number of neural impulses transmitted in a lifetime = 1020


2. taking into account forgetting, still estimated to be 1 billion bits of info (109)!

II. Coding in LTM
A. acoustic, visual, & semantic

B. but strongest evidence for semantic codes

1. false alarms

2. categorization

C. organization
1. recall evidence


III. Duration

A. Bahrick et al. (1975) & classmates

B. Bahrick (1984) & Spanish

1. drop off in 3 years

2. drop off again at 30-35 years out

3. evidence for permastore

C. Conway et al. (1991) & cognitive psychology
1. correlated with grades!
D. recall of spatial info at Ohio Wesleyan University (Bahrick, 1983) 1. learning data
  a. for street names: occurred at a steady rate over 36 months

b. for buildings & landmarks: steep and within the first year
 

2. retrieval data
  a. street names were forgotten quickly (after about 10 years)

b. buildings & landmarks: after 46 years, 40% was still retained

IV. Forgetting in LTM
A. Ebbinghaus: an early case study
1. learned nonsense syllables
a. 830 hours, 6600 lists, 85,000 syllables!
2. forgetting curve
a. rapid forgetting that levels off
3. savings
a. relearning was faster than learning on the first trial

b. the better remembered, the easier to relearn

B. decay
1. difficult to test, but probably occurs

2. compare sleep vs. waking

C. paired-associates learning 1. pair two words together, use first as a cue to recall the second

2. two phenomena

a. proactive interference
1. old learning interferes with new learning
b. retroactive interference 1. new learning interferes with old learning
3. with interference, more targets become associated with a single cue
V. retrieval from LTM
A. mnemonic devices

B. principles of retrieval

1. categorization

2. encoding specificity (Tulving & Thomson, 1973)

a. paired-associate learning with either highly-related, weakly-related, or no cues

b. weakly-related cues can be effective for retrieval when present at encoding

3. context effects
a. Godden & Baddeley (1975)
1. 2 X 2 design

2. underwater vs. on land

3. results support state dependent learning
 

b. state-dependent learning
1. odors

2. music

3. rooms - psychological context

4. drugs

5. moods (but less reliable)

c. affects recall rather than recognition
C. spacing effect
1. better learning when study is spaced rather than massed
a. encoding variability
D. cue overload
1. cues are most effective when they are unique & distinctive

2. Linton’s (1982) "diary" study

a. difficulty in locating routine memories


VI. Levels of processing (Craik & Tulving, 1975)

A. the problem
1. Crowder (1993) and recall of presidents

2. shows a "serial position" effect in LTM

a. but, we can’t use STM to explain recency

b. distinctiveness of encoding

B. depth of processing defined
1. different kinds of analysis from sensory to semantic

2. memorability = analysis at a deeper (more meaningful) level

3. maintenance vs. elaborative rehearsal

C. evidence
1. encoding = retrieval (Craik & Tulving, 1975)
a. physical, acoustic, vs. semantic encoding tasks

b. incidental vs. intentional learning

c. only type of processing predicted memory

2. self-reference effect (Craik & Tulving, 1975)
D. concerns
1. mundane conclusion

2. vague & untestable

3. circular definition


VII. Neurocognition & LTM

A. localized & distributed
1. hippocampus, cortex, thalamus, amygdala
B. anterograde amnesia
1. affects LTM, not WM/STM

2. affects memory regardless of modality

3. spares memory for general knowledge

4. spares skilled performance

5. when a skilled is learned, it is learned as hyperspecific memory

C. retrograde amnesia
 
1. all amnesics show some level of retrograde amnesia

2. features

a. time span for memory loss can vary greatly
  1. extensive memory loss for Korsakoff’s, Alzheimer’s, Parkinson’s & Huntington’s

2. temporally limited retrograde amnesia, with full or partial recovery, for head injured and ECT patients
 

b. most recent memories are most likely to be lost

c. spares information that was overlearned

d. spares skill learning

D. separate memory systems?



TOP | Back to PSY 307 Schedule