Memory experiments of Ebbinghaus Examination of forgetting curve with savings method

1
Memory experiments of Ebbinghaus Examination of
forgetting curve with savings method

• most forgetting happens immediately after
  learning
• -gt approx. 50 in first 40 min

2
Forgetting on Brown-Peterson taskwith consonant
trigrams

• Repeat HLM
• Distractor task Count in 3s from 492
• What were the letters?
• many trials with different delays in same
  session
• after 6 sec only 40 correct recall

3
Forgetting on Brown-Peterson taskwith consonant
trigrams

• findings suggest short-lived memory component
  different from type of memory studied by
  Ebbinghaus

4
Something special about short lists

• Ebbinghaus lists with up to 7 nonsense syllables
  are forgotten at different rate than lists with
  more items
• G. Miller (1956) immediate memory span
• Magical number seven, plus or minus two
• holds for digits, letters, words
• (digit span, letter span, word span)
• -gt suggests that short lived memory
  component has limited capacity

5
Further examination of immediate memory span

• integration of information through chunking
  possible
• e.g. letters in meaningful word
• letters in acronyms (USA, IBM)
• area code in phone number
• chunks are basic storage units in short-lived
  memory component

6
Chunking produces similar forgetting curve on
Brown-Peterson task with words and trigrams
7
Further research on chunking

• optimal chunks can be unique for different
  individuals (e.g. running times dates)
• learning of new chunking methods possible but
  labor-intensive (e.g. digit-letter mapping
  system)
• when chunking leads to larger absolute number of
  digits being retained, capacity still remains
  unchanged
• -gt 7 2 chunks
• -gt change in coding of info instead

8
Consequences of findings with Brown-Peterson task
and with memory span in 1960s

• new notion memory is not unitary
• there may be at least two separate stores that
  work with different cognitive mechanisms
• they seem to differ in terms of forgetting rates
  and capacity
• -gt longer lasting component may not have any
  capacity limitations
• e.g. Ebbinghaus could remember lists of 40
  and more nonsense syllables
• everyday observations suggest that
  memory capacity for life events unlimited

9
W. James Primary vs secondary memory

• Primary memory
• it was never lost its date was never cut off
  in consciousness from that of the immediately
  present moment in fact, it comes to us as
  belonging to the rear-ward portion of the present
  space of time, and not the genuine past
• -gt extended present
• Secondary memory
• the knowledge of a former state of mind after
  it has already once dropped from consciousness
  or rather it is the knowledge of an event, or
  fact, of which we have not been thinking, with
  the additional consciousness that we have thought
  or experienced it before

10
Different proposals for two-store models

• W. James (late 1800s)
• primary memory vs secondary memory
• still in consciousness vs lost from
  consciousness
• Atkinson Shiffrin (1968)
• short-term memory (STM) vs long-term memory
  (LTM)
• Baddeley (1980s)
• further theoretical development of concept of
  STM -gt renamed working memory (WM)

11
Consequences of findings with Brown-Peterson task
and with memory span in 1960s

• controversy over need to postulate two stores
• philosophy-of-science argument relevant
• more parsimonious theory better than more
  complicated one if it can explain the same
  findings (principle of Ochams razor)
• -gt is there sufficient number of critical
  findings that single-store theory cannot
  explain??

12
Forgetting mechanisms in STM vs LTM

• predominant explanation of forgetting on
  list-learning tasks (à la Ebbinghaus) and others
  used in behaviourist verbal-learning research
  interference
• e.g. for paired-associate learning task
• First list Second list
• car ball car fridge
• tree screen tree foot
• table paper table ocean
• door sports door hair

13
Forgetting mechanisms in STM vs LTM

• retroactive interference on paired-associate
  learning task
• What was the missing word from first
  list? tree ?
• door ?
• learning of second list impairs recall of first
  list
• interpretation in behaviourist learning theory
• associations between pairs in first list weakened
• application in everyday life
• previous postal code after move

14
Forgetting mechanisms in STM vs LTM

• suggestion for forgetting on Brown-Peterson task
  trace decay (fading)
• - occurs as soon as stimulus absent
• - can be prevented through active rehearsal
• - different from interference on LTM tasks
• -gt numbers in distractor task known to produce
  minimal interference for letters
• -gt forgetting observed on Brown-Peterson task
  reflects trace decay
• - interpretation later challenged by Keppel
  Underwood

15
Other evidence to support distinction between STM
and LTMSerial-position curve in free recall task
recency effect
primacy effect
16
Effects of presentation rate and delay on serial
position curve (Glanzer Cunitz, 1966)
17
Interpreting effect of presentation rate and
delay in Glanzer Cunitzs experiment

• Presentation rate
• more rehearsal allows for better transfer from
  STM into LTM in primacy portion
• Delay
• trace-decay in STM eliminates recency
• -gt different experimental manipulations have
  different effects on recency and primacy
  portion of curve
• -gt suggests that primacy and recency effects
  reflect operation of different stores (LTM,
  STM)

18
Other evidence in support of two-store
interpretation of serial position curve

• detrimental effect of old age only on primacy
  portion
• beneficial effect of word familiarity only on
  primacy portion
• (more familiar words recalled better than less
  familiar ones)

19
What makes a list difficult for STM (Baddeley,
1966)?
unrelated
semantically similar
acoustically similar
detrimental effect of acoustic but not semantic
similarity on immediate recall of short word lists
20
What makes a list difficult for LTM (Baddeley,
1966)?
detrimental effect of semantic but not acoustic
similarity on learning of long word lists
21
Interpretation of difficulty findings in
Baddeleys experiment

• detrimental effect of acoustic similarity on
  recall of brief lists suggests that info is
  coded phonologically (in terms of sound) in STM
• detrimental effect of semantic similarity on
  recall of longer lists suggests that info is
  coded semantically (in terms of meaning) in LTM

22
Other evidence suggesting semantic code for LTM

• prose passage experiment by Sachs (1967)
• e.g. sentence heard as part of brief story
• she watched the kids on their way home
• Was the following sentence part of the story?
• - she watched the birds on their way home
• (no easy)
• - the kids were watched by her on their way home
• (no difficult)
• - she observed the kids on their way home
• (no difficult)

-gt good recognition of semantic but poor
recognition of verbatim information after delay
23
What kind of memory is impaired in neurological
patients suffering from amnesia?

• amnesia can occur as result of sudden brain
  injury (e.g car accident) or stroke
• - patients do not remember anything from hours /
  days before incident
• - patients have difficulty keeping track of
  daily events
• - patients have difficulty learning names of new
  people (e.g. physicians and nurses)
• - patients have difficulty learning way around
  in new environment (e.g. hospital)
• amnesic patients have memory deficits but
  otherwise normal intelligence -gt can be tested
  on experimental tasks used in memory research
• critical question does amnesia affect STM, LTM
  or both?

24
What kind of memory is impaired in neurological
patients suffering from amnesia?

• general findings in amnesic patients
• normal STM capacity on digit-span task
• normal forgetting curve on Brown-Peterson
  task
• could suggest that only LTM affected by amnesia

25
Serial position curve in amnesic patients
(Baddeley Warrington, 1970)

• most aspects of list learning impaired but
  recency portion in immediate recall normal in
  patients
• -gt further support for idea that amnesia affects
  LTM but not STM

26
Additional evidence showing that memory deficit
is specific to LTM in amnesiaFindings with
span 1 task (Drachman Arbit, 1966)
lists larger than STM span extremely difficult to
learn for patients
27
Can other types of brain damage lead to impaired
STM?

• patient K.F. studied by Shallice and Warrington
  (1970)
• brain damage different from that in amnesic
  patients
• immediate memory span limited to 2-3 digits
• poor performance on Brown-Peterson task,
  especially with auditory presentations
• normal performance on list learning tasks with
  long lists
• normal primacy but impaired recency effect
• -gt patients memory deficits seem to be limited
  to STM LTM unaffected by damage

28
What do patient studies tell us about distinction
between STM and LTM?

• observed pattern of deficits across studies
• one type of neurological condition leads to
  impairments in LTM but not STM
• other type of neurological condition leads to
  impairments in STM but not LTM
• -gt evidence for double-dissociation of STM and
  LTM deficits
• -gt suggests that LTM and STM require normal
  functioning of different brain structures
• -gt suggest that LTM and STM have distinct neural
  basis