Baddeley

1
Baddeleys working-memory model
Two slaves and one master
2
Central ExecutiveHow can it be defined?

• component that coordinates processing of info in
  buffers (sketchpad, phonol. loop) in the service
  of task at hand
• -gt master of two slave systems
• executive processes include
• - focusing attention on relevant info and
  processes and inhibiting irrelevant (often
  automated) ones
• - planning sequence of subtasks to accomplish
  goal
• - scheduling processes in complex task (requires
  switching of attention between different parts)
• - updating and checking contents of WM to
  determine next step in sequence of parts
• example getting to university on snowy winter
  day

3
Central executiveNeuropsychological evidence in
support

• patients with brain damage in prefrontal cortex
  have problems with digit span backwards but not
  digit span forward
• cognitive impairment is different from WM deficit
  in patients K. F. and P.V. who have problems with
  digit forward and digit backward task due to
  defective phonological loop (following different
  brain damage)
• -gt suggests that maintenance of info in phonl.
  loop and manipulation by central executive rely
  on different parts of brain

4
Illustration of human brain
Left hemisphere
damage producing defect in central
executive damage producing defect in
phonological loop
Top-view of brain
5
Central executiveNeuropsychological evidence in
support

• patients with brain damage in prefrontal cortex
  are impaired on many problem solving tasks that
  build on WM
• -gt sometimes referred to as dysexecutive
  syndrome
• e.g. problems with Raven progressive matrix
  test problems with Trails B but not with
  Trails A test

6
Trail A test that relies little on central
executive
11
4
1
10
12
6
13
14
7
2
3
9
15
8
16
17
5
18
7
Trail B test that relies heavily on central
executive
a
4
1
b
e
6
f
i
7
2
3
9
d
h
8
c
5
g
8
Why does Trails B performance require central
executive?

• important feature that distinguishes Trails B
  from Trails A doing more than one thing at once
• -gt switching between 2 different task goals
  while maintaining count on sequence not in use
  in WM
• patients with dysexecutive syndrome make errors
  on tasks in everyday life that similarly rely on
  the central executive in WM
• e.g. following simple cooking sequence
• typical error putting powder in before filter
  

9
How can role of central executive in WM be tested
in normals?

• with dual-task paradigm
• use of random generation as secondary task
• e.g., random key pressing while performing
  calculations
• random key pressing has strong negative effect on
  mental arithmetic when calculation cannot be done
  with visual or writing support (study by Logie et
  al presentation last week)
• why does random generation have negative effect
  on WM tasks?
• requires active manipulation (i.e. generation)
  of info in WM

10
Problems with use of random generation to study
central executive

• task very complex
• -gt makes it difficult to specify which
  executive processes are blocked in primary
  task exactly
• e.g., for mental arithmetics
• processes affected could include any / several
  of the following
• - planning sequence of subtasks to accomplish
  goal
• - scheduling processes in complex task (requires
  switching of attention between different
  parts)
• - updating and checking contents of WM to
  determine next step in sequence of parts

11
Baddeleys working-memory model

• currently most widely accepted model that guides
  research on short-lasting memory processes
• not without criticisms though (see textbook p. 76
  79)
• e.g., definition of central executive?
• but model is constantly evolving to address
  criticisms and new data
• e.g. additional component recently introduced
  to explain how WM interacts with LTM episodic
  buffer

12
Atkinson Shiffrins modal memory model
13
Long-term memory (LTM)Is it unitary or are
there different components?

• Consider the following questions and see their
  diversity
• Did you have water at dinner last night or some
  other drink?
• How did you feel after your break-up with your
  first boyfriend / girlfriend?
• How many siblings do you have?
• Do all birds fly?
• Who sang the song Time after time?
• Is welderly an English word?

14
Long-term memory (LTM)Conceptual distinction of
two different components

• Tulving 1972/1983
• distinction between episodic memory and semantic
  memory
• episodic memory remembering the past
• - memory for episodes / events distinct in time
  and space
• - associated with mental time travel
• - conscious recollection of personal past
  experience
• (me-ness)
• semantic memory knowing about things learnt in
  the past
• - memory for world knowledge
• - timeless does not rely on mental time
  travel
• - does not involve conscious recollection
• - shared with others

15
Episodic memory as one type of LTM

• typical research method study-test paradigm
• Study try to memorize the following words for
  a memory test
• Test what were the words you saw 20 min ago?
• free recall
• cued recall
• recognition
• most tasks typically used in older research on
  LTM (Atkinson Shiffrin era) involve episodic
  memory

16
Information-processing approach to memorybasic
framework

• 3 stages of processing for manipulation of mental
  representations
• Encoding (acquisition of info)
• Storage (retention of info)
• Retrieval (recovery of info)

Encoding
Retrieval
time
Storage
17
Episodic memory Encoding processesExamples Try
to memorize following sentences
BMW just introduced a new 5 series model. The
German social-democratic party has shifted to the
right under its current leader Schroeder The
BOLD response is susceptible to movement
artefacts.
18
Episodic memory Encoding processesHow do we
acquire information about episodes?

• Role of repetition / practice
• Role of attention
• Role of intention
• Role of level of processing / elaboration
• Role of organization

19
Episodic memory Encoding processesRole of
practice

• Role of practice / repetition (rote learning)
• in a strict sense, does not apply to memory for
  episodes because each episode is unique in time
  and cant be repeated
• BUT repetitions can be construed as reminders
  for study episode that allow for re-encoding

20
Early research by Ebbinghaus on practice
effectsnumber of repetitions of list on first
day has dramatic effects on time needed to learn
it to perfect performance level on second day
21
Episodic memory Encoding processesRole of
practice

• much research undertaken to look at effects of
  distributed versus massed practice (starting
  with Ebbinghaus)
• -gt general result memory performance better
  when
• repetitions are separated (distributed
  practice)
• But amount of benefit from distributed practice
  depends on many factors, including
• interval between repetitions within study
  session
• -gt related to attentional capture of info
• rest-interval between successive study
  sessions
• -gt related to consolidat ion
• (biological processes that let info
  sink in)

22
Episodic memory Encoding processesRole of
practice

• despite general benefits of practice /
  repetition in episodic memory
• many studies with study-test paradigm have shown
  that repetition only of limited value for
  successful encoding
• e.g., Baddeleys studies on BBC change in radio
  frequency
• -gt what other factors are relevant then??

23
Episodic memory Encoding processesRole of
attention

• Does encoding of an episode require attending to
  it in the first place?
• Experiment by Murdock (1965) with dual-task
  paradigm
• - free recall task with list of unrelated words
• - secondary card-sorting task during learning
• - secondary task varies in attentional demands
• (different of sorting categories)
• Results
• of items recalled decreases as of sorting
  alternatives increases
• -gt encoding of information into episodic memory
  is attention demanding not all or none
  phenomenon

24
Episodic memory Encoding processesRole of
attention

• dramatic evidence from studies in surgical
  patients undergoing anesthesia (e.g. Millar
  Watkinson, 1983)
• subsequent to surgery, patients cant recall
  words presented auditorily during anesthesia
• -gt attending is critical for episodic memory at
  encoding

25
Episodic memory Encoding processesRole of
attention

• what is the role of attention during encoding?
• putting info into conscious awareness
• -gt we can only remember those aspects of an
  episode that we were consciously aware of in the
  first place
• e.g. remembering conversations from dinner
  party
• why is conscious awareness critical?
• Moscovitch hypothesis brain structures critical
  for encoding (i.e., setting up new memory
  representations) only receive info about
  consciously experienced aspects of environment

26
Episodic memory Encoding processesRole of
intention

• Do we need to make an active effort to memorize
  an episode in order to remember it later?
• e.g. what was the last movie you saw at a
  theatre?
• relates to distinction between incidental and
  intentional encoding

27
Episodic memory Encoding processesRole of
intention

• experiment by Mandler (1967) with pile of 20
  word cards
• encoding conditions
• incid. 1 sorting of word cards by category
• intent. 2 sorting of word cards plus
  memorization
• incid. 3 arrange word cards in straight columns
  
• intent. 4 memorization of words on cards only
• Results
• no difference in free recall for 1, 2, 4
• BUT subjects show poor recall in 3
• -gt intention to remember not a critical factor
• -gt instead type of processing of words is
  relevant

28
Episodic memory Encoding processesHow do we
acquire information about episodes?

• Role of repetition / practice
• Role of attention
• Role of intention
• Role of level of processing / elaboration
• Role of organization

29
Episodic memory Encoding processesLevels of
Processing Framework

• Craik Lockhart (1972)
• challenged Atkinson Shiffrins multi-store
  model
• memory performance is not simply determined by
  what store info is in (e.g. STM vs LTM)
• critical how info is processed at time of
  encoding -gt level / depth of processing
• different levels for processing of words
• perceptual feature analysis
• processing of sounds (phonology)
• processing of meaning (semantics)

Increasing depth
30
Episodic memory Encoding processesLevels of
Processing Framework

• Craik Lockhart (1972)
• empirical test of framework with incidental
  orienting tasks (level of processing induced by
  experimenter)
• shallow (perceptual / surface)
• Is this word in upper-or lower case letters?
  dog
• phonological (sound)
• Does this word rhyme with mat? hat
• deep (semantic)
• Does this word fit into the sentence
• the man ate his meal

31
Episodic memory Encoding processesLevels of
Processing Framework
typical advantage of deep over
phonological over shallow encoding on recognition
memory test
32
Episodic memory Encoding processesLevels of
Processing Framework

• important implication
• encoding processes are essentially those
  involved in the conscious perception and
  comprehension of episodes (no intention
  required)
• BUT encoding processes are attention-demanding
  nonetheless
• -gt attention and intention are not the same!

33
Episodic memory Encoding processesLevels of
Processing Framework

• further important component of framework
• elaborative rehearsal (versus maintenance
  rehearsal)
• Def. degree of semantic elaboration during
  encoding determines subsequent success in memory
  performance
• e.g. tiger
• what color does it have?
• how many legs?
• where does it live?

-gt elaborated representation is more distinct,
i.e. includes more features that distinguish it
from other words -gt more features can
subsequently serve as retrieval cues to get to
specific word in memory
34
Episodic memory Encoding processesLevels of
Processing Framework

• Experiment by Klein Saltz (1976) on
  elaborative rehearsal
• two incidental encoding conditions
• 1 categorize words on one semantic dimension
• egg does word refer to something pleasant
  or unpleasant?
• 2 categorize words on two semantic dimensions
  beer does word refer to something pleasant
  or unpleasant?
• does word refer to something animate
  or inanimate?
• Results more elaborate encoding in condition 2
  produces better memory performance

35
Episodic memory Encoding processesLevels of
Processing Framework
Why is semantic processing at encoding so
powerful for episodic memory?? -gt allows for
integration with existing (semantic)
knowledge e.g. try to learn name of person at
party either when you know or dont know his
friends e.g. What were the three sentences
you saw earlier? -gt learning difficult when
info cannot be integrated
36
Episodic memory Encoding processesIntegration
with existing knowledge
Try to recall the missing parts of the following
sentences BMW just introduced a new ??. The
German ?? party has shifted to the right under
its current leader Schroeder The ?? is
susceptible to movement artefacts. -gt existing
knowledge can act as cue at time of retrieval
37
Episodic memory Encoding processesLevels of
Processing Framework

• research on importance of integration with
  existing knowledge
• Experiments by Chase Simon (1973) on episodic
  memory for chess- positions (permitted vs
  arbitrary) on boards presented to experts vs
  novices in lab
• experts studied chess for approx. 25,000 h
• ( 12 years full-time job with 40 h/week)
• beginners 100 h total
• Results
• chess masters remember positions much better than
  beginners, but only if pieces positioned in
  configurations permitted by chess rules

38
Episodic memory Encoding processesLevels of
Processing Framework

• studies on chess also point to uniqueness of
  elaboration in different individuals
• -gt unique personal background (autobiography)
  relevant
• e.g. chess configuration previously
  encountered in real world game
• -gt what individual remembers about episode is
  determined by how episode is experienced by
  that particular person at encoding
• -gt subjective nature of encoding (me-ness)

39
Episodic memory Encoding processesLevels of
Processing Framework

• important consequences of unique elaborations
  during encoding on subsequent remembering
• -gt Schacters example of different individuals
  memory for
• Magrittes painting The Menaced Assassin in
  Museum of Modern Art
• Schacter
• memory recollections are largely at the mercy
  of our elaborations. we remember only what we
  have encoded, and what we encode depends on who
  we are

40
(No Transcript)
41
Episodic memory Encoding processesLevels of
Processing Framework
Why is LoP framework so attractive?? based
on simple principle memory as by-product of
ongoing cognitive activity (respecting true
incidental nature of most episodic learning)
offers new way of asking questions
e.g., is amnesic patients poor memory due to
difficulties with elaborative rehearsal?

42
Levels of Processing FrameworkCriticisms

• heuristic value of framework undeniable
• theoretical problems
• definition of levels
• e.g. when applied to stimuli other than words
• -gt danger of circularity
• optimal type of encoding may differ depending
  on type of memory test (situation)
• e.g., recognizing word vs hand-writing
• -gt encoding that is beneficial for one purpose
  may be less beneficial for another