Transcript Components of memory - University of Leicester

Components of memory
Components of memory and implicit memory
PS2011 & PS2016: Cognitive Psychology
John Beech
1
Components of memory
Overview of the 3 memories
We process information very briefly in its raw form
in “sensory memory”. Large capacity. Rapid
decay.
Short-term store of limited capacity for period of 1530 sec. Rehearsal maintains it. Forgetting slow.
Long-term memory has large capacity with very
slow decay
2
Components of memory: Atkinson &
Shiffrin (1986)
3
Components of memory: iconic
memory
Sensory memory
Neisser (1967): “iconic memory” (E.g.
cigarette or sparkler in dark) and “echoic
memory” (e.g. saying “what?” and then
knowing what was said).
4
Components of memory: iconic
Sperling (1963) – 3-row matrix of letters:
Matrix shown for 100msec and S
recalled “whole report” of 4 letters on
average
Impression of seeing whole array
Matrix shown 100ms again, then tone
signalled top, middle or bottom of array
to recall.
If 67% of row recalled, he argued that
67% of whole array must therefore be
available.
Varying probe length showed steep
decline in memory to asymptote after
only .25 sec.
K
X
R
V
P
V
J
M
C
5
Components of memory: iconic
memory
Properties of sensory memory
Large capacity.
Limited only by acuity of retina
Information is mainly unprocessed:
6
Components of memory: iconic
memory
Sperling (1960): 2 x 4 matrix of letters and digits.
When probed for position got usual partial report
superiority. But when probed for digits or
letters, no partial report superiority. Therefore
only the visual representation and its position is
stored, not the identity of the items. It appears to
be primarily a visual store.
3
W
J
6
K
8
5
H
7
Components of memory:
working memory
Listening to or reading and understanding text
involves short-term memory to hold information
temporarily while it is being processed.
Remembering a phone number before writing it
down involves this memory. Known as working
memory and can be used visually, spatially, in
mental arithmetic, in reasoning and problem
solving as well. It is important for transferring
information into long-term memory. It works as a
temporary holding store during processing. WM
is important for several reasons….
8
Components of memory:
working memory
Efficient use of WM probably enhances cognitive
functioning
Kyllonen & Cristal (1990):
Suggest people with high IQs have extra storage
capacity in working memory.
WM appears to be linked to mood
Teasdale et al. (1993):
Suggested depressed person keeps generating
negative thoughts. In turn affects mood. This
sequence is in WM. Should be able to kill cycle
by occupying WM with other things.
9
Components of memory:
working memory
Necessary for reading
comprehension
Perfetti (1985):
WM has an important role
in skilled reading of
text. It maintains
information in memory
during the processing of
reading.
10
Components of memory:
working memory
Experiments often artificial when examining
it.
Peterson & Peterson (1959):
Consonant trigrams (e.g.TDK) followed by a
number (e.g. 765). Counts back in 3s (765,
762, 759…) and recalled trigram on signal
Rapid decline to <10% after 18s.
Thus if rehearsal inhibited, decay is rapid…
11
Components of memory: Peterson & Peterson (1959)
12
Components of memory:
working memory
Patients with hippocampal lesions (removal of
their hippocampus) cannot transfer
information to long-term memory. But they
can remember information before their
operation. The hippocampus appears
important in transferring information from
WM to LTM
13
Components of memory:
working memory
Evidence for STM-LTM distinction from
Glanzer & Cunitz (1966):
Recalling list of words give U-shaped effect
Expt 1: varied presentation rate. Improved recall on
start and middle bits, but not on the last part.
Slowness helps LTM but not STM.
Expt 2: Immediate recall vs delayed interference
(counting): affects only last part showing an effect
on “recency”.
14
Components of memory: working memory
E vid e n ce fo r o p e ra tio n o f
in te rfe re n ce : S e ria l p o sitio n cu rve
90
80
70
60
50
Recall accuracy
40
30
20
10
0
1
2
3
4
5
6
7
8
9
S e ria l p o sitio n
10 11 12
15
Components of memory: working memory
E ffe c t o f re c a ll d e la y o n th e s e ria l p o s itio n c u rv e
• D e la y w ip e s o u t th e re c e n c y e ffe ct,
b u t n o t th e p rim a cy e ffe ct.
16
Components of memory:
working memory
Changing presentation rate (Expt 1) affects
processing in LTM, but does not affect
STM. By contrast, distraction that prevents
rehearsal (Expt 2) affects STM but not
LTM.
17
Components of memory:
working memory and its capacity
Miller (1956) suggested memory is 7 plus or
minus 2 “chunks”. Chunks implies not
necessarily discrete items.
Postman (1975): if letters organised into 3letter syllables, can remember 21 letters.
Letters and digits accord with 7+/- 2, but
fewer syllables, still fewer words and fewer
sentences.
18
Components of memory:
working memory and its capacity
Chase and Ericsson (1978): tested a
person keen on track events.
“4729” became 4 min 7.29 sec, as
a finishing time. Initially he
remembered about 7 finishing
times, so could repeat 20-30 digits.
After practice eventually he could
recall string of 79 digits. After 6
months, switched to letters and
back to 6-letter span.
19
Components of memory: further
work on the recency effect
‘Recency’ used to be considered synonymous with
working memory. Last items still in WM and easy
to recall. But can produce recency in other ways.
It is possible to get a recency effect after distractors
are given: Bjork & Whitten (1974): participant
given series of items and each and every one is
followed by a 20 sec distractor. Would expect
the recency effect to be wiped out, but it is not!
Very tedious experiment.
20
Components of memory: further
work on the recency effect
One can show effect of recency in long term
memory
Crowder (1993): People asked to recall
American presidents. There is a clear
primacy effect and a strong recency effect
for Ford, Carter, Reagan and Bush.
Effect also happens for lots of events in
episodic memory: holidays, rugby players
recalling their matches, recalling car
parking. It’s a long-term recency effect.
21
Components of memory: further
work on the recency effect
Could be explained by distinctiveness.
Because of vast store in LTM, we
remember better first and last items.
If this is the case, the recency effect may be
affected by location in WM, but is also
affected by something else. Possibly
distinctiveness.
22
Conceptualising WM
Boxes used to present processes as schematic
presentation
Does one really have something like a desktop?
Transferring information into filing cabinets?
This is too static. WM could be activation of long-term
storage areas.
Analogy of an orchestra. Most silent but small section
playing. The sound analogous to WM. Overall
orchestra represents LTM
23
Summary so far of memory
components
Sensory memory appears to have a large
capacity but information decays very
rapidly.
It is a visual process, as it is affected by
darkness.
It is unprocessed as processing makes the
iconic effect disappear
24
Summary so far of memory components
Working memory (WM) is longer in duration and
transfers information into long-term memory.
• It probably can serve to improve IQ, it can keep a
depressed person depressed, it can help reading
comprehension
• Evidence suggests it declines rapidly if one is
distracted
• Patients with hippocampal damage do not seem to be
able to transfer information to LTM
• WM and LTM can be independently manipulated
experimentally (Glanzer & Cunitz)
25
Summary so far of memory components
Criticisms
The recency effect can be shown to occur even after
the use of a distractor and also in LTM.
Using boxes may distort our understanding. Some
memory processes may be more dynamic, as
illustrated in using the orchestra analogy.
26
The Baddeley & Hitch (1974) model
27
The Baddeley & Hitch (1974) model
At the time they thought that experimental work on
STM had not examined the interaction between
thinking about something (processing) and
memory. So they coined the term “working
memory”.
Central part is a central executive that has conscious
control over everything.
Has a number of slave systems. These are mundane as
they store, but not much else. If anything mentally
demanding, the slaves can’t do it, so hand over to
the central executive.
Just like when we jot things down
28
The Baddeley & Hitch (1974) model
Two slave systems are the visual-spatial
scratch pad and the rehearsal loop.
There is a rehearsal loop from the executive to
the phonological buffer and back.
The inner voice (subvocalisation) can run
independently once initiated as speech can
be automatic. The buffer is passive and
information decays in it. This leaves
executive free to process.
29
Some WM findings
E.g. If had to say “the” repeatedly, the loop couldn’t
operate. Performance falls markedly as inner
articulation disrupted.
If repeating “the” and shown complex shapes, no
effect.
Similarly, inner ear can be disrupted by noisy
conditions. Span goes down with noise.
Executive and phonological buffer have disrupted
communication.
30
Some WM findings
There is a word-length effect. Monosyllabic words
have a greater span than 3-syllable words.
Obviously, it takes longer to pronounce 3-syllable
words.
Those with anarthria have partial or complete loss
of articulate speech. Tongue paralysed. They also
show word length effect. Actual muscle
movements not needed for subvocalisation. Areas
involved in planning movement of speech
musculature perhaps involved.
31
Summary and evaluation of WM
It is of course a theory. We can only infer the
existence of an inner ear or voice.
Various types of findings support the theory. For
instance, we can use concurrent articulation, the
word-length effect. Findings from
neuropsychology.
Good example of a theory in which data could have
gone against it. No line of enquiry has been
decisive. But has built up an overall picture.
Working memory is a very important system and has
a key role in many cognitive aspects and even in
mood.
However, the concept of an executive is problematic.32
Long term memory
This is a store with unlimited capacity that
holds information for a long period. All
that can be remembered is held for virtually
a life time, or at least while the brain is
capable of retrieval.
It is believed that LTM is organised into
components which contain different types
of information.
33
Long term memory: its parts
34
Long term memory: its parts
Declarative vs. procedural memory (Squire
1987). Procedural memory (or implicit
memory) is not available for reporting. It
has the effects of learning skills and
behavioural responses and is unconscious.
Declarative memory (or explicit memory)
has facts, information, ideas, images, etc.
We can retrieve this information and are
conscious of it.
35
Evidence for the declarative vs
procedural distinction
Amnesiacs who have lost much
declarative memory (people,
places, events) still retain skills –
can swim, can play golf, etc.
36
Semantic vs. episodic memory
(Tulving, 1972, 1983)
Semantic memory is about facts, concepts and
ideas; whereas episodic memory is about
personal experiences and is temporally
dated.
Evidence: Lefrancois (2000) KC plays chess
and knows that he can, but cannot
remember playing a game.
37
Explicit vs implicit memory (or
declarative vs procedural memory)
Definition
Explicit memory is basically when conscious
memory is used. Implicit memory applies
when conscious memory is absent.
Explicit memory: direct memory tests e.g.
recall and recognition.
Implicit: priming – indirect testing.
38
Implicit and explicit memory
Tulving, Schacter & Stark (1982)
Ss learned rare words (e.g. “honeycomb”)
1hr or 1wk later gave (_o_ey_o_b). ½ were from
learned list and ½ new. Accuracy good over time.
Fragment completion task (an indirect test)
indicates implicit memory persists quite well.
Recognition memory (testing explicit memory) of
the learned list good after 1hr (60%) but poor after
1wk (20%).
Implicit memory persists quite well over time, but
recognition memory, testing explicit memory,
declines over time.
39
Data driven vs conceptually driven
Roediger’s (1990) theory
“Data-driven” means evoked by external
stimuli while “conceptually driven” means
produced by the person.
Data driven processes are not consciously
evoked, so often test implicit memory.
Conceptually driven ones test explicit
memory.
40
Implicit vs. explicit memory
Jacoby (1983): Phase 1: Ss shown word
sequence.
XXXX, DARK
No context: shown each word
without context. They had to
read aloud dark.
HOT, COLD
Context: shown each (cold in
this case) with antonym.
LOW, ????
Generate: Saw only the antonym
and had to say aloud high.
41
Jacoby (1983): Phase 2
During Phase2 one test involved recognition
memory (testing explicit memory) and the
other involved perceptual identification
(implicit memory).
The perceptual task involved brief presentation
of the 1st phase words or new words. Words
previously seen (in the no context and
context conditions) had been “primed”, so
they were identified more easily. As the
words were not seen in the generate
condition, we would expect poorer
performance in the generate condition.
Explicit memory was best in the generate
condition, while implicit memory was best
in the no context condition.
Implicit vs.
explicit memory
XXXX,
DARK
No context:
shown each
word without
context. They
had to read
aloud dark.
HOT,
COLD
Context:
shown each
with antonym.
LOW,
????
Generate: Saw
only the
antonym and
had to say
aloud high.
42
Implicit vs. explicit memory
Thus explicit memory, requiring conscious
(and active) processes, was best in the
“generate” condition, needing the most
conceptualisation. But conceptualisation is
actually detrimental to implicit memory,
which works best in the absence of
conscious memory.
43
Implicit vs. explicit memory
Aspect
Implicit
Explicit
Conscious
memory
Persistence
Absent
Present
Good
Poor
Roediger
Data-driven
Conceptually
driven
Good
Generation task Poor
Priming task
(no context)
Good
Poor
44
Implicit vs. explicit memory: evaluation
The data-driven vs conceptually driven distinction is
probably an important one.
The distinction involves an element of judgement.
Some would argue that most tasks could involve a
mixture of the two.
There could be different types of implicit memory.
Priming performance on perceptual identification
and in fragment-completion was found to be
unrelated by Witherspoon & Moscovitch (1989),
even though they are both testing implicit memory.
45
Implicit vs. explicit memory: a summary
Implicit memory is used when one does not use
conscious processes.
We can understand the implicit vs. explicit memory
distinction in terms of data-driven and
conceptually driven processes. Data-driven
processes are “driven” by encountered stimuli not
consciously processed.
There are likely to be different sorts of implicit
memory.
46