Transcript Implicit memory

Chapter 5
Learning and Long-term Memory
GP Mem(1)
1
A Little History
Ebbinghaus (1885)
- nonsense syllable
method (e.g., DAX,
BUP, GEJ)
Tested his own
memory
Each list learned to
criterion of 2 errorless
tests.
GP Mem(1)
2
IV: Retention time (minutes, hours, days)
Dependant Measures
1) Number recalled
- fast loss at beginning, then levels out.
GP Mem(1)
3
2) Savings # of trials to relearn
Large Savings at short retention,
Rapid decreased with longer retention
Then leveled out.
E.g., After 6 days 30% savings.
GP Mem(1)
4
3) Overlearning
- learn to criterion then rehearsed thirty times more.
Greatly increased savings.
E.g., At 6 days 64% savings.
Appears Rehearsal leads to Memory.
GP Mem(1)
5
Does Rehearsal transfer information to LTM?
Glenberg, Smith and Green (1977) Study.
- subjects presented 4 digit numbers
- each digit followed by word to repeat for 2, 6 or 18 secs.
Told to remember digits (word repetition just distracter)
Tested for the words.
Increased rehearsal did not lead to better recall from LTM.
Incidental Learning Task – Participants did not intend to
learn the words.
GP Mem(1)
6
Recall vs. Recognition
Glenberg, Smith and Green Study (Revisited)
Incidental recall test showed no effect of
rehearsal.
Incidental recognition test showed that the
more rehearsed words were recalled better.
cogltm(1)
7
Rehearsal increased the familiarity of the
words, but not the ability to recall the words.
Recognition depends on familiarity
Recall requires that you find the words in
LTM and then recognize that they are the
words you are trying to recall.
cogltm(1)
8
Depth of Processing (Craik and Lockhart)
Transfer to LTM requires processing for meaning.
Incidental vs. Intentional Learning.
LTM is like a library!
To encode information, there must be a place for it.
Added to similar information.
Elaborative Rehearsal - thinking about meaning.
- relate to other things you know.
cogltm(1)
9
Depth of Processing (Craik and Lockhart)
Transfer to LTM requires processing for meaning.
Incidental vs. Intentional Learning.
When you experience a stimuli you retrieve information from
LTM to make sense of it. The new information is added to
the old information.
To figure out these are dogs
You retrieve your general
knowledge about dogs.
Then you add the new
information, “Some dogs play
cards”! New is stored with old.
cogltm(1)
10
GP Mem(1)
11
Slamecka and Graf (1978) conducted studies
in which they compared students’ ability to
recall information when they had been
provided with cues, and when they were
required to make-up their own memory cues.
They found that when students generated their
own cues they were more likely to recall the
information they were required to encode.
This is called the Generation Effect.
GP Mem(1)
12
Bransford et al. (1979) showed that long-term
memory depends on the kind of elaboration as well
as the amount.
Minimally elaborated similes
“A nurse is like a mosquitoes because they both
draw blood.”
Multiple elaborative similes
“A nurse is like a raccoon because they both have
heads, legs, jaws, ears”
Minimally elaborated were better recalled were
better recalled perhaps because they are more
distinctive.
GP Mem(1)
13
Self-Reference Effect
Tendency for individuals to have better
memory for information that relates to oneself
in comparison to material that has less
personal relevance.
GP Mem(1)
14
Many researchers became interested in exactly what
produced deep processing:
Elaboration: Craik and Tulving (1975) found complex semantic
processing produced better recall
e.g. “The great bird swooped down and carried off the struggling
___” produced better recall than “She cooked the ___”
Distinctiveness: Eysenck and Eysenck (1980) found that words processed
phonetically even were better remembered when they were distinctive or
unusual (e.g., Yatch, Comb, or Paradigm).
Effort: Tyler et al. (1979) found there was better recall with difficult
anagrams than simple anagrams
e.g. “ORTODC” had better recall than “DOCTRO”
GP Mem(1)
15
Von Restorff Effect (named after Hedwig
von Restorff, 1933) predicts that salient items,
those that "stands out like a sore thumb",
(called distinctive encoding) will be more
likely to be remembered than other items.
Salience can be anything that makes the item
stand out!
GP Mem(1)
16
What is deeper processing????
“Deeper” processing connects the material to
be recalled to other material, integrating it
with already stored information.
- Mental work
- Organization
cogltm(1)
17
But what is Deeper? The term
“deep” is not well defined.
• Depth of processing also tends to mean
more time spent processing; it might be
that it is the length of time spent
processing which affects the memory
trace, not the depth of processing
• There may be more effort involved in
“deeper” processing and the greater
effort might account for the better
recall.
GP Mem(1)
18
Focuses only on encoding.
Transfer-Appropriate Processing
Memory will be best when the processes
engaged in during encoding match those
engaged in during retrieval.
What to do if you catch on fire!
GP Mem(1)
19
The Testing Effect
GP Mem(1)
20
Implicit and Explicit Learning
• Explicit learning: deliberate learning that
one is consciously aware of
• Implicit learning: learning that is acquired
without awareness
– Can we learn without awareness?
Research into Implicit Learning
Sequence Learning Serial reaction time task
• The task was to hit the key (corresponding to
where the “X” appeared) as fast as possible.
Computer screen
X
• Unknown to participants, the appearance was
based on an underlying sequence (8-12 items
long)
Response key pad
• Over time, people might get faster due to
anticipation of this sequence
• The question is, would the improvement be
accompanied by explicit knowledge of the
sequence?
Serial reaction time task –findings
• In many cases, the
sequence group were
unable to report the
sequence, despite vast
improvements in RT
• Seemed to have
something to do with the
complexity of the
sequence, and the length
of practice
• What is with B10?
800
700
600
500
400
300
200
100
0
B1
B2
B3
B4
B5
B6
Sequence
B7
B8
B9
Random
The sequence is broken.
When questioned the participants show no
conscious awareness of the sequence!
B10
B11
B12
Retrospective problem – people may be
consciously aware of what they have learned
at the time of the task, but have forgotten at
then end of the experiment when asked for
their explicit knowledge.
Wilkinson & Shanks (2004)
Inclusion task (try to guess next location)
Exclusion task (don’t try to guess)
- Better when asked to guess indicating that
there is some conscious knowledge!
GP Mem(1)
24
Artificial Grammars
Study the following letter strings for 5 minutes:
A-D
C-D-B-D
A-C-A-D-B
C-A-D
A-C-A-C-A-D-B-D-B
C-A-D-B
A-D-B
A-B-B-B
C-A-B-B
C-D-B-D-B
C-A-B
A-C-A-C-D-B
GP Mem(1)
25
TEST PHASE : For each of the following, simply
tick those that are legal and place a cross against
those that are illegal
C-B
B-D-B
C-A-D-B-D
C-B-D-B
C-A-B-D-B-D
C-D-B-A
C-D
A-B-D
C-A-D-C-D-C
A-B-D-B-D
GP Mem(1)
26
The grammar contains a fixed number of letters, and
a fixed number of rules that determine the order and
length in which “words” (strings of letters) can be
created. For example, suppose I create a grammar
consisting of three letters, A, B, C, and I make three
rules: If there is an A, it must always precede C;
each letter can appear only once, except B, which
can be repeated up to three times. Under this
grammar, I can create several words, such as A-C,
A-B-C, B-C, B-B-C, A-B-B-C, C-B, and so on.
However, other combinations such as, A-A-C, and
C-B-A, are illegal, since they break the rules of the
grammar.
GP Mem(1)
27
Diagrams
GP Mem(1)
28
Shows a form of implicit learning, since one measure
has shown that you have learned something
(recognition task), while a second measure of your
awareness of what you have learned (stating the
rules of the grammar) indicates poor explicit
learning.
It may be that this is how many people acquire
grammar in natural language. Do you think that
people who have not been taught the rules of
grammar can still detect when a rule is broken, yet
be unable to state what the rule is?
GP Mem(1)
29
Differences between Implicit and
Explicit Learning
1. Robustness – unaffected by disorders.
2. Age independent
3. Low variability (individual differences)
GP Mem(1)
30
3. IQ independent
4. Commonality of process – all species
GP Mem(1)
31
Age and Intellegence Differences
in Implicit and Explicit Memory
Howard & Howard (1992)
Serial Reaction time study with older and younger adults.
Both groups showed implicit learning – did not differ
from each other.
Explicit test (predict the next position) – younger group
was better.
Gebauer & Mackintosh (2007) - same type of results
found in relation to intelligence.
GP Mem(1)
32
GP Mem(1)
33
Evidence from Neuropsychology
Parkinson’s Disease
Damage to Basal Ganglia
which is thought to play a role
in implicit memory.
Show larger impairment of
implicit than explicit learning.
GP Mem(1)
Substantia nigra
34
Parkinson’s patients however also have
impaired explicit learning when the task
is difficult.
Conclusions: Basal Ganglia plays a role
in implicit memory – and perhaps a
smaller role in explicit memory.
GP Mem(1)
35
Double Dissociation – Damage to Medial
Temporal lobes and Hippocampus
Amnesiacs have damage to
Medial Temporal lobes and
Hippocampus. Show Implicit
learning and no Explicit
Learning.
GP Mem(1)
36
Brain Imaging
In general these are consistent with the
evidence from brain damaged patients.
GP Mem(1)
37
Long Term Memory systems
Declarative Memory – conscious recollection
Memories can be “declared” or “described”
Nondeclaritive Memory – no conscious
recollection
Claparede "Sometimes people hide
pins in their hands".
GP Mem(1)
38
Divisions of Declarative LTM
Episodic/Semantic (Tulving, 1972): Researchers
seemed to be studying two kinds of memory when
you looked at the experiments closely.
Episodic Organized around time cues and
experience cues. E.g., when remembering a word
list, it is not just the words we are recalling but that
they are tied together by the fact that they were
learned at the same time from the same list. (what,
where, when).
Autobiographical Memory similar to is episodic but
there’s a sense of your presence associated with the
GP Mem(1)
39
memories.
Semantic Memory: Fact knowledge. This is like
compiled episodic memories. Memories are
organized around meaning cues and are divorced
from your experience of the episode in which they
were learned.
GP Mem(1)
40
Nondeclarative Memory
Implicit memory: Memory that does not involve
conscious recollection of information.
Procedural memory: Skill learning and habits.
(e.g., keyboard skills, bike riding)
Serial reaction time task is a test of procedural
knowledge.
GP Mem(1)
41
Structure of LTM
Non-Declarative
Declarative
Implicit
Memory
Episodic
Memory
Procedural
Memory
Semantic
Memory
Episodic Memory
Does not work like a Video Tape.
What if it did??
Jill Price remembers all the sad and bad things in her life –
the death of loved ones, for instance, like it’s happening
right now. Time heals all wounds, but not for Jill Price.
GP Mem(1)
43
Solomon Shereshevskii
The Mind of a Mnenonist (Luria, 1968)
Eidetic memory - Solomon remembered nearly everything that ever
happened to him in his entire life, even his mother's face as she stood over
him in his crib as an infant. He could remember a passing car's license
plate number from twenty years past. He was able to memorize lists of
70 numbers after being read them once, then repeat them forwards and
backwards and retain them for upwards of 15 years.
Synesthesia - if a doorbell rang, he wouldn't just 'hear' it. He wouldn't
even 'hear' and 'feel' it. He would hear, see, taste, feel and smell the
doorbell. Anything that he experienced was not only stamped in his mind
for good, it was stuck there with a connection to all five of his senses.
GP Mem(1)
44
Solomon tried out several different methods of making his
memories less invasive. One method was to write down all
the things that he no longer wanted to remember and
then burn up the paper he'd written them on. Another was
to imagine covering up all the memories he didn't want with a
blank canvas in his mind. He was eventually able
to will some of his memory away, but not much.
Shereshevskii eventually quit his reporting gig and bounced
around from job to job for a while. He ended up becoming a
professional mnemonist, wowing folks with neat memory
tricks and getting paid for it.
Want to know more?
GP Mem(1)
45
Reconstructive Memory
Remembering the past and imagining the future
are neurologically similar. Both utilize area
associated with semantic information about one’s
life, self-referential processing.
GP Mem(1)
46
Remembering
Imagining
Asked to either conjure up a memory or
imagine a future scenario about some
particular cue-word the same sites in the brain
light up in both cases.
GP Mem(1)
47
Research with Amnesiacs
Amnesic patients with primary damage to the
hippocampus bilaterally were markedly impaired
relative to matched control subjects at imagining
new experiences. “The patients' imagined
experiences lacked spatial coherence, consisting
instead of fragmented images in the absence of a
holistic representation of the environmental
setting” (Hassabis, 2007).
GP Mem(1)
48
Recognition Memory
Remember/ Know Task (Tulving)
• Know = no contextual details; just a
feeling of familiarity with the item.
• Remember = contextual details (e.g.,
associations).
On recognition task participants indicate
whether they “know” a word was on the list or
“Remember” that it is.
GP Mem(1)
49
“What” and “Where/How” Pathways
Binding together of the two types of
information is needed for recollection, but not
for familiarity.
Binding appears to depend
on the hippocampus.
GP Mem(1)
50
Skinner & Fernandes
(2007) – amnesiacs
with hippocampus
damage show
impaired recollection
compared to controls,
but comparable
familiarity judgments.
Figure 5.5 Mean recollection and familiarity estimates for healthy controls, patients
with medial temporal lobe (MTL) lesions or brain damage, and patients with non-MTL
lesions. Reprinted from Skinner and Fernandes (2007). Copyright © 2007, with
permission from Elsevier.
Semantic Memory
Consists of general knowledge about the world,
concepts and language.
Some Examples of Semantic Memory
The meaning of letters
The concept of what a cat is
The sounds that letters make
The idea of what a car is
How letters put together can make a word
Knowledge that Annapolis is the capital city in Maryland
Word definitions
The dates of when World War II began and ended
Knowing what a dinosaur is
Recalling who wrote "Walking Through Woods on a Snowy Night"
Knowledge of what a phone is used for
Knowing how to make a pulley
Recognizing color names
Knowing where Mongolia is located on a map
Remembering what a computer is
The knowledge that fish swim in water
Knowledge that owls are nocturnal
Knowing deer can be hunted
Knowledge of what constitutes a chair
Knowing what a broom is used for
To know where to find a parrot
Remember what foods are popular in Cambodia
Knowledge of what a book is
GP Mem(1)
Recollection of customs of regarding weddings in Syria
52
Tulving suggests episodic
memory is more recently
evolved, later developing
(i.e., infantile amnesia)
and earlier deteriorating
than semantic memory.
Amnesias often retain
semantic memory while
having greatly impaired
episodic memory.
GP Mem(1)
53
Schemas
Organized packets of knowledge about the
world.
GP Mem(1)
54
Read the following;
‘When the man entered the kitchen, he
slipped on a wet spot and dropped the
delicate glass pitcher on the floor. The
pitcher was very expensive and everyone
watched the event with horror.’
(Bransford, 1979)
Now write out the paragraph you have read,
word for word.
How much can you remember?
The Constructive approach
Bartlett (1932) proposed the constructive
approach of Memory
Existing knowledge (schemas) are used to
understand new information. Memories are
changed to fit in with what we already know
We tend to see and in particular interpret
and recall what we see according to what we
expect and assume is 'normal' in a given
situation
Schemas
• These are mental structures that represent an aspect of
the world, such as an object or event.
• Schemas help us to make sense of the world, by
providing short cuts to identifying things that we
come across (our building blocks of knowledge).
Schemas
 For example,
 It has a large metal door
 Buttons and knobs
 Gets hot inside
 Has hot metal rings on top
 It’s probably a stove. You don’t need to have seen
this particular stove before to identify it. Your
schema for “stoves” allows you to be able to
identify all stoves so long as they don’t differ too
much from your mental schema.
Bartlett 1932 Study
Aim To investigate whether memories can be altered due
to own values/culture
Method
Participants read a story called ‘The war of the ghosts’.
Story based upon Native American culture
Participants asked to recall the story repeatedly over
days, weeks and months.
Results
Participants distorted the story when asked to recall it
Recall errors conformed to recaller’s culturally expectations
(e.g. discussion about the spirits, swapped word canoe for
boat). Rationalization.
Conclusion
Precise information is lost over time. Memory for schema
are not. More rationalization errors over time. Memory is
distorted to try and fit with a person’s current schemas (at
retrieval)
Anderson and Pichert (1978)
Participants read a description of a home from
one of the two following perspectives.
• Burglar Group
• Homebuyer Group
Recalled the story again either from the same
or the opposite.
Schemata and Memory
(Anderson & Pichert, 1978)
Proportion Recalled
Recall #1
Identity
Items
Schemata and Memory
(Anderson & Prichert, 1978)
Change in proportion recalled
Distracter task....then....switch perspectives and try to
recall again
First identity/second identity
Items
Figure 5.6 The “graduate
student’s” room used by
Brewer and Treyens (1981)
in their experiment.
Copyright © 1981 Elsevier.
Photo reproduced with kind
permission of Professor
Brewer.
Participants spent
35 in this room.
Given Recall and or Recognition Test
Both things that we expect to see and those
that, for whatever reason, we are surprised to
see (that stand out) tend to get noticed and
remembered.
Schemas effect memory.
GP Mem(1)
66