PSY 368 Human Memory

Memory Implicit memory

Outline

• Theories accounting for Implicit vs. Explicit memory

• Experiment 2 Signal detection analysis

• Process-dissociation procedure, working through our example (probably not time, so after break)

Memory Tasks

indirect direct

incidentalimplicit memoryexpts.

Levels of Processing

expts.

intentional? explicit

memoryexpts.

Test Instructions

Study Instruction

s

Implicit Memory: Often defined as "memory without awareness”

• Also “Non-declarative” & “procedural” (Squire, Knowlton, & Mesen, 1993)

Implicit/Explicit Dissociations

• Many demonstrations of different effects depending on whether implicit or explicit tasks are used• Amnesic patients• Levels of processing manipulations

• Pleasantness vs. vowel comparisons

• Generation effect• Divided attention• Picture-word superiority

• Note. Most of we’ve talked about concern repetition priming effect (study “horse” and respond “horse”)

• Four major approaches have been proposed• The Activation view• Multiple Memory systems view• Transfer appropriate processing view• Bias View

Accounting for Implicit/Explicit Dissociations

• Four major approaches have been proposed• The Activation view• Multiple Memory systems view• Transfer appropriate processing view• Bias View

Accounting for Implicit/Explicit Dissociations

The Activation View

• Priming on indirect tests is attributable to the temporary automatic activation of preexisting representations.• Because it is automatic, it occurs without elaborative

processing and thus has little to no contextual information

• Weak Point • Can not explain priming over long time periods

• Some implicit priming over days or even weeks (e.g., Sloman, et al, 1988)

• Can not explain priming without pre-existing representations

• The least popular of the four views

• Four major approaches have been proposed• The Activation view• Multiple Memory systems view• Transfer appropriate processing view• Bias View

Accounting for Implicit/Explicit Dissociations

Multiple Memory Systems

• Many dissociations between direct and indirect tests of memory arise because the tests tap different underlying memory systems.

Squire (1987)

Multiple Memory Systems

• Many dissociations between direct and indirect tests of memory arise because the tests tap different underlying memory systems.

Tulving (1984)

Multiple Memory Systems

• What is a system?• It is NOT a process• It is NOT a task

• Some different ways that systems have been defined

Schacter and Tulving (1994)

Multiple Memory Systems

• What is a system?• Functional Dissociations

• Task that taps into system A that has no effect (or a different effect) in System B

• Different neural substrates• System A involves different brain areas than System B

(brain damage cases and neural imaging studies)

• Stochastic independence• Performance on System A task uncorrelated with

performance on a System B task

• Functional incompatibility• Could involve different rates of forgetting

• Function carried out by System A can not be done by System B

Schacter and Tulving (1994)

Multiple Memory Systems

• What is a system?

Schacter and Tulving (1994)System Other Name Subsystems Characteristics

Procedural Nondeclarative Motor skills Non-conscious operation(indirect)

Cognitive skills

Simple conditioning

Simple associative learning

Perceptual representation

Nondeclarative Visual word form

Auditroy word form

Structural description

Primary memory

Working memory Visual Conscious operation(direct)

Auditory

Semantic Generic Spatial

Factual Relational

Knowledge

Episodic Personal

Autobiographical

Event memory

If you “know how to do something”

Allows you to automatically recognize things

See earlier in the semester

Factual information (chpt 10)Memory of events

• Brain areas• Brain imaging studies

found that different areas of the brain are used when completing implicit and explicit tasks

• Note: more than one structure involved in each type of memory

Buckner et al (1995) PET study

Multiple Memory Systems

• Brain areas• Brain imaging studies

found that different areas of the brain are used when completing implicit and explicit tasks

Gabrieli et al (1995) Case study of MS

MRI of MS’s brain

• Intact performance on explicit tests of recognition and cued recall

• Intact performance on implicit test of conceptual memory

• Impaired performance on implicit tests of visual perceptual memory• Suggests a specific deficit

in visual implicit memory

• Studied lists of words

• Perceptual identification and recognition task

Multiple Memory Systems

• Brain areas• Different kinds of implicit

tasks seem to involve different areas• Perceptual vs.

conceptual tasks appear to use different brain areas

• Conclusion: brain area involvement may be a function of type of processing and type of memory

Buckner & Koutstall (1998) fMRI study

Multiple Memory Systems

• Stochastic Independence• Hayman and Tulving (1989)

• Measure correlation between explicit and implicit task performance

• If not correlated (independent), then tasks measure different processes

Multiple Memory Systems

• Forgetting• Tulving et al. (1989) showed a difference in

forgetting rate for recognition and fragment completion

• Confirmed with other tasks (stem completion)

• Strengths• Fits well with dissociations found

• In patients• In experiments

Multiple Memory Systems

• Problem• Hard to find consensus on what the systems are• May be “too easy” to posit a new system

• Four major approaches have been proposed• The Activation view• Multiple Memory systems view• Transfer appropriate processing view• Bias View

Accounting for Implicit/Explicit Dissociations

• The key to good performance is similarity of processes involved in encoding vs. retrieval, be it implicit or explicit, perceptual or conceptual test

• Implicit and explicit may refer to different processes, yet the key to performance is matching processes.

Processes at encoding

Processes at test

Overlap determines retrieval success

Transfer Appropriate Process

A consequence: conceptual processing is the common core in free recall and implicit conceptual tasks, hence performance on these two types of task should be equal.

Transfer Appropriate Process

Assumes:• Performance depends of match between

processing at study and processing at test.Analogous to encoding specificity.

• Two-types of Processes (Jacoby, 1990)• Data-driven (perceptual) – processing of

physical features.• Conceptually-driven (semantic) – processing

for meaningTypically confounded, however, it is possible to un-confound test-type from process-type

• Jacoby (1990) proposed that implicit vs. explicit memory is confounded with two different kinds of memory processes (associated with two kinds of information)

Mixing Implicit and Explicit Effects

Memory system

Mode of Processing

Declarative(Episodic)

Procedural(Priming)

Perceptual(Data-driven)

Perceptual identification

Word Fragment

Completion

Meaning based(conceptually-driven)

Free RecallRecognition

Data-driven

(Perceptual):

• fragment completion

• stem completion

• anagram completion

• lexical decision

• perceptual identification

Conceptually-driven

(Semantic):

• word association

doctor ??

• category-instance generation“name a mammal”

• general knowledge“The capital of the US is …?”

Transfer Appropriate Process

• Goal to demonstrate• data-driven processing can affect direct tests• data-driven processing do not necessarily

affect indirect tests

Blaxton (1989)

Transfer Appropriate Process

Data-driven Conceptually-driven

Direct Graphic-cuedRecall

Free Recall

Indirect FragmentCompletion

GeneralKnowledge

Target word: bashful• graphic-cued recall: looks like “bushful”• free recall• frag completion: b_sh_u_• General knowledge: “Name one of the 7 dwarfs”

Blaxton (1989)

Data-driven Conceptually-driven

Direct Graphic-cuedRecall

Free Recall

Indirect FragmentCompletion

GeneralKnowledge

Transfer Appropriate Process

S’s saw or heard lists of words (key IV here)

Predictions• Systems view: modality match should affect only

indirect tests (if indirect tap separate system, then modality should affect them in the same way)• for both implicit tests: visual > auditory

• for both explicit test: visual = auditory

Blaxton (1989)

Transfer Appropriate Process

Data-driven Conceptually-driven

Direct Graphic-cuedRecall

Free Recall

Indirect FragmentCompletion

GeneralKnowledge

Same pattern of results regardless of modality

Visual better than auditory for both

Predictions• TAP View: modality match should affect data-

driven tasks only. (priming depends on match between study/test processing match & not on indirect vs direct): • for both data-driven tests: visual > auditory• for both conceptually-driven tests: visual = auditory

Blaxton (1989)

Transfer Appropriate Process

Data-driven Conceptually-driven

Direct Graphic-cuedRecall

Free Recall

Indirect FragmentCompletion

GeneralKnowledge

Visual should be better than auditory

Visual and auditory should be about the same

ResultsPriming Effect (V > A) for

data-driven tasks only:

• indirect: frag completion

• direct: graphemic-cued recall

Not all indirect tests display priming effect.

• Gen Know (indirect, conceptual): V = A

Blaxton (1989)

Transfer Appropriate Process

ConclusionsSupport view that

processing rather than system is what is important

• Four major approaches have been proposed• The Activation view• Multiple Memory systems view• Transfer appropriate processing view• Bias View

Accounting for Implicit/Explicit Dissociations

The Bias View

• Proposed to account for repetition priming effects. • Prior presentation of an item can bias subsequent

processing of the item on later presentations (if you see it once, you are more likely to interpret in the same way later)• Multiple systems attributes this to 3 separate systems, but

doesn’t really offer an explanation

• TAP’s answer is considered circular (you respond faster the second time because of transfer appropriate processing, which was developed to account for priming effects)

The Bias View

• Bias View’s account for repetition priming effects.

1. First See one of old woman and young woman

2. Second See ambiguous woman

3. People are more likely to interpret the ambiguous picture as the same person as the unambiguous picture

1. First See Old Woman Young Woman

2. Second See Ambiguous -> Old Woman

Ambiguous -> Young Woman

• Bias entails both cost and benefits• Cost : There will be an advantage if prior

processing is appropriate for the current task• Benefits : There will be a disadvantage if prior

processing is inappropriate for the current task

Comparing the theories

• Strengths• Processing perspective• No “need” for separate

systems (true of Bias view too)

• Bias View may be seen as a complement to the TAP view

• Weaknesses• Doesn’t explain impact of

conscious awareness• Trouble with finer grain

distinctions between tasks

TAP Multiple Systmes• Strengths

• Good fit for deficit data (but may be too easy to propose “new systems”)

• Weaknesses• Has troubles with some

data showing differential decline in memory performance with aging

• Sometimes difficult to make specific predictions in advance

Implicit Memory Summary

• Implicit memory is memory without awareness.• Implicit and explicit tasks are not “process pure”• PDP offers a measurement method for processes

• Implicit memory is different memory from explicit memory by experimental dissociations.

• There is 4 main accounts for implicit memory• Probably a complex relationship between systems and

processes

Experiment 2

• Recall that for experiment 2 you each collected data from three participants.

• IV – levels

• Prediction: our instructions would lead participants to shift their criterion for what counts as old vs. new.

• Signal detection analysis

Signal Detection Theory

• Recognition accuracy depends on:• Whether a signal (noise/target memory) was

actually presented• The participant’s response

• Thus, there are four possible outcomes:

• Hits• Correctly reporting the presence of the signal

• Correct Rejections• Correctly reporting the absence of the signal

• False Alarms• Incorrectly reporting presence of the signal

when it did not occur• Misses

• Failing to report the presence of the signal when it occurred

CORRECT

INCORRECT

Signal Detection Theory

• Calculating d’ and C (or β)• Discriminability (d’):

• Step 1) Look up the z-score for the average Hit and False Alarm rates.

• Step 2) Apply the formula d’ = zHIT – zFA, where zFA is the z-score for FAs and zHIT is the z-score for Hits.

• Criteria C (or β):• Take the negative of the average of

zHIT and zFA. This is the criterion value C. • Remember that positive C values

indicate a conservative response bias, while negative C values indicate a liberal response bias.

http://memory.psych.mun.ca/models/dprime/

Experiment 2Target Lure

“Old” Hit15.050.75

False Alarm2.480.12

Neutral

Target Lure

“Old” Hit12.050.60

False Alarm1.140.06

Target Lure

“Old” Hit16.950.85

False Alarm4.380.22

Conservative

Liberal

N=21 per condition

Total possible hits or false alarms = 20AveragesProportions (avg/20)

Experiment 2Target Lure

“Old” Hit15.050.75

False Alarm2.480.12

Neutral

Target Lure

“Old” Hit12.050.60

False Alarm1.140.06

Target Lure

“Old” Hit16.950.85

False Alarm4.380.22

Conservative

Liberal

d’ = 1.85C = 0.25

d’ = 1.81C = 0.65

d’ = 1.81C = -0.13

http://memory.psych.mun.ca/models/dprime/

Experiment 2

Neutral

Conservative

Liberal

d’ = 1.85C = 0.25

d’ = 1.81C = 0.65

d’ = 1.81C = -0.13

stimulus intensity

pro

babili

ty

Noise

Signal(remember)

d’

Experiment 2

Neutral

Conservative

Liberal

d’ = 1.85C = 0.25

d’ = 1.81C = 0.65

d’ = 1.81C = -0.13

stimulus intensity

pro

babili

ty

Noise

Signal(remember)

OldNew

- Criterion +

Experiment 2

Neutrald’ = 1.85C = 0.25

stimulus intensity

pro

babili

ty

Noise

Signal(remember)

OldNew

- Criterion +

Experiment 2

Conservatived’ = 1.81C = 0.65

stimulus intensity

pro

babili

ty

Noise

Signal(remember)

OldNew

- Criterion +

Experiment 2

Liberal

d’ = 1.81C = -0.13

stimulus intensity

pro

babili

ty

Noise

Signal(remember)

OldNew

- Criterion +