Exploring the Structure of Recognition Memory Jeffrey N. Rouder - - PowerPoint PPT Presentation

Exploring the Structure of Recognition Memory

Jeffrey N. Rouder January, 2012

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Multiple Memory Systems

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Multiple Memory Systems

◮ Memory is usually defined broadly, ”systematic change in

response to an event in the environment.”

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Multiple Memory Systems

◮ Memory is usually defined broadly, ”systematic change in

response to an event in the environment.”

◮ Mnemonic consequences of a violent encounter:

◮ Recollection ◮ Change in Attitudes ◮ Change in Physiological Response ◮ Neural correlates of bruising Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Multiple vs. Single System Memory Debate

But what happens if we restrict the domain:

◮ Are there well-circumscribed domains that admit a single

memory mechanism?:

◮ LTP in simple organisms ◮ Perhaps the neural correlates of bruising in a violent encounter. ◮ Perhaps memory for verbal information as lay people

understand it

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Leading Multiple Memory System Account

Two Distinct Process or Systems:

◮ Unconscious, Familiarity Process ◮ Conscious, Recollective Process

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Common Alternative

Memory Strength Theory:

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Common Alternative

Memory Strength Theory:

◮ Each item has an associated latent strength.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Common Alternative

Memory Strength Theory:

◮ Each item has an associated latent strength. ◮ Strengths are single-dimensional quantities (scalars)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Common Alternative

Memory Strength Theory:

◮ Each item has an associated latent strength. ◮ Strengths are single-dimensional quantities (scalars) ◮ Studying increases strength

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Signal-Detection Representation of Strength Theories

−2 2 4 6 0.0 0.1 0.2 0.3 0.4 Latent Strength Density New Item Old Item

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Insight

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Insight

◮ Perhaps both of these models are too complex for behavioral

paradigms

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Insight

◮ Perhaps both of these models are too complex for behavioral

paradigms

◮ Two discrete states: remember or guess

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Insight

◮ Perhaps both of these models are too complex for behavioral

paradigms

◮ Two discrete states: remember or guess ◮ Study increases the probability of remembering.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis

◮ Receiver operating characteristic (ROC) curves are a lossless

way of highlighting structure in data. Primer to follow.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis

◮ Receiver operating characteristic (ROC) curves are a lossless

way of highlighting structure in data. Primer to follow.

◮ Most researchers I know endorse the proposition that the

shape of ROC curves provide evidence for or against specified models of processing.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis

◮ Receiver operating characteristic (ROC) curves are a lossless

way of highlighting structure in data. Primer to follow.

◮ Most researchers I know endorse the proposition that the

shape of ROC curves provide evidence for or against specified models of processing.

◮ I do not. Any one ROC curve is noninformative.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis

◮ Receiver operating characteristic (ROC) curves are a lossless

way of highlighting structure in data. Primer to follow.

◮ Most researchers I know endorse the proposition that the

shape of ROC curves provide evidence for or against specified models of processing.

◮ I do not. Any one ROC curve is noninformative. ◮ Conventional interpretation of ROC curves relies on untenable

parametric assumptions.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis

◮ Receiver operating characteristic (ROC) curves are a lossless

way of highlighting structure in data. Primer to follow.

◮ Most researchers I know endorse the proposition that the

shape of ROC curves provide evidence for or against specified models of processing.

◮ I do not. Any one ROC curve is noninformative. ◮ Conventional interpretation of ROC curves relies on untenable

parametric assumptions.

◮ Talk will be done for the confidence-ratings task, main

conclusions hold without qualification for two-choice paradigms such as old/new recognition memory or yes-no perceptual detection.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Confidence Ratings Recognition Memory Task

◮ STUDY: See a bunch of words ◮ TEST: Present new and old words, one at a time ◮ TASK: Rate each word

• 1. Sure New
• 2. Believe New
• 3. Guess New
• 4. Guess Old
• 5. Believe Old
• 6. Sure Old

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis Primer

1 2 3 4 5 6 Sure Believe Guess Guess Believe Sure New New New Old Old Old

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis Primer

1 2 3 4 5 6 Sure Believe Guess Guess Believe Sure New New New Old Old Old Old Items .14 .10 .12 .14 .14 .36

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis Primer

1 2 3 4 5 6 Sure Believe Guess Guess Believe Sure New New New Old Old Old Old Items .14 .10 .12 .14 .14 .36 New Items .31 .19 .19 .15 .09 .07

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis Primer

1 2 3 4 5 6 Sure Believe Guess Guess Believe Sure New New New Old Old Old Old Items .14 .10 .12 .14 .14 .36 New Items .31 .19 .19 .15 .09 .07 Hit Rate 1 .86 .76 .64 .50 .36

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

ROC Analysis Primer

1 2 3 4 5 6 Sure Believe Guess Guess Believe Sure New New New Old Old Old Old Items .14 .10 .12 .14 .14 .36 New Items .31 .19 .19 .15 .09 .07 Hit Rate 1 .86 .76 .64 .50 .36 False-Alarm Rate 1 .69 .50 .31 .16 .07

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Recognition-Memory Basics: ROC Curves

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate

• Benchmark Phenomena:

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Recognition-Memory Basics: ROC Curves

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate

• Benchmark Phenomena:
• 1. Asymmetry relative

to negative diagonal

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Recognition-Memory Basics: ROC Curves

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate

• Benchmark Phenomena:
• 1. Asymmetry relative

to negative diagonal

• 2. Not a straight line

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

A Signal Detection Account

−2 2 4 6 0.0 0.1 0.2 0.3 0.4 Latent Strength Density New Item Old Item

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

A Signal Detection Account

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate

• Benchmark Phenomena:
• 1. Asymmetry relative

to negative diagonal

• 2. Not a straight line

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Dual-Process Account

◮ Recollection: Yes or No Process (All-or-None)

◮ Yes: Respond Sure Old ◮ No: Use Familiarity Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Dual-Process Account

◮ Recollection: Yes or No Process (All-or-None)

◮ Yes: Respond Sure Old ◮ No: Use Familiarity

◮ Familiarity:

Signal Detection with Equal Variance

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Dual-Process Account

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate

• Benchmark Phenomena:
• 1. Asymmetry relative

to negative diagonal

• 2. Not a straight line

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

PART I: Axiomatic Definition of Single Process

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Core of Latent-Strength Theory

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Core of Latent-Strength Theory

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ SIDE ASSUMPTION: Parametric form of latent strength

(e.g., normal).

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The Core of Latent-Strength Theory

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ SIDE ASSUMPTION: Parametric form of latent strength

(e.g., normal).

◮ These side assumptions have no psychological content. Made

for convenience.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Perfectly Good Latent-Strength Models

−2 2 4 6 0.0 0.1 0.2 0.3 0.4 Latent Strength Density

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Perfectly Good Latent-Strength Models

−2 2 4 6 0.0 0.1 0.2 0.3 0.4 Latent Strength Density

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Perfectly Good Latent-Strength Models

−2 2 4 6 0.0 0.1 0.2 0.3 0.4 Latent Strength Density

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Perfectly Good Latent-Strength Models

5 10 15 0.0 0.1 0.2 0.3 Latent Strength Density

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Perfectly Good Latent-Strength Models

5 10 15 0.0 0.2 0.4 0.6 Latent Strength Density

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Latent Strength Is Too Flexible

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Latent Strength Is Too Flexible

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ THEOREM: The core is unfalsifiable.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Latent Strength Is Too Flexible

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ THEOREM: The core is unfalsifiable.

◮ For any family of ROC curves, there exists some set of latent

strength distributions N, S1, S2, . . . that can perfectly predict the family.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Latent Strength Is Too Flexible

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ THEOREM: The core is unfalsifiable.

◮ For any family of ROC curves, there exists some set of latent

strength distributions N, S1, S2, . . . that can perfectly predict the family.

◮ CONSEQUENCE: Goodness of fit tests are tests of the

contentless side assumptions rather than the core.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Latent Strength Is Too Flexible

◮ CORE: Decisions are made by placing criteria on a latent

strength axis.

◮ THEOREM: The core is unfalsifiable.

◮ For any family of ROC curves, there exists some set of latent

strength distributions N, S1, S2, . . . that can perfectly predict the family.

◮ CONSEQUENCE: Goodness of fit tests are tests of the

contentless side assumptions rather than the core.

◮ MY VIEW: The debate in the ROC literature on multiple vs.

single process is misguided because in every case it is critically dependent on contentless parametric assumptions.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Example of Flexibility

Latent Mnemonic Strength 0.0 0.5 1.0 1.5 2.0 2.5 3.0

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Example of Flexibility

Latent Mnemonic Strength 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Fam Recollection

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

NEEDED: New conceptualization of single process.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

NEEDED: New conceptualization of single process.

◮ More constrained than the core of signal detection

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

NEEDED: New conceptualization of single process.

◮ More constrained than the core of signal detection ◮ Less arbitrary than the side parametric assumptions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

NEEDED: New conceptualization of single process.

◮ More constrained than the core of signal detection ◮ Less arbitrary than the side parametric assumptions ◮ Behavior of a family of ROC curves rather than any one curve.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Approach

SINGLE PROCESS IMPLIES:

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Approach

SINGLE PROCESS IMPLIES:

◮ All ROC curves differ from one another in just one way.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

My Approach

SINGLE PROCESS IMPLIES:

◮ All ROC curves differ from one another in just one way. ◮ ROC curve that differ in just one way are said to be Single

Process Representable (SPR)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Examples of Single-Process ROC Families

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate

Normal- Distribution Signal Detection

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Examples of Single-Process ROC Families

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate

High Threshold Model

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Examples of Single-Process ROC Families

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate

Gamma- Distribution Signal Detection

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Not Single Process Representable

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate

Yonelinas Dual Process Model

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Not Single Process Representable

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate

Unequal-Variance Normal- Distribution Signal Detection (UVSD)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

One Formalization

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate −6 −4 −2 2 4 −6 −4 −2 2 4 Φ−1(f) Φ−1(h)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

One Formalization

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate −4 −3 −2 −1 −4 −3 −2 −1 log(1 − f) log(1 − h)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Linear Single Process Representability

Linear Single Process Representability:

◮ Does there exist a linearizing transform? ◮ Do all the transformed lines have the same slope?

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Linear Single Process Representability

◮ Let ωi be the function that describes the ROC curve for the

ith condition; e.g., h = ωi(f ).

◮ Let Ω be the collection of all such curves under consideration. ◮ Collection Ω is linearly single-process representable (lSPR) if

there exists a strictly increasing function function θ such that for any ωi ∈ Ω θ(ωi(f )) = θ(f ) + ci, 0 ≤ f ≤ 1, where ci is constant across f

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

SPR Is General

ROCs:

◮ Working Memory ◮ Auditory/Visual Perception ◮ Speech ◮ Tumor Detection

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

SPR Across Domains

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate Tone Detection Recognition Memory 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False−Alarm Rate Hit Rate Tone Detection Recognition Memory

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

SPR Across Domains

◮ How can tone recognition rely on the same single process as

recognition memory?

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

SPR Across Domains

◮ How can tone recognition rely on the same single process as

recognition memory?

◮ Abstracted information, common decision system

(much like a statistician who uses a t-test across multiple domains)

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

SPR Across Domains

◮ How can tone recognition rely on the same single process as

recognition memory?

◮ Abstracted information, common decision system

(much like a statistician who uses a t-test across multiple domains)

◮ SPECULATION: ROCs tell us more about common decision

processing than mnemonic or other upstream processing.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing Linear SPR

How does one know if a suitable linearizing function θ exists?

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR

◮ Functional Approximation:

θ(x) ≈ g(x, β) + β0

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR

◮ Functional Approximation:

θ(x) ≈ g(x, β) + β0

◮ Check whether

g(ωi(f ), β) = g(f , β) + ci for all ωi ∈ Ω.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR In Practice

g(x, β1, β2) = Φ−1

• 1 − 2

3β1 − 1 3β2

• x+

β1

• x − 1

3

• I(x − 1/3) + β2
• x − 2

3

• I(x − 2/3)
• .

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR In Practice

g(x, β1, β2) = Φ−1

• 1 − 2

3β1 − 1 3β2

• x+

β1

• x − 1

3

• I(x − 1/3) + β2
• x − 2

3

• I(x − 2/3)
• .

◮ Two-piece linear spline on z-ROC function.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR In Practice

g(x, β1, β2) = Φ−1

• 1 − 2

3β1 − 1 3β2

• x+

β1

• x − 1

3

• I(x − 1/3) + β2
• x − 2

3

• I(x − 2/3)
• .

◮ Two-piece linear spline on z-ROC function. ◮ g = Φ−1 if β1 = β2 = 0

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR In Practice

g(x, β1, β2) = Φ−1

• 1 − 2

3β1 − 1 3β2

• x+

β1

• x − 1

3

• I(x − 1/3) + β2
• x − 2

3

• I(x − 2/3)
• .

◮ Two-piece linear spline on z-ROC function. ◮ g = Φ−1 if β1 = β2 = 0 ◮ Estimate separate parameters (β1i, β2i) for each ωi.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing SPR In Practice

g(x, β1, β2) = Φ−1

• 1 − 2

3β1 − 1 3β2

• x+

β1

• x − 1

3

• I(x − 1/3) + β2
• x − 2

3

• I(x − 2/3)
• .

◮ Two-piece linear spline on z-ROC function. ◮ g = Φ−1 if β1 = β2 = 0 ◮ Estimate separate parameters (β1i, β2i) for each ωi. ◮ SPR holds if β1 and β2 do not vary across curves, i.e.,

β1 = β1i and β2 = β2i for all i.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Conclusions

• 1. The core of single-process signal-detection models is too

general.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Conclusions

• 1. The core of single-process signal-detection models is too

general.

• 2. Single-process representability (SPR) implies that ROCs differ

in one way. May be formally defined and tested (Linear SPR).

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Conclusions

• 1. The core of single-process signal-detection models is too

general.

• 2. Single-process representability (SPR) implies that ROCs differ

in one way. May be formally defined and tested (Linear SPR).

• 3. SPR seems especially well-suited to explore behavior across

several domains.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

PART II: And Then We Began Collecting Recognition Memory and Perception Data

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Consider a Simple Discrete-State Model

◮ Three States: Detect Old Item, Detect New Item, Guess ◮ Conditional Independence: Responses are determined only by

which state one is in.

◮ No false detection. Participant can enter one of two states on

any trial (correct detect state, guess state).

◮ Certainty: Detection leads to a “certainty” response.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete States With Certainty Assumption

Old Item Presented

θ1 θ2 θ3 θ4 θ5 θ6 6 1 2 3 4 5 6 ds 1 − ds

∑

i=1 6

θi = 1

New Item Presented

θ1 θ2 θ3 θ4 θ5 θ6 1 1 2 3 4 5 6 dn 1 − dn

∑

i=1 6

θi = 1

0.0 0.5 1.0 0.0 0.5 1.0

Hit Rate False Alarm Rate

• Jeffrey N. Rouder

Exploring the Structure of Recognition Memory

Discrete States W/O Certainty Assumption

Old Item Presented

φ6 φ5 φ4 θ1 θ2 θ3 θ4 θ5 θ6 6 5 4 1 2 3 4 5 6 ds 1 − ds

∑

i=1 6

θi = 1

∑

i=4 6

φi = 1

New Item Presented

φ1 φ2 φ3 θ1 θ2 θ3 θ4 θ5 θ6 1 2 3 1 2 3 4 5 6 dn 1 − dn

∑

i=1 6

θi = 1

∑

i=1 3

φi = 1

0.0 0.5 1.0 0.0 0.5 1.0

Hit Rate False Alarm Rate

• Jeffrey N. Rouder

Exploring the Structure of Recognition Memory

Testing Discrete State Models

◮ W/O Certainty Assumption, Discrete-State Models predict

that the ROC points are connected by straight lines.

◮ Can account for any single ROC curve, no constraint. ◮ Constraint across the comparison of several ROC curves

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Testing Discrete State Models

◮ i. Show you the paradigm ◮ ii. Then the predictions for discrete-state and typical

latent-strength models

◮ iii. Finally, the data

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Paradigm

◮ Study A List of Words (e.g., “FROG”) ◮ Test: 2AFC, Move Slider

TABLE FROG

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Paradigm

STUDY:

◮ 1 repetition ◮ 2 repetitions ◮ 4 repetitions

TEST:

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Paradigm

STUDY:

◮ 1 repetition ◮ 2 repetitions ◮ 4 repetitions

TEST:

◮ Normal: Old Word + New Word

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Paradigm

STUDY:

◮ 1 repetition ◮ 2 repetitions ◮ 4 repetitions

TEST:

◮ Normal: Old Word + New Word ◮ Sneaky: Two New Words

(0 repetitions).

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Paradigm

STUDY:

◮ 1 repetition ◮ 2 repetitions ◮ 4 repetitions

TEST:

◮ Normal: Old Word + New Word ◮ Sneaky: Two New Words

(0 repetitions).

◮ These sneaky trials allow us to isolate responses under

guessing.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Sure Left Sure Right Word Was Word Was Studied Studied

Latent Strength Theory

No Repetitions Sure Left Sure Right Word Was Word Was Studied Studied

Discrete State Theory

No Repetitions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Sure Left Sure Right Word Was Word Was Studied Studied

Latent Strength Theory

Few Repetitions Sure Left Sure Right Word Was Word Was Studied Studied

Discrete State Theory

Few Repetitions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Sure Left Sure Right Word Was Word Was Studied Studied

Latent Strength Theory

Many Repetitions Sure Left Sure Right Word Was Word Was Studied Studied

Discrete State Theory

Many Repetitions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Sure Left Sure Right Word Was Word Was Studied Studied

Latent Strength Theory

None Few Many Sure Left Sure Right Word Was Word Was Studied Studied

Discrete State Theory

None Few Many

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Latent Strength Theory

Sure "New" Sure "Old" Word Was Word Was Studied Studied No Repetitions Few Repetitions Many Repetitions

Discrete State Theory

Sure "New" Sure "Old" Word Was Word Was Studied Studied No Repetitions Few Repetitions Many Repetitions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Differing Predictions

Latent Strength Theory

Sure "New" Sure "Old" Word Was Word Was Studied Studied No Repetitions Few Repetitions Many Repetitions

Discrete State Theory

Sure "New" Sure "Old" Word Was Word Was Studied Studied No Repetitions Few Repetitions Many Repetitions

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Data: Select Subjects

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

Probability

Study Condition 0 Repetitions 1−2 Repetitions 4 Repetitions "Lure was on List" (Incorrect Response) "Target was on List" (Correct Response) [High Confidence] [High Confidence] [Low Confidence] [Low Confidence]

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Data: Select Subjects

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

Probability

Study Condition 0 Repetitions 1−2 Repetitions 4 Repetitions "Lure was on List" (Incorrect Response) "Target was on List" (Correct Response) [High Confidence] [High Confidence] [Low Confidence] [Low Confidence]

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Data: Select Subjects

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

Probability

Study Condition 0 Repetitions 1−2 Repetitions 4 Repetitions "Lure was on List" (Incorrect Response) "Target was on List" (Correct Response) [High Confidence] [High Confidence] [Low Confidence] [Low Confidence]

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Data: Select Subjects

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

Probability

Study Condition 0 Repetitions 1−2 Repetitions 4 Repetitions "Lure was on List" (Incorrect Response) "Target was on List" (Correct Response) [High Confidence] [High Confidence] [Low Confidence] [Low Confidence]

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Data: Select Subjects

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

Probability

Study Condition 0 Repetitions 1−2 Repetitions 4 Repetitions "Lure was on List" (Incorrect Response) "Target was on List" (Correct Response) [High Confidence] [High Confidence] [Low Confidence] [Low Confidence]

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Experiment 1

A B C

Experiment 2

D E F

Experiment 3

G H I Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Assessment: Absolute Fit vs. Vacuous Multinomial

◮ Compute G 2, log of likelihood ratio test statistic of

discrete-state model vs. a generic or vacuous multinomial for each participant

◮ 8 of 89 total participants (8%) had G 2 bigger than .05 cutoff.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Assessment: Relative Goodness-Of-Fit

−20 −10 10 20 30 40

Deviance Comparison

(Latent Strength − Discrete State)

• C

B A

• F

E D

• I

H G

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Assessment: Stability of Conditional Response Times

Experiment 2 Experiment 3 1 2 3 4

Number of Repetitions

1 2 3 2.5 3 3.5 4 4.5

Mean RT (seconds)

2 2.5 3 3.5 4

Conditional on Guessing Marginal Average Conditional on Detection Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Assessment: Stability of Conditional Response Times

Experiment 2 Experiment 3 1 2 3 4

Number of Repetitions

1 2 3 2.5 3 3.5 4 4.5

Mean RT (seconds)

2 2.5 3 3.5 4

Conditional on Guessing Marginal Average Conditional on Detection Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Our ROCs are Curved

0.2 0.4 0.6 0.8 1

Pr(Right | Left)

0.2 0.4 0.6 0.8 1

Pr(Right | Right)

A

• Experiment 1

Experiment 2 Experiment 3

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 1 2 3 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3

1 2 3 1 2 3 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Discrete-State Constraint in ROCs

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 False Alarm Rate Hit Rate 1 2 3

1 2 3 1 2 3 1 2 3 1 2 3 Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

◮ Shape of ROC curves may be less informative that the

relations between curves.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

◮ Shape of ROC curves may be less informative that the

relations between curves.

◮ Linear SPR and mixture property are attempts to formalize

general, nonparametric relations across ROC curves.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

◮ Shape of ROC curves may be less informative that the

relations between curves.

◮ Linear SPR and mixture property are attempts to formalize

general, nonparametric relations across ROC curves.

◮ A simple “detect/guess” discrete state model carries a lot of

water.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

◮ Shape of ROC curves may be less informative that the

relations between curves.

◮ Linear SPR and mixture property are attempts to formalize

general, nonparametric relations across ROC curves.

◮ A simple “detect/guess” discrete state model carries a lot of

water.

◮ ROC data across many domains may not be any more

complex or varied than that predicted by a simple “detect/guess” model.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

Take Home Points

◮ Shape of ROC curves may be less informative that the

relations between curves.

◮ Linear SPR and mixture property are attempts to formalize

general, nonparametric relations across ROC curves.

◮ A simple “detect/guess” discrete state model carries a lot of

water.

◮ ROC data across many domains may not be any more

complex or varied than that predicted by a simple “detect/guess” model.

◮ This Year: A lot of signal detection experiments across several

domains to see if these properties hold.

Jeffrey N. Rouder Exploring the Structure of Recognition Memory

The End

pcl.missouri.edu

Jeffrey N. Rouder Exploring the Structure of Recognition Memory