Effects of value on rule-based and information-integration category learning across the lifespan

Veronica X. Yan1, Sharon M. Noh2, Tyson Kerr1, Alan D. Castel1, & W. Todd Maddox2

1University of California, Los Angeles; 2The University of Texas at Austin

INTRODUCTION

RESEARCH QUESTIONS CONCLUSION Value attenuates age-related deficits

for rule-based learning

RESULTS

DISCUSSION

Value Effects on Memory

1.  Do value and aging differentially affect rule-based and information-integration category learning?

2.  Can value attenuate age-related deficits in

category-learning?

DESIGN & MATERIALS

Acknowledgments:

This research was funded by the National

Institute of Aging through grant No. AG043425 awarded to W. Todd Maddox and through grant

No. R01AG044335 awarded to Alan Castel

Contact: Sharon Noh smnoh@utexas.edu

Veronica Yan veronicayan@ucla.edu

QR Code:

METHODS Seventy-eight participants recruited from Amazon Mechanical Turk (51 younger adults, 18-34 years; 27 older adults, 60-78 years)

“Imagine you are training to work in a pharmacy, and your job is to learn to categorize various pills”

Participants instructed that they would study both medications and supplements (in reality, they participants studied either medications or supplements), but that “it is especially important to learn the medications accurately, as there may be severe consequences of mixing them up!”. Study: 8 blocks of feedback training (8 exemplars x 4 categories per block), with aggregate feedback after each block Test: Classify 64 new exemplars, without feedback

2x2 between-subjects Value: Low (supplements) vs. High (medications) Category structure: Rule-based vs. Information-integration •  Four categories •  2 relevant

dimensions

.0

.1

.2

.3

.4

.5

.6

.7

.8

.9

1 2 3 4 5 6 7 8 9 10 11 12

Point Value of Word

Prob

abilit

y of

Rec

all Young

Old

Category Learning Across the Lifespan Dual category learning systems (Ashby & Maddox, 2011): •  Two competing, neurobiologically-grounded learning

systems •  Rule-based (RB): Verbalizable, hypothesis-testing,

frontally-mediated •  Information-integration (II): Non-verbalizable

stimulus-response mappings, striatally-mediated Age-related deficits found in both tasks (Maddox, Pacheco, Reeves, Zhu, & Schnyer, 2010), although evidence is relatively mixed for the II task (Filoteo & Maddox, 2004).

Older adults can selectively attend to high-value information, compensating for age-related deficits in memory (Castel, 2008)

Ashby, F.G. & Maddox, W.T. (2011). Human category learning 2.0. Annals of the New York Academy of Sciences, 1224, 147-161. Castel, A. D. (2008). The adaptive and strategic use of memory by older adults: Evaluative processing and value-directed remembering. In A. S. Benjamin & B. H. Ross (Eds.), The psychology of learning and motivation (Vol. 48, pp. 225-270). London: Academic Press. Filoteo, J.V. & Maddox, W.T. (2004). A quantitative model-based approach to examining aging effects on information-integration category learning. Psychology & Aging, 19, 171-182. Maddox, W.T., Pacheco, J., Reeves, M., Zhu, B., & Schyner, D.M. (2010). Rule-based and information-integration category learning in normal aging. Neuropsychologia, 48, 2998-3008.

Optimal Strategy Use

Test Performance RESULTS

EXAMPLE STRATEGIES

REFERENCES

•  Value enhances use of conjunctive rules (greater “effort”?) •  Optimal for rule-based category

learning •  Sub-optimal for information-integration

category learning •  As a result of greater rule-based strategy

use, high value attenuates age-related deficits in rule-based learning, but not in information-integration learning

DISCUSSION

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Younger Older Younger Older

Rule-based Information-integration

Prop

ortio

n C

orre

ctly

C

lass

ified

Low Value

High Value

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Younger Older Younger Older

Rule-based Information-integration

Prop

ortio

n O

ptim

al

Stra

tegy

Use

Low Value

High Value

Conjunctive rules (optimal for RB)

Information-integration (optimal for II)

Unidimensional rules

length

Noi

se/g

rain

length

Noi

se/g

rain

length

Noi

se/g

rain

length

Noi

se/g

rain