effects of value on rule-based and information … · effects of value on rule-based and ......
TRANSCRIPT
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 [email protected]
Veronica Yan [email protected]
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