cardinal number representation in rhesus...
TRANSCRIPT
Kerry E. Jordan and Elizabeth M. BrannonCenter for Cognitive Neuroscience and Department of Psychological and Brain Sciences, Duke University
DMTS Method
Conclusions
Acknowledgements
Cardinal Number Representation in Rhesus Macaques
Thanks to Jessica Cantlon, Dr. Kerrie Lewis, Pierre Rojas, and Jessica Ward for helping with data collection. This project was funded by the John Merck Scholars Program to EMB. Correspondence regarding this poster should be sent to [email protected].
Sample
Subjects and general apparatus:
· 3 female rhesus macaques (Mikulski, Schroeder, and Feinstein)· Tested in primate chairs inside sound-isolated booths · Responded via a touch screen for juice reinforcement
Test
Experiment 1: Monkeys match stimuli based on numerical equivalence
Experiment 2: Number bisection task
Subjects: Mikulski, Schroeder, FeinsteinStimuli: 2 sets, approximately 3000 trials/setSample: numerosities between 1-9Test stimuli: 2 and 8Reinforcement: 70% trials reinforced with sample numbers 2 & 8 15% trials unreinforced with sample numbers 2 & 8 15% trials unreinforced with novel sample numbers 1,3,4,5,6,7,9
Rhesus macaques1) match stimuli based on numerical equivalence, regardless of continuous dimensions2) interpolate novel values along an ordinal continuum3) are sensitive to ratio when making more precise numerical matches
Surface area and element size control
Subjects: Schroeder and FeinsteinSample: numerosities between 1-9 Test stimuli: numerosities between 1-9Full reinforcement
Non-Identity
Introduction
This work addresses the nature of cardinal number representation in nonhuman primates. While the abstract nature of such representations in adult humans is well-documented, less is knownabout the possible evolutionary origins of such central cognitive capacities. This series of delayed match-to-sample (DMTS)studies specifically investigates the abilities of rhesus monkeys to match stimuli based on number.
Stimulus Examples
Element configuration and color differed between the sample and match, unlikeidentical stimuli as illustrated in previous column.
Correct numerical matches wererewarded with a green borderflash and fruit juice. Incorrectnumerical choices resulted in the disappearance of the stimuli and a 5s timeout.
Density control
The density of the match was sometimes greater than, sometimes less than, and sometimes equal to that of the sample.
Performance minimally affected by stimulus controls
Methods
The point at which monkeys are equally likely to classify a numerosity as 2 or 8 is the point of subjective equality, or PSE.
Methods
PSE Weber fractionStim set 1 4.23 0.955Stim set 2 3.45 0.579
Sessions to criterion
Total surface area of elements and individual element size were varied such that the total surface area or element size of the match was sometimes greater than,sometimes equal to, and in some trials less than that of the sample.
Sample was an exemplar of 2 or 8Test stimuli were one exemplar of 2 and one of 8Trained with each stimuli control set to a criterion of 80% accuracy
Probability of choosing 8 systematically increases with number
Experiment 3: Choice Requires Precise Numerical Match
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 2 3 4 5 6 7 8 9Sample Value
Prob
abili
ty o
f cho
osin
g 8
Unreinforced Set 1Unreinforced Set 2Unreinforced AverageReinforced Average
PSE
Accuracy decreases as match:distracter ratio increases
550
650
750
850
0 0.2 0.4 0.6 0.8 1
Match:Distracter Ratio
Rea
ctio
n ti
me
(ms)
40
50
60
70
80
90
100
0 0.2 0.4 0.6 0.8 1
Match:Distracter Ratio
Per
cen
t co
rrec
t
Reaction time increases as match:distracter ratio increases
200
250
300
350
400
450
1 2 3 4 5 6 7 8 9
Latency to touch sample does not vary with sample number
0
10
20
30
40
50
60
70
80
90
100
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59
Perc
ent C
orre
ct
Mikulski (every session)
Schroeder (every 5sessions)Feinstein (every session)
Picture Number Identity
Number Non-Identity
Surface Area and Element Size Control
Density Control
Chance
Criterion
0
5
10
15
20
25
30
35
Num
ber s
essi
ons
Mikulski
Schroeder
Feinstein
Picture Number identity
Density control
Surface area and element size control
Number non-
Identity
* outlier = 100
minimum sessionsrequired
Rea
ctio
n ti
me
(ms)
Sample Number