|
Examining the perception and cognition of primates
often requires the use of non-verbal paradigms such as Match-to-Sample
(MTS) which utilize the subjectís natural categorization
of the stimuli that constitute their world. Advancing our knowledge
of the mental organization and discrimination of stimuli in non-human
primates will help elucidate their capacity for abstract thought.
Five adult brown capuchin monkeys (Cebus apella) were trained to
distinguish high contrast two-dimensional stimuli in the MTS task.
Upon attaining an established criterion of 3 consecutive sessions
at 84% accuracy the subjects were tested for transfer to comparable
and then increasingly complex and less-discriminable stimuli. The
subjects demonstrated a rapid acquisition of the MTS task; all reached
criterion within 9 sessions. Furthermore performance was not affected
by the transfer to novel stimuli indicating that subjects grasped
the MTS paradigm and were not simply shaped by operant conditioning.
Understanding how nonhuman animals think about and perceive the world
requires clever techniques that evade the need for language. Match-to-sample
(MTS) is a common procedure for learning about how non-human primates
categorize the environment. The task requires the subject to observe
a sample stimulus and then indicate (usually through touch) its
match from among a set of test stimuli. Previous studies (e.g. Tavares
& Tomaz (2002)) have shown that the capuchin monkey (Cebus apella)
can perform versions of MTS with novel stimuli after training. Others
(e.g. D'Amato & Colombo (1989); D'Amato & Colombo 1985;
Milner 1973; Salmon 1984) have shown that capuchins may have difficulties
in transferring from training to novel stimuli. Thus the goal of
this research was to examine the acquisition of the MTS task in
Capuchin monkeys and access its potential as a paradigm for future
studies.
Individual subjects were transported to a 72 x 60 x 60 cm wire
mesh test chamber for all training and testing. Stimuli consisted
of 8.5 x 11 in laminated paper with brightly colored or black &
white images. Training: Monkeys were rewarded for gently touching
a transparent stimulus behind which a Fruit Loop cereal reward was
held. Touch to transparent stimuli was transitioned to a set of
3 highly contrastive opaque stimuli without the reward being in
view. Next a touch to each of two identical stimuli held up in succession
was rewarded. A foil stimulus was then presented alongside the second
stimulus on a rod and only a touch to the identical stimuli was
rewarded. Experimental sessions Daily sessions consisted of 32 trials
with the arrangement and identity of sample and foil stimuli randomized
and counterbalanced. Criterion was designated as 84% accuracy on
3 consecutive testing sessions. Subjects were then transfer tested
in this manner on 3 novel sets of 3 stimuli designed to test the
role of color and pattern as cues for matching. The next stage utilized
another 3 sets of 3 transfer stimuli this time consisting of complex
black & white images of insects.
The 5 subjects reached criterion (84%) with the original stimuli
in an average of 7 sessions (min 4 max 9). Average performance on
the transfer tests was above criterion (contrastive 87% n=5; insects
90% n=3). Under all transfer conditions subjects both individually
and as a group chose the matching stimuli at a level significantly
greater than would be expected by chance (Heterogeneity G; p <
.001). Subjects demonstrated a steeper learning curve in transfer
tests than with original stimuli. There were no significant differences
in performance across transfer tests although patterns were evident
in individual subjects.
The capuchin subjects proved more than capable of acquiring and
performing the MTS task. Although accuracy increased overall throughout
the study high initial scores suggest that the matching concept
is easily grasped. Transfer to stimuli sets that differed only along
the variables of size or pattern did not significantly impact accuracy
and scores increased overall for the more complex yet colorless
insect sets. This result may have been partly an artifact of only
the best subjects reaching the insect phase within the possible
timeframe. However it is possible that increased attention to relevant
(the monkeys are very much insect connoisseurs) stimuli improves
matching ability. Success in matching complex stimuli (e.g. insects)
is promising for the use of the MTS paradigm to enhance our understanding
of capuchin cognition. Future studies might include matching: physical
objects to photos to explore the capuchin's ability to map to representations
various photos of the same individual to test their understanding
of identity or different but related pictures to investigate what
categories the capuchins naturally perceive.
Thanks to Marietta Dindo and Cassie Freeman for their knowledge
& experience and aid in testing. My gratitude to the Nuts &
the Bolts for being clever cooperative and endearing creatures.
This research was made possible by HMMI grant #52003727 and the
Emory SURE program 2003.
|
