Abstract 2, Ph.D. Sponsor: Newmont Mining
Frontier of Science 20111, Dept. of Psychology2, UNC
Executive function (EF) is a broad term for the higher order thought processes that are used in
new situations, things like problem-solving, working memory, inhibitory control, and planning. It
Mentor: Marilyn Welsh
can be separated into two divisions, “hot” EF dealing with emotionally and motivationally Iowa Gambling Task
influenced decisions, and “cold” EF dealing with purely cognitive problem solving. This study • Measure the risky decision making of the participants in a computerized, hypothetical gambling
examined the various similarities and differences between “hot” and “cold” executive function, situation.
Fig. 3 Fig. 4 Fig. 5
and attempted to make “cold” tasks “hotter” through giving traditional “cold” EF tasks in an • “Hot” EF task.
incentive condition. A total of 10 volunteer participants from the 2011 FSI program, ranging in • Requires the participant to makes a series of selections from an electronically delivered
age from 15 to 17, were tested through the use of the Iowa Gambling Task (IGT), Tower of (computerized) display of four card decks, A, B, C, or D.
London (TOL), and Letter-Number Sequencing (LNS). The IGT represented a typical “hot” task, • When a card is chosen by a mouse click, it results in a hypothetical win or a loss of some money
while both TOL and LNS were “cold” tasks. Performance on the TOL and LNS under non-incentive and the task taker’s running total of the amount of money won and the amount of money
conditions did not correlate, and neither did the TOL or LNS under non-incentive conditions. borrowed is displayed on the top of the computer screen.
However, the TOL task under incentive conditions did strongly correlate with scores on the IGT,
indicating that the “cold” task became “hotter”. Future studies in this area should include a larger
• 100 trials given in 5 blocks of 20.
• Two “good” decks (C and D) and two “bad” decks (A and B).
The primary purpose of this experiment was to study the effects of incentive and non-incentive
sample size in order to gather more definitive data and gain stronger statistical power. • “Bad” deck has a high reward but occasionally results in a very large loss. The “good” deck has
conditions on traditionally “cold” executive function tasks (Tower of London, Letter-Number
minimal reward but occasionally results in only a small penalty.
Sequence), and to see that, while given during incentive conditions, the tasks would correlate more
• The participant is unable to predict when they will lose or win money; however, they
Figure 1. The closely to a “hot” EF task (Iowa Gambling Task). The TOL and LNS would effectively become
presumably can figure out over time which are good decks and which are bad decks.
“hotter” under the incentive conditions.
prefrontal Tower of London (fig. 2)
The research question addressed how the incentive manipulation would influence the
• “Cold” EF task.
cortex is • The point of the task is to move the balls, one at a time, to build the goal pattern that is
performance (number correct) on the TOL and LNS tasks. The paired-samples t-test performed on
thought to be the data collected showed almost no differences in performance on the TOL or the LNS tasks under
presented to the student on a large card. The balls are set up in a “starting pattern” and the
incentive and non-incentive conditions.
the location of participant is told how many moves (4, 5, or 6) it will take to move the balls into the goal
Hypothesis 1 stated that the scores on the TOL and LNS tasks given under non-incentive conditions
executive • The 15 odd problems were given under incentive conditions, and the 15 even under standard
would be moderately correlated with each other because they are both considered to be “cold” EF
function tasks. Based on the results of the experiment, the hypothesis was not supported by the data.
Correlational analysis revealed a non-significant correlation.
Hypothesis 2 said that the scores on the TOL and LNS tasks given under non-incentive conditions
• “Cold” EF task, specifically working memory
Introduction • The participant is read a combination of numbers and letters and is asked to recall the numbers
first in ascending order and then the letters in alphabetical order.
will be correlated with the “hot” EF task, IGT, at a low magnitude. This hypothesis was also
unsupported by the data, with only an insignificant correlation for both the TOL and LNS. However,
Executive function (EF) consists of the higher order thought processes that are essential in novel an interesting trend was seen between the TOL and Block 1 of the IGT, where a strong negative
• There are seven items ranging from 2-letter/number sequences (e.g., B-7) to 8-letter/number
situations, such as problem-solving, working memory, inhibitory control, planning, etc. (Gilbert & correlation was seen. This finding suggests that participants who performed better on the TOL
sequences (e.g., S-2-L-8-B-1-G-7).
Burgess, 2008), which are thought to occur in the frontal lobe region of the brain (fig. 1) (Gilbert & under non-incentive conditions did worse (made more risky choices) on the beginning 20 trials of
• 7 trials were given under each condition
Burgess, 2008). The prefrontal cortex consists of the front most part of the frontal lobes, and is the IGT.
thought to be the part of the brain that is responsible for executive function. The prefrontal cortex Similar to Hypothesis 2, Hypothesis 3 stated that the scores on the TOL and LNS tasks, but given
can be divided into several sections: the dorsolateral (responsible for “cold” EF), the orbitofrontal under incentive conditions, will be correlated with the “hot” EF task, IGT, at a moderate to high
(responsible for “hot” EF), and the frontopolar (theorized to be the site of cognitive branching) magnitude. Correlational analysis on the LNS under incentive conditions against the individual IGT
(Hongwanishkul et al., 2010). Executive function can be further divided into two groups, both Figure 2. A trials and total net score showed a very small, non-significant correlation that ranged from being
“hot” and “cold” executive function, to describe the emotional and cognitive aspects associated positive to negative. However, the TOL under incentive conditions showed a significant correlation
with each respectively (Prencipe et al., 2011).
typical Tower to Blocks 2-5 and the net total of the IGT. These are the most important findings of this study, and
The “hot” EF deals with emotionally and motivationally influenced decisions, and “cold” EF of London although much more research does need to be performed, it does indicate that under incentive
concerns cognitive problem solving. “Hot” executive function involves traditional EF when there is problem conditions, the TOL (a “cold” task), did become “hotter” by correlating more closely with the IGT (a
a strong motivation to perform well, especially emotional or motivational influence “hot” task).
(Hongwanishkul et al., 2010). “Cold” (or “cool”) executive function involves the purely cognitive Future studies involving “hot” and “cold” executive function can improve on many of the flaws
aspect of EF, where there is little or no emotional or consequential influence. seen in this experiment. Most importantly, a much larger sample must be tested, in order to
The purpose of this study was to observe executive functions (EF) under incentive and non- collect more data, and eliminate many of the problems seen with testing a small sample, such as
incentive conditions, using a traditionally “hot” task (Iowa Gambling Task) and “cold” tasks (Tower outliers strongly effecting the correlations, etc.
of London, Letter-Number Sequencing). Through the use of incentive and non-incentive settings, it
is predicted that these “cold” tasks will become “hotter”. The data collected from our study will be
Results Although this study shows little practical implications, mainly due to the small sample size, it does
show the need for further research in this area of study. “Hot” and “cold” executive function is still
used to answer these three hypotheses, as well as our research question: The descriptive statistics (means and standard deviations) for the experimental measures can be being studied, and larger experiments are critical to understanding the many similarities and
Research Question (no apriori hypothesis): How will the incentive manipulation influence the found on Table 1 below. differences between them. Finding a strong correlation between the TOL under incentive
performance (number correct) on the TOL and LNS tasks? The research question addressed how the incentive manipulation would influence the conditions and IGT (while there was none between the TOL under non-incentive conditions and
Hypothesis 1: The scores on the TOL and LNS tasks given under non-incentive conditions will be performance (number correct) on the TOL and LNS tasks. The paired-samples t-test showed no the IGT) is a very interesting result, and further experimentation could support or hurt this finding.
moderately correlated with each other because they are both considered to be “cold” EF tasks. significant differences in performance on the TOL or the LNS tasks under incentive and non-
Hypothesis 2: The scores on the TOL and LNS tasks given under non-incentive conditions will be incentive conditions.
correlated with the “hot” EF task, IGT, at a low magnitude.
Hypothesis 3: The scores on the TOL and LNS tasks given under incentive conditions will be
Hypothesis 1 stated that the scores on the TOL and LNS tasks given under non-incentive conditions
would be moderately correlated with each other because they are both considered to be “cold” EF Acknowledgements
tasks. Correlational analysis found a non-significant positive correlation, r (8) = 0.268, p = 0.227 I’d like to thank Dr. Welsh for all her tremendous help on this project and paper, including the design of the study in the first place.
correlated with the “hot” EF task, IGT, at a moderate to high magnitude. She was an excellent mentor, and always answered any questions I had during the research process. I would also like to thank
(fig. 3). The scatter-plot indicates a positive association, however many outliers caused the Nathan Kirkley and Zabedah Saad for their help on editing this paper. My best regards go out to Lori Ball and UNC, for maintaining a
correlation to be non-significant. science program as awesome as FSI. Lastly, I’d like to thank Newmont Mining for sponsoring me to attend this program; it has been
Methods/Measures Hypothesis 2 stated that the scores on the TOL and LNS tasks given under non-incentive conditions
will be correlated with the “hot” EF task, IGT, at a low magnitude. Correlational analysis once again
a truly life-changing experience.
A total of 10 teenagers ranging from ages 15-17 (mean = 16.1, std. deviation = 0.57) voluntarily
participated in this study. The group contained 5 boys and 5 girls. All participants are active
found a non-significant positive correlation between the non-incentive TOL and IGT. This suggests
that better scores on the TOL were related to better (less risky) choices on the IGT. The exception References
to this finding was the first block of 20 IGT trials in which there was a significant negative Baddeley, A. (2010, February 23). Working memory. Current Biology, 20(4).
members of the 2011 Frontiers of Science Institute. D
Best, J. R., & Miller, P. H. (2010, November/ ecember). A Developmental Perspective on Executive Function. Child Development,
correlation between the TOL and IGT, r (8) = -0.744, p = 0.007 (fig. 4). This means that, for the first 81(6).
20 trials of the IGT, participants who did well on the TOL task made significantly worse (more risky) Brock, L. L., Rimm-Kaufman, S. E., Nathanson, L., & Grimm, K. J. (2009). The contributions of ‘hot’ and ‘cool’ executive funtion to
The participants were pre-divided into 2 groups, those given tasks under incentive conditions first, children’s academic achievement, learning-related behaviors, and engagement in kindergarten. Early Childhood Research Quarterly,
choices on the IGT. In general, there was a low correlation between the LNS under non-incentive
and those given tasks under incentive conditions second, in order to provide a counterbalanced (24).
conditions and IGT, meaning that higher scores on the LNS resulted in poorer choices on the IGT. A
Carlson, S. M., & Moses, L. J. (2001, July/ ugust). Individual Differences in Inhibitory Control and Children’s Theory of Mind. Child
order of testing. The group in which incentive tasks were given first was tested on 15 Tower of
The correlation between the LNS score and the first block of 20 IGT trials showed a moderate Development, 72(4).
London problems, 7 Letter-Number Sequencing problems, and the Iowa Gambling Task. The Crone, E. A. (2009). Executive functions in adolescence: inferences from brain and behavior. Developmental Science.
negative correlation, r (8) = -0.536, p = 0.055.
participants were informed that their performance on the tasks would determine the number of Gilbert, S. J., & Burgess, P. W. (2008, February 12). Executive function. Current Biology, 18(3).
Hypothesis 3 stated that the scores on the TOL and LNS tasks given under incentive conditions will Hongwanishkul, D., Happaney, K. R., Lee, W. S. C., & Zelazo, P. D. (2010, June 8). Assessment of Hot and Cool Executive Function in
entries to receive a prize. After the Iowa Gambling Task (IGT) the group was given another 15
be correlated with the “hot” EF task, IGT, at a moderate to high magnitude. Correlational analysis Young Children: Age-Related Changes and Individual Differences. Developmental Neuropsychology, 28(2).
Tower of London problems and 7 more Letter-Number Sequencing problems, where they were told Kerr, A., & Zelazo, P. D. (2004, June). Development of “hot” executive function: The children’s gambling task. Brain and Cognition,
yielded several significant positive correlations between the TOL under incentive conditions and
that the performance would not count toward entries for the prize. For the second group of 55(1).
IGT Block 2, r (8) = 0.598, p = 0.034, IGT Block 3, r (8) = 0.726, p = 0.009, IGT Block 4, r (8) = 0.725, p Prencipe, A., Kesek, A., Cohen, J., Lamm, C., Lewis, M. D., & Zelazo, P. D. (2011). Development of hot and cool executive function
participants, 15 Tower of London problems and 7 Letter-Number Sequencing problems were given
= 0.009, IGT Block 5, r (8) = 0.633, p = 0.025, and the IGT net total, r (8) = 0.776, p = 0.004 (fig. 5). during the transition to adolescence. Journal of Experimental Child Psychology, (108).
first, with performance not counting towards entries. Then the Iowa Gambling task, 15 different Russo, N. (2003). Executive function and autism (Doctoral dissertation, McGill University, Montreal).
Correlations between the LNS under incentive conditions and the IGT were all non-significant,
Tower of London problems, and 7 more Letter-Number Sequencing problems were given, with Seguin, J. R., Arseneault, L., & Tremblay, R. E. (2007). The contribution of “cool” and “hot” components of decision-making in
ranging from negative to positive. adolescence: Implications for developmental psychopathology. Cognitive Development, (22).
participants being told that performance on these tasks would count towards the drawing.