Home Uncategorized ESSAYGWIZARD

Uncategorized

ESSAYGWIZARD

AGAIN, THE DISSERTATION ATTACHED IS THE PROFESSORS EXAMPLE THAT IS JUST FOR A FEW STRUCTURAL ELEMENTS IT DOES NOT NEED TO BE EVEN MORE THAN A HANDFUL OF PAGES.

(ESSAYGURU – I’d like my topic to be on making the decision of taking a new job while taking a difficult class…if you can make that work, it would be perfect as the professor knows my life vs. school struggles at the moment.)

ASSIGNMENT:

The science of decision making is a young one. There is still much to discover about how/why people make the decisions they do. For this assignment I want you to perform a decision making experiment that will expand our knowledge of how people make decisions. This assignment should include:

1.) A well written description of the type of decision you are studying and what we already know about decisions of this sort, i.e. research of this type of decision.

2.) A good enough description of your experiment that I could replicate it if I wanted to.

3.) A thoughtful analysis of the results of your experiment. E.g. a graph or table that makes it clear what you found.

4.) A discussion of what your results mean and what they tell us about how people make decisions.

5.) A list of future studies that could further our understanding of Decision Making based on what you discovered or improvements to your methods.

Example:

One student did an experiment that looked at weather people at a dog park picked up the dog poo more when other people were around or if they were by themselves (the student apartment was overlooking the park). In this experiment they demonstrated the importance of social decision making and how people’s willingness to follow rules changed in different circumstances.

See my dissertation (attached below)
for general structural guidance
. My report will obviously be more in depth but hopefully you can get guidance from the
structure
and how I laid out the past research (
Introduction
), how I did the experiments (
Methods
), what I found (
Results
) and what these results mean (Disscussion and future suggestions).

A TALE OF TWO SYSTEMS: EXECUTIVE FUNCTION IN ULTIMATUM GAME DECISIONS
by

Aaron Daniel Kuechler Tesch

_______________

A Dissertation Submitted to the Faculty of th

DEPARTMENT OF PSYCHOLOGY

In Partial Fulfillment of the Requirements
For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2009

THE UNIVERSITY OF ARIZONA

GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation
prepared by Aaron Daniel Kuechler Tesch

entitled A Tale of Two Systems: Executive function in ultimatum game decisions
and recommend that it be accepted as fulfilling the dissertation requirement for the Degree

of Doctor of Philosophy

____________________________________________________________Date: 11/11/0

Alan Sanfey

____________________________________________________________Date: 11/11/09

William Jake Jacobs

___________________________________________________________Date: 11/11/09
Anouk Scheres

___________________________________________________________Date: 11/11/09

Lee Ryan

Final approval and acceptance of this dissertation is contingent upon the candidate’s
submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and
recommend that it be accepted as fulfilling the dissertation requirement.

____________________________________________________________Date: 11/11/09
Dissertation Director: Alan Sanfey

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an
advanced degree at the University of Arizona and is deposited in the University Library
to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided
that accurate acknowledgment of source is made. Requests for permission for extended
quotation from or reproduction of this manuscript is whole or in part may be granted by the
head of the major department of the Dean of the Graduate College when in his or her
judgement the proposed use of the material is in the interests of scholarship. In all other
instances, however, permission must be obtained from the author.

SIGNED: Aaron Daniel Kuechler Tesch

Table of Contents
Figure Index………………………………………………………………………………………………………….

Table Index…………………………………………………………………………………………………………..

Abstract………………………………………………………………………………………………………………..

Introduction…………………………………………………………………………………………………………..

Greek rationality…………………………………………………………………………………………………….9
Problems with purely rational theories………………………………………………………………………

Emergence of non-rational theories…………………………………………………………………………..

Two-system models…………………………………………………………………………………………………

Neurological two-system models………………………………………………………………………………

Possible Problems with Two-system models………………………………………………………………

The Somatic Marker Hypothesis………………………………………………………………………………

Can low-level systems be separated from high-level cognitive systems?……………………….

A two-system model for second player ultimatum game decisions………………………………..

Experiment 1 – Working Memory Easy Study…………………………………………………………….

Experiment 1 – Methods………………………………………………………………………………………

Experiment 1 – Results………………………………………………………………………………………..

Experiment 1 – Discussion…………………………………………………………………………………..

Experiment 2 – Working Memory Hard Study……………………………………………………………

Experiment 2 – Methods………………………………………………………………………………………

Experiment 2 – Results………………………………………………………………………………………..

Experiment 2 – Discussion…………………………………………………………………………………..

Experiment 3 – Within Subject Variable Memory Study………………………………………………

Experiment 3 – Methods………………………………………………………………………………………

Experiment 3 – Results………………………………………………………………………………………..

Experiment 3 – Discussion……………………………………………………………………………………

Experiment 4 – Incongruent (Incongruent) Feedback Manipulation Study……………………..

Experiment 4 – Methods………………………………………………………………………………………

Experiment 4 – Results………………………………………………………………………………………..

Experiment 4 – Discussion……………………………………………………………………………………

Experiment 5 – Priming vs. Loading Study…………………………………………………………………

Experiment 5 – Methods……………………………………………………………………………………….

Experiment 5 – Results…………………………………………………………………………………………

Experiment 5 – Discussion…………………………………………………………………………………….

Overall Results………………………………………………………………………………………………………..

Overall Discussion…………………………………………………………………………………………………..

Appendix A…………………………………………………………………………………………………………….

References………………………………………………………………………………………………………………98

Figure Index

Figure-1. Time line for Experiment 1 & 2………………………………………………………………….

Figure-2. The difference between Easy Working Memory and Control conditions………….31
Figure-3. Correlation of ultimatum game acceptance rates and working
memory task error rates…………………………………………………………………………………………..36
Figure-4. Memory pattern entry key………………………………………………………………………….

Figure-5. Experiment 3 timeline……………………………………………………………………………….

Figure-6. N-back instructions…………………………………………………………………………………..

Figure-7. Example Global-Local Stimuli…………………………………………………………………..

Figure-8. Memory ability vs. memory difficulty in

Experiment 3.

………………………………..48
Figure-9. Ultimatum game acceptance rates vs. feedback condition………………………………

Figure-10. Ultimatum game acceptance rates vs memory difficulty level……………………….

Figure-11. The relationship between perceived memory difficulty and

acceptance rates.

……………………………………………………………………………………………………..

Figure-12. The relationship between performance on N-back tasks and ultimatum game acceptance
rates……………………………………………………………………………………………………………………….

Figure-13. The relationship between visual digit span performance and
acceptance rate in the Correct Feedback condition……………………………………………………….56
Figure-14. Order effects for Experiment one and the Easy Working Memory………………….

Figure-15. Memory performance vs. memory difficulty for the Incongruent
Feedback condition………………………………………………………………………………………………….67
Figure-16. Correlations between % correct and % accepted vs. conditions……………………..

Figure-17. Ultimatum game acceptance rates vs. all feedback conditions……………………….

Figure-18. Memory performance vs. memory difficulty separated by
Priming and Loading conditions………………………………………………………………………………..75
Figure-19. Bar graph of memory performance and condition vs. ultimatum
game acceptance

rate.

……………………………………………………………………………………………….

Figure-20. Memory performance and memory difficulty vs. ultimatum game
acceptance rates……………………………………………………………………………………………………….

Figure-21. Memory difficulty and offer level vs. ultimatum game
acceptance rates……………………………………………………………………………………………………….

Figure-22. Bar graph of SSRT defectors and compliers vs. ultimatum game
acceptance rates……………………………………………………………………………………………………….

Figure-23. Ultimatum game acceptance rates compared for all conditions………………………88

Table Index
Table-1. Example table of % accepted and % correct for each memory
difficulty level…………………………………………………………………………………………………………

Table-2. The correlations between executive function for the Correct
Feedback condition………………………………………………………………………………………………….

Table-3. The correlations between executive function for the No Feedback
condition………………………………………………………………………………………………………………..

Table-4. Priming vs. loading timeline………………………………………………………………………..74

Abstract

Theories that formally describe decision-making have traditionally posited that decisions are

made by rational actors. However, it is generally accepted that humans often make irrational decisions

because of quick emotional judgements. In order to reconcile these two inconsistent ideas psychologists

have developed two-system theories that hypothesize decisions are made by two opposing cognitive

systems, representing the rational and emotional processing of decisions. Evidence for a two-system

model of decision-making can be observed in ultimatum game responder decisions. It is thought that

rational processing of these choices will produce acceptance of unfair offers and emotional processing

will encourage rejection of unfair offers. Emotional priming has been shown to decrease ultimatum

game acceptances and trans-cranial magnetic stimulation of rational brain areas, i.e. DLPFC, show

increases in ultimatum game acceptances. This study investigated the possibility of using behavioral

tasks that are known to activate rational brain areas to promote/disrupt ultimatum game acceptances.

The possible relationship between ultimatum game acceptances and executive functions was also

examined. Although there were promising indications that working memory loading may increase

ultimatum game acceptances in between-subject experiments, a within-subject investigation found little

support for this method of promoting/disrupting rational ultimatum game decisions. There were also no

relationships found between switching or inhibition executive functions and ultimatum game responder

decisions. A moderate positive relationship was found between updating executive function and

ultimatum game acceptance rates but this relationship was dependent on working memory task

feedback, a within-subject design and active loading of the working memory system. However, its

possible that these findings only apply to within-subject paradigms and future between-subject studies

are advised.

Introduction

Classical economic theory posits that the universe followed rational laws and that human

thought should therefore be rational. However, this premise is challenged every day by the irrationality

and emotionality of human decisions and behavior. One common explanation for this imbalance

between what we should do and what we actually do is to propose that we have within us two

diametrically opposed processes. One process behaves rationally and the other behaves instinctually.

This type of theory is called a two-system model theory. This paper will examine the plausibility of

promoting/disrupting the rational system with a working memory task while people are dealing with

unfairness. Decision processes that assess unfairness are responsible for many if not most of our

decisions we make every day, and thus discovering how they work should be one of the pillars of

decision-making theory.

Greek rationality

Socrates believed that all behavior is directed by a rational pursuit of a person’s goal. Or more

elegantly said, “nobody acts against his [or her] better knowledge” (Frede, & Striker, 1996 p. 7).

Socrates’ intellectual children, Plato and Aristotle later modified this thinking to include elements of the

soul that were free to be irrational. These later theories can be thought of as the first two-system

models. Plato believed in a purely rational spirit world where our souls inter-mingled with pure ideas,

i.e. Forms, only to be corrupted by the non-spirit world. For example, in Plato’s story of Meno, a slave

boy spontaneously discovers a geometric principle when pressed by Socrates (Klein, 1965). Plato

makes the argument that this geometric principle as well as principles like virtue, are imbedded in us

and we only need to be reminded for our souls to recall their pure spirit world “Form”. Plato believed

that this spirit world is absolutely analytical where pure rationality, i.e., mathematics, reigns supreme.

So Plato would explain non-rational behavior as part of the mental corruption where the physical world

has strayed from true decisions. Later philosophical traditions like stoicism also adopted Socrates’

belief in a purely rational universe and thus postulated that to find truth they had to purge any physical

world corruption, i.e., emotion, from their thinking (Hadas, 1961). While these different Greek

philosophies differ on the extent that people act rationally they all agree that pure analytic rationality

represents the ideal of how we should make decisions. This belief continues to be at the core of all

normative economic theories.

Problems with purely rational theories

The belief in the purity and trueness of rational decisions lead to analytical, i.e., normative,

models of decision-making. In these models, people are thought to make decisions by calculating the

value of the possible outcomes and multiplying by the probability of each outcome and then rationally

choosing the option with the highest calculated value. For example, if someone had to pick between an

option of getting $1 or a 50% chance of getting $3 they would always pick the second option because it

has a calculated expected value of $1.50, which is higher than the $1 expected in the first option.

A major challenge to this purely rational theory of economic behavior was made by Daniel

Bernoulli (1738/1954) who pointed out that people do not fully equate money and value. He observed

that when given the chance to buy an option with infinite expected value, i.e., the amount to win

multiplied by the probability of winning is infinite, people often refuse. If people used purely rational

calculations of what option has the highest expected value and then choose the option with the highest

value this should never happen. To demonstrate the disconnect between what people actually decide

and what pure rationalist calculations would predict, Bernoulli described the St. Petersburg paradox. In

this paradox the reader is presented with a bet that would give a dollar, or an equivalent form of

currency, if a flip of a coin came up heads and for every subsequent heads, would be given an

additional 2n-1 dollars (where n is the number of heads) until the coin came up tails. So if the coin only

came up heads once you would get one dollar; if it came up heads two times you would get 1+2 2-1

dollars; if it came up three times you would get 1 + 2 2-1(2)+ 2 3-1 (4)… dollars and this pattern would

continue until the coin came up tails. This bet has an infinite calculated expected value because each

possible coin flip has a calculated expected value of a half dollar, and since there are endless possible

coin flips the total expected value is calculated to be infinite. However, it has been observed that

contrary to the prediction of the calculated expected value, people do not put an infinite value on this

coin-flipping gamble since they are rarely willing to place a value of over $25 (Martin, 2004).

Bernoulli explained that the reason people are unwilling to pay very much for this bet is that money is

not treated as an exact measure of value. Money can only buy so much happiness, i.e., value, so having

an infinite amount of money would not mean an infinite amount of happiness.

The unwillingness to pay for a bet with an infinite expected value demonstrates that contrary to

normative economic theory people do not treat money as an exact measure of value. Bernoulli’s theory

was formalized into new normative decision-making theories that posit a single information processing

system that computes decisions by logical analysis of all available information (Von Neumann &

Morgenstern, 1944). These theories are described as utility theories because they posit that people will

calculate the expected utility, i.e., how much it is worth to them, of the possible outcomes of a decision

and choose the option with the highest expected utility.

Utility theory has been challenged by the insight that humans have a limited ability to calculate

expected utility. For example, Simon (1956) pointed out that strict utility maximization couldn’t be

used to describe actual decision-making behavior because humans have a limited ability to process all

of the factors that would be needed to fully evaluate all decision options. In other words the limited

capacity of human memory and cognitive processing power makes a strict utility calculation

impossible. To solve this conundrum it has been suggested that people avoid endless utility calculations

by using cognitive short cuts called heuristics (Gigerenzer, Todd & The ABC Research Group, 1999).

Tversky and Kahneman’s (1974) Prospect Theory also challenged the idea that people make decisions

based on strict utility calculation by suggesting that psychological weighting of probabilities and value

do not conform to utility theories. These challenges have weakened the dominance of economic

theories that postulate the strict rational calculation of utility to explain decisions. Nevertheless, strict

rationalist utility theories are often still lauded as the best way to predict behavior (Camerer, 2009).

Emergence of non-rational theories

The dominance of the theories that assumed rationality in our behavior and thinking was

probably most famously disputed by Sigmund Freud. Freud challenged the convention of assuming

rational decisions when he divided the mind into the canonical Id, Ego, and Superego (Freud, 1949).

This theory recognized that much of our behavior is directed by primitive non-rational, i.e., emotional,

motivations as well as the rational facets of the mind. This theory conflicted with to the philosophy of

stoic restraint that was promoted as the ideal for all behavior in the Victorian society of Freud’s time

(Gay, 1998). It should be noted that Freud’s reasoning relied mainly on distorted descriptions of case

studies and narcissistic self-reflection. However, it should also be noted that more methodological

psychologists such as William James had also considered primitive emotional instincts as important

motivators of behavior (Eysenck, 1986).

The idea that behavior was influenced by non-rational facets of the mind was exploited by

advertising pioneers like Walter Dill Scott. For example, in Scott’s advertising manifesto “The Theory

and Practice of Advertising” he details how advertisements should appeal to the non-rational

motivations of consumers (Scott, 1903). In other words, advertising should try to speak to costumers

primitive drives, i.e., sex sells.

Two-system models

The introduction of theories that focus on non-rational processing of decisions was an

important step towards the realization of theories that can predict behavior more accurately. However,

emotional non-rational theories like their purely rational counterparts also fail to predict decision-

making with much certainty. Therefore, in an attempt to improve these disparate theories hybrid

theories have been hypothesized. These hybrid theories postulate that decisions are made through the

competition of rational and emotional cognitive systems. The most common formulation of a hybrid

theory pits non-rational processes, i.e., emotional, against volitional, i.e., rational, processes (Posner &

Snyder, 1975; Schneider & Shiffrin, 1977; Metcalfe & Jacobs, 1996). Many versions of these

dichotomous decision-making processes have been proposed, but the general concept was well

described by Kahneman (2003), who stated that there are two evaluation systems termed System 1 and

System 2. In this conceptualization non-rational processes are represented by a System 1, which is fast,

emotional, automatic, uses heuristics, and can work in parallel with other systems. Volitional processes

are thought to be represented by a System 2, which is relatively slow, completely rational, unemotional,

can override System 1, and processes information serially. It should also be noted that some decision-

making theorists have suggested that the operation of these two systems can be better described as a

continuum of emotional and volitional processing rather than decisions being dominated by one or the

other (Cohen et al., 1990). An example of how these systems are thought to interact can be

demonstrated in the hypothetical reaction to being bumped in a hallway. An initial emotional reaction,

i.e., System 1, to being bumped might be to hit the person back, and for some people this is what

happens. However, a volitional reaction, i.e., System 2, often overrides this initial emotional reaction.

We often realize the bump may have been a mistake, and as a result we ignore it or say “excuse me”.

Neurological two-system models

Le Doux (1994) proposed one example of a neurological two-system model, which was based

on a theory that there are two distinct visual processes in the brain. One process, often called the high

road follows a route from the retina, through the lateral geniculate nucleus, i.e., visual thalamus, and

onto the visual cortex. The other process, often called the low road, follows a route from the retina to

the lateral geniculate nucleus then directly to emotional areas like the amygdala. Öhman and Soares

(1994) demonstrated the unconscious visual processing of the low road by presenting participants with

phobic stimuli, i.e., pictures of spiders, too quickly to be consciously perceived, and then detecting

emotional responses after the pictures were presented. We are most often unaware of the visual

processing that follows the low road because it bypasses conscious awareness of sight. This

unconscious emotional visual processing can be demonstrated in the classroom by placing a replica of a

bug at the entrance of the room. As people enter the room may get an emotional jolt before they realize

this bug is not real and thus does not pose a threat. This demonstration makes it clear that the bug is

being processed by two different neurological circuits. One circuit is fast, emotional, and unconscious

while the other is relatively slow and can override the emotional reaction once it is assessed that there

is no threat.

The separation of conscious and unconscious visual processing is also be supported by the

phenomenon known as blindsight. Hints that there might be sight without the visual cortex can be

demonstrated by removing the visual cortex, i.e., occipital lobe, of monkeys and then observing the

monkeys respond to visual stimuli (Kluver, 1949; Blackmore, 2004). Human neuropsychological

patients were also discovered who claimed they can not see but can also identify objects, color and

movement using visual information despite their conscious blindness (Weiskrantz, 1986). Like the

high and low road visual processing theory the existence of blindsighted individuals suggests that there

is more to vision then a single unified neurological system and implies that there are both conscious

and unconscious vision systems.

A similar dichotomy can be found in the expressions of the face. Duchenne (1855) discovered

humans have specialized facial muscles that activate during different emotional states. These muscles

have been linked to separate emotional and volitional systems (Damasio, 1994). For example, when a

photographer tells little kids to smile they often will produce a forced smile, but if they tell the kids a

silly joke, i.e., “say monkey butts”, the kids will produce a natural smile. Two distinct cognitive

systems seem to be responsible for facial expressions. The first system is unconscious and responds to

emotional stimuli and the second is a conscious volitional system. It should be noted that humans can

and do train their volitional expression system to override emotional expressions. However, this

volitional expression system is relatively slow so emotional expressions, i.e., micro-expressions, often

are produced before being overridden (Ekman, 2003).

Evidence for two-system models involved in decision-making has also been mounting. In

numerous neuroimaging studies brain areas whose activation seems to be correlated with both the

emotional and volitional systems have been identified. For example, McClure et al. (2004) found that

frontal brain areas (∂ areas) were more active than a set of limbic areas (ß areas) during decisions to

delay payment to get a larger reward. However, there were no differences between the activation of

these brain areas when participants were making decisions to take smaller rewards sooner rather than

waiting for a greater reward. This study demonstrated that neural activation in brain areas associated

with rational volitional deliberation correlated with decisions to delay reward. This finding supports a

two-system decision-making theory where dichotomous cognitive systems compete to dominate

decisions.

This neurological two-system model framework may also help to explain the inconsistent

decisions participants make during delay discounting paradigms (Kirby & Marakovic, 1995). For

example, some participants would take an immediate reward ($1 now) over a delayed reward ($2 in a

month) but would not take a relatively short delayed reward ($1 in a month) over a long delayed reward

($2 in two months). Both of these sets of choices have the same delay and reward differences between

the options. Differences in the emotionality of decisions and therefore the relative activation of ß areas

may explain the decision inconsistencies observed by Kirby and Marakovic (1995). The reduced

activation of the ∂ areas when participants refuse to wait for a larger reward suggests that these choices

may result from less self control or a relatively weak System 2.

A two-system model may also help explain decisions we make about moral behavior. Greene et

al. (2001) found that when participants were making deontological decisions, i.e., concerned most with

rightness vs. wrongness, a set of brain areas associated with emotion were relatively more active then

areas associated with rationality. These deontological decisions happened most often when participants

were making personal moral judgments, i.e., being in direct contact with a person your decision would

kill. However, when participants were making utilitarian decisions, i.e., concerned with the best

outcome, areas of the brain associated with rationality and working memory were more active. These

utilitarian decisions are most common when making impersonal moral decisions, i.e., pulling a lever

that will result in a distant person’s death. These findings are supported by the observation that priming

disgust can increase deontological reactions to moral problems, because disgust activates emotional

brain areas (Schnall, Haidt, Clore, & Jordan, 2008; Sanfey et al., 2003). It has also been found that

damage to emotional processing brain areas, i.e., ventromedial prefrontal cortex, increases utilitarian

moral decisions further supporting the idea that there is an emotional component to our moral decision-

making (Koenigs et al., 2005). These studies of moral decision-making seem to demonstrate that some

moral decisions can be explained by a two-system model.

The neurological studies described above demonstrate that some brain areas have activation

patterns that correspond to the theoretical conceptualization of System 1 vs. System 2 decision-making.

Sanfey and Chang (2008) suggest that there might be distinct neurological systems that fit a System 1

vs. System 2 model of decision-making. That being said, it remains unclear exactly, which brain areas

and thought patterns are part of a System 1 or System 2. More work is needed to elucidate the outlines

of both the anatomical and cognitive components of these systems.

Possible Problems with Two-system models

Two system models that pit a non-rational emotional system against a rational volitional system

are not universally accepted and some evidence suggests that these models miss a number of important

exceptions to the standard two-system model explanations. One highly successful theory that may

challenge a strict distinction between an rational cognitive system and an emotional cognitive system is

the Somatic Marker Hypothesis. The Somatic Marker Hypothesis suggests emotions can help us make

more rational decisions. There are also many examples of how low-level systems like System 1 change

the functioning of higher level systems, i.e., System 2. These findings suggest that all decisions may

not be clearly described by distinct and hierarchical System 1 and System 2.

The Somatic Marker Hypothesis

One of the most influential neuropsychological decision-making case studies is the story of

Phineas Gage. In 1848 Phineas Gage had a metal tamping rod pass through his frontal lobe. His

subsequent personality changes and bad decision-making inspired much neuropsychological

speculation. One line of speculation focused on Phineas’ apparent bad decision-making after his

accident (Sanfey, Hastie, Colvin, & Grafman, 2003). For example, before his accident Phineas was a

responsible man, demonstrated by his attainment of the position of foreman for his work crew.

However, after his accident he became irresponsible, as evidenced by his refusal to observe social

norms, and building up gambling debts.

Phineas Gage is not thought to be an isolated case study because many other cases have shown

similar behavioral changes after frontal lobe damage. For example, Eslinger and Damasio (1985)

described a patient EVR who was a responsible accountant before the growth of a large orbitalfrontal

meningioma. After the removal of EVR’s tumor his poor decision-making led him to bankruptcy and

divorce. However, despite EVR’s poor life decisions other mental faculties, i.e., IQ, memory, speech,

attention and MMPI scores, seemed to be normal. In fact, the behavioral pattern of increased

irresponsibility with no change in intelligence has been found to be common in people with

ventromedial prefrontal cortex (VMPFC) lesions (Bechara et al., 1998).

Another important finding that clarifies the role of the VMPFC is the carful dissection of the

prefrontal cortex. Much of the frontal lobe lesion literature has categorized patients as having a general

frontal lobe syndrome and not distinguished between lesions in the dorsal lateral prefrontal cortex

(DLPFC) and those in the VMPFC (Bechara et al., 1998). However, Bechara et al. (1998) found a

functional separation between these two brain areas. In a paradigm called the Iowa Gambling Task,

subjects picked between 4 card decks two with low probability large rewards (A and B) and two with

high probability small rewards (C and D). Decks C and D were advantageous; choosing them would

result in gaining money over time. Decks A and B were disadvantageous; choosing them would result

in loosing money over time. The researchers found that patients with VMPFC damage performed much

worse on a gambling task, i.e., picked risky card decks with lower expected rewards, than patients with

DLPFC lesions and non-brain damaged controls. This suggested that the VMPFC is required for good

decision-making during situations that require avoiding risk.

It can also be noted that people with VMPFC lesions also have difficulty processing emotions

(Bechara & Damasio, 2005). The paired impairment of both decision-making and emotions in people

with VMPFC lesions led to the theoretical proposition that these two deficits might be a consequence

of the disruption of a single cognitive circuit. This idea was formalized in Damasio’s Somatic Marker

Hypothesis, which proposes that an important part of the cognitive processing of decisions requires the

processing of visceral emotions (Damasio et al., 1991). Hohmann (1966) reports that paraplegics seem

to have a loss of subjective emotions in proportion of the loss of their visceral sensations, suggesting

that visceral emotions are an important component of the processing of subjective emotions. The

Somatic Marker Hypothesis explains the impairment of decision-making observed in people with

VMPFC lesions by referring to the discovery that VMPFC patients do not feel emotional sensations

before choosing risky gambles. The lack of emotional responses is theorized to lead these patients to

continue taking risks, which people with fully functioning emotional systems do not take.

The Somatic Marker Hypothesis is supported is supported by Bechara et al. (2000a) and

Bechara et al. (2000b), who used the Iowa Gambling Task to show that people with VMPFC lesions do

not have anticipatory skin conductance responses when choosing risky choices, i.e., disadvantageous

decks A and B. The lack of an anticipatory skin conductance response when picking risky decks

demonstrates how damage to the VMPFC can interrupt a somatic state. The Somatic Marker

Hypothesis explains the disadvantageous choices made by people with VMPFC lesions by suggesting

that without anticipatory somatic states like skin conductance responses, they don’t fear the future

consequences of picking disadvantageous decisions, i.e., card decks or other real life monetary

decisions.

In an earlier study, Bechara, Damasio, Tranel, and Damasio (1997) found that these types of

choices are often made unconsciously and conscious knowledge of the location of advantageous decks

often doesn’t translate into advantageous choices. Bechara et al. (1997) also found that good decisions,

e.g. preferring advantageous decks, can be made by controls before they can consciously report their

strategy. These studies provide evidence to suggest that conscious knowledge of contingencies is

neither required nor sufficient to make advantageous decisions about risk.

Another prediction of the Somatic Marker Hypothesis is that damage to brain areas associated

with the production of somatic states should result in disadvantageous choices. For example, damage to

the amygdala, an important somatic state generator, should have decision-making created problems. To

test this prediction Bechara, Damasio, Damasio and Lee (1999) used the Iowa Gambling Task to

compare somatic markers, i.e., skin conductance responses, and the decision-making ability of people

with bilateral amygdala damage to people with VMPFC damage. They found that both groups picked

disadvantageous choices reflecting their impaired decision-making. However, when actually

experiencing a bad outcome the people with VMPFC damage, but not the people with amygdala

damage, had a skin conductance response. These results suggest that the amygdala is important for the

production of all skin conductance responses while the VMPFC is only important for anticipatory skin

conductance responses. These findings suggest that in order to make advantageous decisions we need

the ability to produce anticipatory skin conductance responses, which help us avoid making

disadvantageous decisions.

Earlier research by Tranel, Bechara, Damasio, and Damasio (1996) further elucidates the role of

the VMPFC in cognitive processing by studying the possible role of the VMPFC in fear conditioning.

They found that people with VMPFC lesions have inhibited fear responses to conditioned fear stimuli.

This suggests that the VMPFC is involved in Pavlovian conditioning of phobic stimuli as well as

anticipatory somatic markers.

Other studies have helped narrow down the VMPFC regions that could be responsible for

anticipatory skin conductance responses and thus normal decision-making during the Iowa Gambling

Task. In one such study Tranel, Bechara, and Denburg (2002) had people with VMPFC lesions

concentrated on either the left or right hemisphere complete the Iowa Gambling Task. They found that

people with VMPFC lesions restricted to the right hemisphere performed similarly to controls, and that

people with VMPFC lesions restricted to the left hemisphere preferred disadvantageous choices

although not to the same degree as people with bilateral VMPFC lesions. This suggests that the region

of the VMPFC that is most important for producing anticipatory somatic states is the left VMPFC.

Anatomical distinctions like the one described above might also help explain why not all people with

VMPFC lesions show preferences for disadvantageous choices during paradigms like the Iowa

Gambling Task (Sanfey et al., 2003b).

The Somatic Marker Hypothesis might also be a partial explanation of Prospect Theory

(Bechara & Damasio, 2005). For example, It may explain why losses have a greater psychological

weight than gains because losses are more emotional. Anticipation of losses is thought to induce greater

somatic states, which translates into a feeling of greater psychological weight for losses. The Somatic

Marker Hypothesis might also explain the decaying value curves described in prospect theory (Tversky

& Kahneman, 1974). If the value of a reward is proportional to the visceral response that it elicits, then

impairment of visceral responses can therefore limit the subjective value of a reward.

The Somatic Marker Hypothesis seems to demonstrate that emotional systems (System 1) can

support the rational decisions that are attributed to the purely rational systems (System 2). This can be

seen as a challenge to the idea of a discrete System 1 and System 2 because the Somatic Marker

Hypothesis demonstrates emotional support for volitional decision-making. However, it should be

noted that proponents of two-system modeling can interpret the findings that lead to the Somatic

Marker Hypothesis as simply a demonstration of the functioning of one of the systems (A. Sanfey,

personal communication, October 22, 2009).

Can low-level systems be separated from high-level cognitive systems?

In the construction of the standard two-system model it is assumed that System 2 dominates

System 1. System 2 is thought to evaluate System 1’s output and can volitionally inhibit decisions that

do not agree with its output. However, this assumes that System 1’s output is arrived independently

from System 2. Fodor (1983) hypothesized that all cognitive functions are modularized like the formal

description of System 1 and System 2 described above. Fodor’s theory states that fully formed

cognitions are arrived in each module independently (Francesco, 2001). However, this theory has been

challenged by perception studies like Peterson and Grant (2003), which demonstrated that even the

most simple cognitive functions like processing visual figure-ground distinctions are influenced by

other modules like memory. An example, of memory’s influence in figure-ground perception can be

demonstrated by the arrow in the FedEx Logo. There is an arrow between the ‘E’ and the ‘x’ but unless

pointed out people can rarely see it and after reading this report it will be very hard not to see this

arrow. This example demonstrates that past experience, i.e., memory, can influence even the most basic

cognitive functions and that there are many connections between different cognitive modules, which

violate the assumption that these processes are completely independent. Thus it is quite possible that

System 1 and System 2 are not fully independent cognitive modules.

A two-system model for second player ultimatum game decisions

One area where there has been a great effort to investigate a possible two-system model is in the

investigation of the decision processes that underlie decisions made during the ultimatum game (Guth

et al.,1984). In this game there are two players. The first player, i.e., the proposer, proposes a split of

$10 between themselves and the second player, i.e., the responder. The responder then has to decide

whether accept or reject the proposed split of the money. If the responder rejects the proposal then

neither player receives any money, but if the responder accepts the proposal then the money is split as

the proposer suggested.

The rational solution to this game is that the proposer should give as little money as possible to

the responder because the responder should accept any non-zero offer. Research has found that contrary

to this prediction, most offers are fair, i.e., $5 offers, offers less than $5 are rejected about 50% of the

time, and lower offers are rejected in higher proportions than higher offers (Camerer, 2003).

Participants report that they reject offers because they are perceived as unfair (Pillutla & Murnighan,

1996; Xiao & Houser, 2005). To discover the brain areas associated with these decisions, Sanfey et al.

(2003a) looked for brain activations in ultimatum game responders when rejecting or accepting offers.

They found greater activation in the anterior insula than in the dorsolateral prefrontal cortex (DLPFC)

when participants were rejecting offers and when accepting offers. When there was greater activation in

the DLPFC than in the anterior insula. Ultimatum game acceptance rates were also negatively

correlated with the magnitude of activity in the anterior insula.

These results are consistent with a two-system model of decision-making. The anterior insula is

known to activate when participants are in pain or disgusted, suggesting that this area might be part of

an emotional System 1 and its activation may represent the negative feelings associated with unfairness

(Derbyshire, 1997; Calder, 2001). Activation in the anterior insula is also correlated with emotional

social decisions and empathic responses (Rilling et al., 2007; Singer et al., 2006). In contrast, activation

associated with the DLPFC has been linked to goal maintenance, working memory, and executive

function thus alluding to the possibility that its activation represents part of the rational System 2

(Duncan, et al., 1996; Smith & Jonides, 1999; Manoach et al., 1997; Miller & Cohen, 2001).

Further study of the brain systems used in the accept/reject decisions of the responders during

the ultimatum game have increased our understanding of these processes. Harle and Sanfey (2007)

found that priming the incidental moods of negative affect and disgust can decrease ultimatum game

acceptance rates. This result may be explained as a priming of the emotional System 1 and possibly of

the anterior insula. Investigations of the volitional System 2 involved in ultimatum game responder

decisions have also been done. Van‘t Wout et al. (2005) and Knoch et al. (2006) used transcranial

magnetic stimulation (TMS) on the DLPFC and found that TMS increased ultimatum game acceptance

rates. However, the nature of TMS makes it unclear whether the TMS in these studies increased or

decreased the functioning of the DLPFC. Despite the uncertainty associated with the affect of the TMS

on DLPFC functioning, these studies demonstrated that there is a causal link between the functioning

of the DLPFC and responder decisions during the ultimatum game.

Experiment 1 – Working Memory Easy Study

Rationale

To further investigate the role of the volitional system in the ultimatum game responder

decisions the current experiment determined if activating the DLPFC, a brain area associated with

System 2 decisions, with a behavioral task would change the decision processing in the volitional

System 2 during ultimatum game decisions. In order to activate the DLPFC a task needed to be found

that could reliably activate this brain area. The DLPFC has been found to be reliably activated by many

working memory tasks (Casey et al., 1998, Courtney et al., 1998, Sakai, Rowe, & Passingham, 2002).

Therefore, the studies described here used a working memory task that was adapted from Sakai, Rowe,

and Passingham (2002) to load the DLPFC. Adding a visual working memory task to the ultimatum

game paradigm will allow the observation of any changes in ultimatum game acceptance rates due to

behavioral loading of the DLPFC. Any change in ultimatum game acceptance rates would be an

indication of a change in the rational System 2 processes for the ultimatum game responders.

Experiment 1 – Methods

Participants

A total of 72 participants (43 female and 29 male) were recruited from the pool of psychology

undergraduate students at the University of Arizona (age 18-51 years, M = 20.75, SD±5.02). Each was

randomly assigned to one of two experimental conditions (Control, N=35; Easy Working Memory,

N=37). To ensure that participants were sufficiently motivated to make real decisions, they were paid

10% of their actual earnings in the UG task, so that most participants received between $2 and $7 in

cash, in addition to 2 course credits (1 hour worth). For most participants this class credit represented

about a fifth of the total experimental credits that were required in the Introductory Psychology class or

some extra credit for higher level classes.

Procedure

Dot task. In the first preliminary experiment in this set of studies the experimental participants

were asked to do a simple working memory task, i.e., Easy Working Memory condition, where they

had to remember the location of three black dots in a 3 x 3 grid. Participants in the Control condition

had no requirement to memorize the presented dots. After the memory stimuli presentation each

participant received a varied set of ultimatum game offers, 4 offers of $5, 2 of $4, 4 of $3, 5 of $2, and

5 of $1, and decided to accept or reject them. To test the extent to which working memory was loaded,

each ultimatum game trial was followed by a test where the participant was asked to pick, out of four

possible patterns, the pattern of shapes they saw before that ultimatum game trial (See Figure-1 for a

timeline).

Ultimatum Game Task. Participants were first given a description of the ultimatum game. These

instructions emphasized that each trial will be independent and that their decisions would not affect

future offers. Participants were reminded that would be compensated in proportion to the money they

earned during the experiment. Participants were told that they were playing as ultimatum game

responders against proposers located at another university who were free to offer any split of $10 they

would like. After the participant indicated that they understood the rules of the game they were given

two test trials and again offered a chance to clear up any confusion.

Participants then played the game in the role of responder, receiving one-time monetary offers

from 20 different human proposers. The entire experimental task therefore consisted of a single block

of these 20 offers, each involving a $10 split. Participants in the Easy Working Memory condition (see

above) were given additional instructions. On each trial, participants first saw the memory test stimulus

(see above), followed by a picture of their partner for that trial. Following the presentation of the

picture, participants then saw the proposer’s offer and had a maximum of 20 seconds to decide to either

accept or reject. Upon deciding by way of a key press, the outcome of the trial was presented for 5

seconds, and the next offer sequence followed. All participants saw the same set of offers that differed

in unfairness level, with the trials presented in a randomized order for each. Participants saw a varied

set of offers, 4 offers of $5, 2x$4, 4x$3, 5x$2, and 5x$1 (see Figure 1).

Experiment 1 – Results

Memory Task. As expected, the memory test proved quite easy, with an error rate of 5.54%.

This accuracy rate was significantly better than chance performance, t(36) = 60.391, p < 0.001.

Decision-Making. The percentage of offers accepted for each type of offer ($1-$5) was

calculated for each participant. A repeated measures ANOVA was performed with one between-subject

factor (Control, WM) and one within-subject factor with five levels (acceptance rates for offers of $5,

$4, $3, $2, and $1). There were significant main effects for both experimental condition, F(1,70) =

5.54, p = 0.021, and offer amount, F(4,70) = 85.38, p < 0.001. Additionally, there was a significant

interaction between these factors, F(4,70) = 4.133, p < 0.05 (see Figure-2).

One concern that has yet to be addressed in this analysis is the possibility that participants may

change their willingness to accept ultimatum game offers over the course of the study. To investigate

this possibility a correlation was done between order of decisions and of average ultimatum game

acceptance rates. There were no order effects observed in the correlation between trial order of the

decision and average ultimatum game acceptance rates for either the Easy Working Memory condition,

r(18) = -0.04, not significant or for the Control condition, r(18) = 0.12, not significant.

Figure-1. Time line for Experiment 1 & 2. Each trial started with the presentation of three dots/shapes
(750ms dot/shape presentation, 250ms delay between). The participants then saw a picture of the
person proposing a split of $10. They then saw the proposed split and accepted or rejected that offer.
The outcome of the decision was then confirmed. Participants then had to recall the pattern made by the
three dots/shapes out of four possible patterns. Then they were informed if their memory choice was
correct or incorrect.

Figure-2. The difference between Easy Working Memory and Control conditions. Shows the average
percent of ultimatum game acceptance rates for each of the possible offers for participants who were
loaded with a simple working memory task and participants who were not loaded with a working
memory task. Error bars show ± SEM.

Experiment 1 – Discussion

These results can be interpreted as evidence that the easy working memory task primed the

DLPFC resulting in the strengthening of the volitional System 2, i.e., an increase in ultimatum game

acceptance rates. There also doesn’t seem to be any order effects of note in this between-subject

ultimatum game paradigm.

In the experiment above a working memory load was imposed on the DLPFC, with a visual

working memory task, during the ultimatum game trials. To understand the possible implications of this

loading, an understanding of the theoretical basis for cognitive loading is required. Loading is the

activation of attentional or memory systems during the task of interest. An example of loading was

observed by Lavie (2004) who reported that increasing working memory load, i.e., remembering digit

order, can increase the detrimental effects of distractors during an attention task. Lavie (2005) also

reported that increasing attentional load, i.e., identifying the number of syllables in a word vs.

identifying the case, can decrease these distractor effects.

Inherent in the theory of cognitive loading is the possibility that too much loading, i.e.,

overloading, can result in decreased functioning of the cognitive system. For example, overloading was

observed by Shiv and Nowlis (2004) who found that when participants trying to keep a relatively large

list of numbers in their working memory were more likely to pick unhealthy treats over healthy

options. They explained this result by postulating that both the participant’s working memory and

volitional systems were overloaded leaving the decision to the emotional brain. Therefore, actively

loading the volitional part of the brain, i.e., DLPFC, may be used to either promote, i.e., prime, or

disrupt, i.e., overload, System 2. Experiment 2 will investigate the possibility of overloading System 2

with a difficult working memory task.

In order to investigate whether loading System 2 is responsible for the promotion of volitional

decision processes and not simply a result of the overlap between the decision and working memory

systems, it needed to be determined if we find these relationships without active loading. To discover

these effects of loading an appropriate control is needed and harder working memory task should be

used so the effects of loading the DLPFC can be distinguished from other effects due to the particular

ultimatum game paradigm that was used. This need for a good control could possibility be filled by a

priming condition. Priming is the focusing of attention by activating psychological systems BEFORE

presenting stimuli. For example, Bower, Gilligan, and Monteiro (1981) primed sad or happy moods by

having people think about happy or sad events. Mood congruent details of a story they read after

priming were remembered better than mood incongruent details. A study that examined the effects of

either loading and priming the working memory system during ultimatum game trials would allow us

to distinguish the specific effects loading has on System 2 functioning. A comparison between a

priming condition and loading condition is done in Experiment 5 below.

Experiment 2 – Working Memory Hard Study

Rationale

In the Experiment 2 participants had to complete a working memory task that loaded the

DLPFC but was designed to be more difficult then the memory task in Experiment 1. It was thought

that a more difficult task may overload the DLPFC and we may find a decrease in ultimatum game

acceptance rates.

Experiment 2 Methods

Participants

A total of 39 participants (22 female and 17 male) were recruited from the pool of psychology

undergraduate students at the University of Arizona (age 18-26 years, M = 19.38, SD±1.72).

Participants were paid 10% of their actual earnings in the UG task, in addition to 2 course credits (1

hour worth). For most participants this represented about a fifth of the total experimental credits that

were required in the introductory psychology class or some extra credit for higher level classes.

Hard Dot Task. In this task the participants where asked to remember the location, in a 3 x 3

grid, shape (circle, square, and triangle) and color (yellow, red, and blue) of three shapes. As in the

working memory easy study, i.e., Experiment 1, participants had to recall the pattern of shapes, out of

four possible patterns that differed on one dimension, they saw before an ultimatum game trial after the

trial.

Ultimatum Game Task. The participants received the same task and instructions as in

Experiment 1 with the exception that the memory patterns the participants had to remember consisted

of Hard Dot Task stimuli as described above.

Experiment 2 Results

Participants in this experiment did not have significantly different ultimatum game acceptance

rates than either the Control or Easy Working Memory conditions. However, it was found the number

of working memory errors in the Hard Working Memory condition had a significant negative

correlation, r(37) = -0.44, p < 0.05, with ultimatum game acceptance rates (Figure-3).

One concern that has yet to be addressed in this analysis is the possibility that participants may
change their willingness to accept ultimatum game offers over the course of the study. To investigate
this possibility a correlation was done between order of decisions and of average ultimatum game

acceptance rates. Their is a positive trend in the relationship between order of decisions and of average

ultimatum game acceptance rates, r(18) = -0.35, p = 0.13. Fisher’s r to z transformation comparisons

show that there is no difference between the order effects correlation in Experiment 1 and Experiment

2, z = 0.95, not significant.

Figure-3. Correlation of ultimatum game acceptance rates and working memory task error rates. This
figure shows the significant negative correlation (r = -0.44) between a participant’s working memory
error rate (y-axis) and ultimatum game acceptance rates (x-axis).

Experiment 2 Discussion

The correlation between ultimatum game acceptance rates and memory task performance may

have been the result of the harder working memory tasks overloading some of the participants

volitional System 2 decreasing their ultimatum game acceptance rates. The other participants did not

have had their volitional System 2 overloaded and the memory task primed or had no effect on their

deliberations. In support of this explanation it was also found unfair offer rejections were more likely

on the trials where the participant had an error on the working memory task. There is also a small non-

significant tread suggesting that participant’s threshold for unfairness decreases over the progression of

the experiment.

During this experiment independent measures of working memory were not collected. In fact

only measure of working memory performance gathered was the participant’s working memory task

error rate. This error rate measurement if a fairly crude working memory because a participant can do

well through partial recognition and there is a 25% chance of simply guessing the correct pattern. The

distinction between the difficulty of the working memory tasks used in the Experiment 1 and

Experiment 2 is also crude. A working memory task that can vary on difficulty in a predictable way

would allow for a clearer working memory difficulty comparison. This possibility was addressed in

Experiment 3.

It is also unclear if loading working memory during an ultimatum game trial enables an accurate

independently verifiable assessment of working memory or the other executive functions, which are

thought to be indicators of the rational and volitional mind. It should also be noted that the experiments

above only included visual working memory tasks so there would be no verbal working memory

interaction with reading during the ultimatum game task. It would be useful to get independent

measures of verbal working memory and other measures of executive functioning to identify the extent

of their relationships with System 2. This analysis could help outline the cognitive components of the

System 2, which is thought to be involved in ultimatum game decisions. In order to do this analysis

executive function measures were taken in the experiments described below.

In the series of studies described above feedback was given to participants on each trial about

their working memory performance on that trial. It is possible that the participants who were being told

that they were getting many memory trials wrong reacted by being put in a sad mood. Harle and Sanfey

(2007) showed sad moods can decrease ultimatum game acceptance rates suggesting an alternative

explanation of the results in Experiment 1 & 2. Participants who did worse on the memory task

received more negative feedback possibly resulting in a sad mood, which may have resulted in

decreased ultimatum game acceptance rates. The No Feedback condition in Experiment 3 will explore

the plausibility of this emotional explanation for the results of Experiment 1 & 2.

It is also informative to investigate the possibility and the extent to which getting incongruent

feedback, i.e., taking control of the feedback away, on memory tasks might have on ultimatum game

decisions (W. J., Jacobs, personal communication, May 27, 2009). These tests can be done by

manipulating the working memory feedback the participants receive and observe any changes in their

ultimatum game acceptance rates. Experiment 4 investigates the effect that lack of control can have on

ultimatum game decisions.

As mentioned above in the discussion of Experiment 1 it is necessary to control for the

possibility that the effects observed is due to loading of the working memory system as opposed to

solely being a result of the overlap, i.e., non-causal co-variation, of decision and memory systems.

Experiment 5 tested this possibility by having a condition where the working memory system was

actively loaded during the ultimatum game and another condition where the working memory task is

completed before the ultimatum game trial. This experiment allowed the discovery of the effects of

overloading the working memory system on ultimatum game responder decisions.

Experiment 3 – Within Subject Variable Memory Study

Rationale

This experiment will address many of the major limitations Experiment 1 & 2 described above.

First, in order to determine the nature of the within participant relationship between working memory

load and ultimatum game decisions participants were asked to remember visual patterns of varying

difficulty during ultimatum game trials. If the relationship observed in the easy working memory

condition applies to this new paradigm there’s an expectation that there will have relatively high

acceptance rates for trials where participants have an easy working memory task, i.e., they only had to

remember 2-3 dots. The fact that each participant experienced a range of memory task difficulty also

allowed an assessment of the concomitant variation between ultimatum game decisions and the

difficulty of the working memory load. If there is a complicated relationship between these two factors

the variation of task difficulty will allow analysis of this complexity. An example of a possible complex

relationship is that working memory load may increase ultimatum game acceptance rates with low

difficulty but it decreases ultimatum game acceptance rates with high difficulty.

The second major limitation of Experiments 1 & 2 that will be addressed in this experiment is to

find the extent of the relationship between spatial working memory, verbal working memory and

executive functioning systems with the responder decisions made during the ultimatum game. In

Experiment 2 it was observed that acceptance rates had a moderate positive correlation between

working memory performance and ultimatum game acceptance rates so this type of relationship was

expected for all of the executive function measures because these factors have been found to have

significant overlap (Miyake et al., 2000). However, since Miyake et al. (2000) also demonstrated that

there is significant separation between the different executive functions the possibility of distinct

relationships between executive functions and ultimatum game acceptance rates was always a

possibility. Either way this experiment was designed to assess the extent of the overlap between the

executive function systems, i.e., working memory (the visual pattern memory task), inhibition (Stop

Task), and switching (Global-Local), and the volitional System 2 (ultimatum game acceptance rates).

Baddeley et al. (1998) also reported that there is a separation between verbal and spatial working

memory. To observe a possible distinction between the relationship of these two types of working

memory and ultimatum game acceptance rates an independent measure of verbal working memory, i.e.,

N-back, was also taken.

Another limitation of Experiment 1 & 2 this experiment will address is the possibility that

getting feedback from the memory task emotionally primes ultimatum game decisions. To investigate

this possibility participants will either receive memory feedback or will not receive memory feedback.

This hypothetical emotional priming may also interact with the loading of the working memory system

or change the nature of the relationship between ultimatum game acceptance rates and executive

functioning so measures of executive functioning will be taken in all conditions. If a difference

between the feedback and no feedback conditions were found it would suggest that emotions as a result

of memory task feedback are partially responsible for the effects observed in Experiment 1 & 2.

Finding a relationship between these two conditions would also underscore the role that emotional

processes have in ultimatum game decisions. However, if no difference is found between the No

Feedback and Correct Feedback conditions it would confirm that the ultimatum game acceptance rates

seen in Experiment 1 & 2 were solely due to effects of manipulating the volitional System 2 and not a

artifact of unintentional priming of the emotional System 1.

Experiment 3 – Methods

Participants

In the Correct Feedback condition there were Thirty-Six participants (20 Female and 16 Male).

In the No Feedback condition there were Forty-six participants (30 Female and 16 Male). All

participants were recruited from the University of Arizona online subject pool. They were paid 10% of

the money they earned during their experience in addition to three (1.5 hours worth) class credits they

received. In general this represented about a third of the required experimental participation

participants were required to do in order to pass the class most of them were enrolled in. There may

have also been students who were participating in studies in order to get extra credit in non-

introductory level classes.

Procedures

Visual digit span memory task. A the start of each trial, participants were presented with a series

(2-6 dots) of black dots within a 3 x 3 grid (750ms dot/shape presentation, 250ms delay). Following

each ultimatum game trial participants were asked to replicate the patterns of dots they saw before that

ultimatum game trial on a keyboard, i.e., 3 x 3 grid, consisting of r-t-y in the first row, f-g-h in the

second row, and v-b-n in the third row (Figure-4). Preceding the recall of the visual pattern participants

saw a two second presentation of a feedback screen. In the Correct Feedback condition the feedback

screen showed a message in the middle of the screen indicating they were correct, i.e., Correct, if the

pattern they entered matched the pattern they saw, or if the pattern they entered did not match the

pattern they saw a message indicating they were incorrect, i.e., Incongruent. All participants saw a

randomized set of memory patterns that were of equal complexity. Each participant saw 5 memory

patterns of 2, 3, 4, 5, and 6 dots in length presented in a random order. Each memory pattern difficulty

was paired with each offer amount. This task has been given the title of the visual digit span test. The

computer program used for this task was adapted from a working memory task developed by

Schneider, Eschman, and Zuccolotto (2002)

Ultimatum game task. This task used the same paradigm as in Experiment 1 with some minor

exceptions (see Figure-5 for a timeline). A computerized version of the ultimatum game programmed in

E-prime experimental software was used. Each trial started with the working memory task presentation

followed by a picture of a proposer and then the proposal of how to split the $10. Proposer pictures

were taken from a pool of pictures including the proposer pictures used in Sanfey et al. (2003a) or from

similar experiments where participants had given their consent to use their pictures. Participants saw a

varied set of offers, 5 offers of each offer level, $5, $4, $3, $2, and $1 presented in a random order. All

participants saw the same set of offers that represented all possible combinations of offer amounts and

working memory difficulty. So five memory difficulty levels multiplied by five offer levels created

twenty-five trials for each participant. After the participants decision they were given a chance (max 5

seconds) to accept or reject the offer. The outcome of the trial was then presented for 5 seconds. In the

very unlikely event they did not make a decision within 5 seconds they would have receive a message

stating they received $0 and the other player received the amount they chose to keep. After the outcome

the participants were tested on the working memory task as described above. Then after the completion

of the ultimatum game trials participants completed in a randomized order three executive functioning

measures, i.e., N-back, Global-Local, and Stop Task described below.

Figure-4. Memory pattern entry key. This figure shows the picture participants saw to tell them how to
enter the pattern of dots for the visual memory task. So if the participant saw a dot in the upper right
and then the upper left they would then have to type the Y key, R key, and finally the Enter key to
submit their remembered dot pattern.

Figure-5

Figure-5. Experiment 1 timeline. Each trial began with 2-6 dots presented (750ms dot presentation
followed by 250ms delay) to the participants on a computer screen at different locations in a 3 x 3 grid.
They then saw a picture of the proposer. Participants then saw the proposal and were given five seconds
to either accept or reject it. A two second ultimatum game outcome screen then showed the participants
the result of the ultimatum game trial. Participants then had to replicate the pattern of dots they saw
prior to the ultimatum game trial on a 3 x 3 portion of the keyboard (r-t-y first row, f-g-h second row,
and v-b-n third row) They then saw a two second outcome of the memory task showing the participants
whether they got the memory task correct or incorrect.

Independent Verbal Working Memory Task (N-back). To get an independent measure of verbal

working memory ability participants were asked to do an N-back task. This e-prime delivered task was

adapted from Gold et al. (2006). Since its development by Kirchner (1958), the N-back task has

become a standard measure of working memory and is often used to assess the strength of working

memory systems (Miyake et al., 2000). A meta-analysis of fMRI studies that used N-back tasks shows

that this task activates the DLPFC, the ventral lateral prefrontal cortex, lateral premotor, and dorsal

cingulate cortex (Owen, McMillan, Laird, & Bullmore, 2005).

The N-back task used in this experiment consisted of four sessions. Each session was preceded

with instruction clearly explaining that participants were expected to hit a target button (‘1’) or a non-

target button (‘0’) when letters are presented on the screen (See Figure-6 for the instructions given

during the Three-Back session). Before the first session participants were given a chance to practice

this activity on 17 practice trials. They were then given feedback showing the percentage of correct

target hits and asked if they would like to repeat the practice trials. During the Zero-Back session the

participants were asked to hit the target button when an ‘X’ is displayed and the non-target button when

any other letter is presented. The One-Back session asked the participants to hit the target button if the

previous letter matches the current letter so the participants have to remember the letter that precedes

the current letter in order to respond correctly. The Two-Back session and Three-Back session were the

same as the One-Back except the participants were asked to hit the target if the current letter

presentation matches the letter presented two trials ago or three trials ago respectively. See Figure-6 for

example instructions adapted from the Two-Back session. During each session 30 letters are presented

(10 targets) for 500ms where the participants are open to respond with the target buttons. Each letter

presentation was followed by a 3000ms break. A total correct memory hit rate was then calculated as a

measure of how good participants were at remembering and updating information. Participants who

had a high hit rate, i.e., targets hit/total target presented, are thought to have a high ability to update

information.

Figure-6. N-back instructions. Shows an example N-back timeline and instructional picture for the third
session task. Each letter is presented for 500ms followed by a 3000ms break.

Executive Functioning Switching Task (Global-Local). To get an executive function measure of

a participants ability to switch between tasks a Global-Local task was given. Navon (1977) developed

this task to distinguish the visual processing of global and local objects. The Global-Local task has

come to be a standard way of measuring people’s ability to switch between cognitive tasks (Miyake et

al., 2000). The switching executive functioning tasks are associated with activation in the medial

prefontal cortex, DLPFC, and a possible hemispheric asymmetry in the processing of global vs. local

visual features (Martinez et al., 1997; Sylvester et al., 2003).

The Global-Local task used in this experiment began with a short introduction explaining to the

participants that their goal was to hit the key of the letter that was visible either at the global level, i.e.,

big letter, if the letters were blue or local level, i.e. small letters, if the letters were black (See Figure-7).

The participants were asked to answer as fast as possible. The participants were then given twelve

practice trials where they had to hit the key that corresponds the global letter if the figure was blue or

local letter if the figure was black. Example stimuli can be observed in Figure-7. If the participant was

comfortable with the task and indicated that they understood the task then the experimenter started the

experimental trials. However, if they indicated they were not comfortable they were given extra sets of

practice trials until they were comfortable with the task. The experimental trials consisted of twenty-

four randomized presentation of blue or back, S’s or H’s as described above. A switching cost was

calculated by subtracting the average reaction time for trials where the previous trial had the same color

and therefore the participant didn’t have to switch attention from the average reaction time for trials

where the participant did have to switch their attention from global to local or local to global, i.e., the

color switched. The difference between the reaction time for trials that required an attention switch and

the reaction time for trials that did not require an attention switch indicates the participants ability to

switch attention and is called “switching cost”. A low switching cost indicates a high ability to switch

between cognitive tasks.

S S
S S
SSSSSS
S S
S S

Figure-7. Example Global-Local Stimuli. This figure shows an example stimuli from the Global-Local
task. This is a black stimuli so the participant would be asked to report the local letter, in this case it
would be an S.

Executive Functioning Inhibition Task (Stop Task). To get a measure of the participant’s ability

to inhibit actions a Stop Task was used. Since the first use of the Stop Task by Vince (1948), the Stop

Task has become a standard way of measuring peoples ability to inhibit actions, especially in research

examining disorders like attention-deficit hyperactivity disorder (ADHD), which is defined by patients

inability to inhibit actions (Band, 1997; Miyake et al., 2000). Inhibition tasks have been shown to

activate the medial prefrontal cortex, the right inferior prefrontal cortex, and the left caudate (Rubia et

al., 1999).

The Stop Task that was used in this experiment began with instructions explaining to

participants the simple task of reporting whether a picture was on the right (hit ‘L’) or left (hit ‘A’) as

fast as possible. There were fifty-four of these practice trials. The participants then received new

instructions, which stated they were again to report whether a picture was on the right or left as

described above but if the picture turned into a stop sign they should not hit the button. There were 250

trials, there were 50 stop trials, in the experimental session with a break in the middle, after 125 trials.

The stop signal began at 300ms after the presentation of the picture stimuli and was adjusted up 50 ms

if the participant could not inhibit their response and down 50ms if the participant inhibited the

response successfully. A Stop Signal Reaction Time (SSRT) was calculated for each participant by

subtracting the average speed of the stop signal from the average reaction time for non-stop trials. The

SSRT measures how long it takes an inhibition signal to prevent behavior. A short SSRT is an

indication of a fast inhibition system.

BADS Self-Rating Dex Questionnaire. The executive function measures described above have

two major limitations. First, there is no measure that encompasses a general

executive function.

Second, there is no self assessment of executive function. To fill these needs each participant was given

the Self-Rating Dex Questionnaire portion of the Behavioral Assessment of Dysexecutive Syndrome

(BADS, See Appendix A). This questionnaire has been used as a quick measure of a participant’s self

assessment of executive functioning and therefore filled the need for this type of measurement

(Bodenburg & Dopslaff, 2008; Wilson et al., 1998). Higher values on this measure indicate lower

executive function.

Experiment 3 – Results

An one-way generalized linear model using a binary logistic distribution demonstrated that as

the memory difficulty increased the amount of memory correct decreased, Wald Chi-Square (4) =

388.234, p < 0.05 (Figure-8). There is a trend suggesting the participants in the Correct Feedback

condition may be able to remember visual patterns better then the participants in the No Feedback

condition, Wald Chi-Square (1) = 3.103, p = 0.078. There is also a trend for the interaction between

condition and memory difficulty, Wald Chi-Square (1) = 7.444, p = 0.114.

Figure-8. Memory ability vs. memory difficulty

in Experiment 3.

This figure shows that the
participants got less memory trials correct as the number of dots they had to remember increased. Error
bars are ± SEM.

A generalized linear model analysis was done using a binary logistic distribution to determine if

memory difficulty level, i.e., number of dots required to remember during that trial, amount of money

offered, or visual digit span performance could predict whether the participants accepted the ultimatum

game offers. This model also investigated the possibility that the presence of feedback could change

ultimatum game acceptance rates. There was a significant main effect showing the amount of money

offered increased acceptances, Wald Chi-Square (4) = 234.14, p < 0.05. There was no main effect for

either memory difficulty level, Wald Chi-Square (4) = 0.23, not significant, visual digit span

performance, i.e., getting that trials memory test correct/incorrect, Wald Chi-Square (1) = 0.56, not

significant, or feedback, i.e. being in the Correct Feedback or No Feedback condition, Wald Chi-Square

(1) = 0.38, not significant (Figure-9). There were no significant interactions between memory difficulty

and offer level, Wald Chi-Square (16) = 8.56, not significant, memory difficulty and visual digit span

performance Wald Chi-Square (4) = 0.85, not significant, visual digit span performance and money

offered, Wald Chi-Square (4) = 4.255, not significant, feedback and memory difficulty level, Wald Chi-

Square (8) = 2.129, not significant, feedback and visual digit span performance, Wald Chi-Square (2) =

1.748, not significant, or feedback and money offered, Wald Chi-Square (8) = 8.331, not significant.

See Figure-9 for a visualization of the Feedback group comparisons and offer level. See Figure-10 for a

visualization of the memory difficulty level and group comparisons. It should be noted that although

there was a small difference in the number of participants in the Correct Feedback condition and No

Feedback condition the variance observed in each group is comparable so there is no good statistical

reason to be concerned about spurious results due to this difference.

Figure-9. Ultimatum game acceptance rates vs. feedback condition. This figure shows ultimatum game
acceptance rates separated for each memory task feedback condition and money offer level interact to
predict ultimatum game acceptances. Error bars ± SEM.
A

Figure-10. Ultimatum game acceptance rates vs memory difficulty level. This figure shows average
ultimatum game acceptance rates in both the Correct Feedback condition (A) and the No Feedback
condition (B) for each offer level and memory difficulty level. Error bars are ± SEM.

The generalized linear model analysis above assumes that the difficulty of remembering 2 to 6

dots is the same for all participants. To create a measure that takes into account the perceived difficulty

rather than an assumed measure of difficulty, i.e., number of dots required to remember, the difference

between a participants average visual digit span performance and the average correct answer at each

memory level was calculated. This technique is described by Field (2009), p. 740 as “group mean

centering” of the data1. For example, if a participant had an overall average of 60% memory correct but

only got 20% of the 6 dot trials correct a relative -40% was recorded for the perceived memory correct

measure for the participant’s 6 dot trials. This perceived memory correct measure was correlated with

1 This type of analysis was inspired by Luke Chang at a lab meeting and consultation a day after the lab meeting during
the month of May 2009.

ultimatum game acceptance rate for that memory level. So if the participant mentioned above accepted

3 out of 5 ultimatum game offers during 6 dot trials, they would have a 60% acceptance rate paired

with a measurement of -40% perceived memory correct. A significant relationship between these

factors would suggest that the perceived difficulty of a memory trial changed the likelihood that the

participants would accept the ultimatum game offer. This analysis demonstrated no correlation between

perceived memory correct measure and ultimatum game acceptance rates in the Correct Feedback

condition, r(34) = -0.07, not significant. There was also no correlation between perceived memory

correct measure and ultimatum game acceptance rates in the No Feedback condition, r(44) = 0.043, not

significant. It should be noted this type of analysis inflates the degrees of freedom because each

participant produces five non-independent data points so the measured degrees of freedom has been

corrected to reflect this violation of the assumption of data independence. See Figure-11 for a

scatterplot showing the lack of a relationship between perceived memory correct and ultimatum game

acceptance rates in the Correct Feedback condition.

0.2

0.4

0.6

0.8

-1 -0.5 0 0.5 1
Perceived Memory Correct

A
cc

ep
ta

nc
e

R
at

Figure-11. The relationship between perceived memory difficulty and acceptance rates. This figure
shows a scatterplot demonstrating the non-existent relationship between perceived memory correct and
Ultimatum game acceptance rates in the Correct Feedback condition.

Another way to explore the possibility of a relationship between memory performance and

ultimatum game acceptance rates is to use a measure that summarizes the strength of this relationship

for each participant and to see if the combined summaries suggest an overall relationship. To fill this

need a beta value, i.e., correlation, between ultimatum game acceptance rates and memory performance

was calculated for each participant. These beta values represent the strength of this relationship for

each person (See Table-1 for and example). A large beta value would indicate that there is a strong

relationship between ultimatum game acceptance rates and memory performance, for that participant. A

one sampled t-test was performed to investigate the possibility of a significant general relationship

between memory performance on the visual digit span. This one sampled t-test, testing for the

possibility that these beta values are significantly different then 0, demonstrated no relationship

between ultimatum game acceptance rates and memory performance for the Correct Feedback

condition, t(35) = -1.453, not significant. A one sampled t-test analysis demonstrates no significant

relationship for the No Feedback condition, t(42) = 0.700, not significant. It should be noted that this is

a fairly crude measure of this relationship because a simple beta value may over or underestimate the

strength of the relationship for each person.

Data summery for participant 500

# of Dots % UG Accepted % Dot Task Correct

2 80 80

3 100 100

4 40 80

5 80 60

6 80 20

Table-1. Example table of % accepted and % correct for each memory difficulty level. This table shows
a summary of an example participant’s, i.e., Participant 500, ultimatum game acceptance rates and
correct memory rates for each memory level. A beta value can be calculated by correlating %
ultimatum game acceptance rates and % correct memory trials rates for each memory difficulty level.
Participant 500 had a beta value 0.06 suggesting that for this participant there was not a large
relationship between memory and

ultimatum game acceptance rates.

To prepare the executive function measures for analysis, some data had to be culled because of

participant non-compliance and computer malfunctions. Two data points were treated as missing in the

N-back data for the Correct Feedback condition. One of these participants was an extreme outlier, i.e.,

more than 3 standard deviations below the mean, suggesting that this participant did not take the N-

back task seriously. The N-back score of another participant was treated as missing because of a

computer malfunction making it impossible to recover their score. In the No Feedback condition three

participants also had their N-back data lost by computer malfunction. During the Stop Task some

participants who did not respond for an extended period of time had a average stop delay that was

negative, this made the calculation of the stop signal reaction time (SSRT) impossible for these

participants. For this reason these four participants in the Correct Feedback condition and five in the

No Feedback condition SSRT’s could not be included in the analyses.

In order to compare the relationships of executive functions and ultimatum game acceptance

rate to the significant positive correlation found in Experiment 2, correlations were done between these

measures. Ultimatum game acceptance rates positively correlated with some of the executive function

measures of verbal working memory ability, i.e., correct trials in the one-back task, r(32) = 0.206, not

significant, two-back task r(32) = 0.054, not significant, and r(32) = 0.370, p < 0.05, for the Correct

Feedback condition. Ultimatum game acceptance rates positively correlated with some of the executive

function measures of verbal working memory ability, i.e., correct trials in the one-back task, r(32) =

0.407, p < 0.01, two-back task r(32) = 0.187, not significant, and r(32) = 0.146, not significant, for the

No Feedback condition (See Figure-12). There is a strong trend in the relationship between

performance on the visual digit span task and ultimatum game acceptance rates, r(36) = 0.313, p =

0.063 (See Figure-13 for a scatterplot), for the Correct Feedback condition, however there was no trend

in this relationship, r(43) = 0.069, not significant, for the No Feedback condition. There were no

significant correlations between the of number of correct trials in performance on a executive function

measure of switching, i.e., switching cost, for either the Correct Feedback condition, r(36) = -0.06, not

significant, or the No Feedback condition, r(43) = 0.08, not significant. There were also no significant

correlations between an executive function measure of inhibition, i.e., SSRT, and ultimatum game

acceptance rates for either the Correct Feedback condition, r(32) = -0.11, not significant, or the No

Feedback condition, r(42) = 0.11, not significant.

Figure-12. The relationship between performance on N-back tasks and ultimatum game acceptance
rates. This scatterplot shows the general positive relationships between ultimatum game acceptance
rates and 1-Back (Red), 2-Back (Pink) and 3-Back (Blue) performance.

Figure-13. The relationship between visual digit span performance and acceptance rate in the Correct
Feedback condition. This figure shows a slight positive trend in the relationship between participant’s
ultimatum game acceptance rate and their performance on the visual digit span task for the Correct
Feedback condition.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

% Correct

%
U

G
A

cc
ep

te
d

To determine whether if the executive function measures were representing a cohesive general

executive function system in the Correct Feedback condition correlations were also done to examine

the relationship between these measure to each other (Table-2).

Table-2.

Executive Function Correlations

EF Measures Switching Cost N-back SSRT Visual Digit Span BADS

Switching Cost 1.000 0.030 0.082 0.148 0.2

N-back 1.000 -0.303 0.416* 0.050

SSRT 1.000 -0.105 -0.018

Visual Digit Span 1.000 0.037

BADS 1.000

Table-2. The correlations between executive function for the Correct Feedback condition. This figure
shows the correlations for the four executive function measures used in this study with the ultimatum
game acceptance rate in the Correct Feedback condition. * = p < 0.05.

To see if the executive function measures were representing a cohesive general executive

function system in the No Feedback condition correlations were also done to find the relationship

between these measure to each other (Table-3).

Table-3.

Executive Function Correlations .

EF Measures Switching Cost N-back SSRT Visual Digit Span BADS

Switching Cost 1.000 -0.167 -0.145 -0.225 0.290

N-back 1.000 -0.155 0.249 0.050

SSRT 1.000 -0.252 -0.018

Visual Digit Span 1.000 0.037

BADS 1.000

Table-3. The correlations between executive function for the No Feedback condition. shows the
correlations for the four executive function measures used in this study with the ultimatum game
acceptance rate in the No Feedback condition. * = p < 0.05.

It should be noted that because of the large number of tests represented in the these tables

caution should be taken so these relationships are not over interpreted. It is inherently dangerous to do

so many tests without regard to the increasing alpha each test represents. A more conservative way to

analyze this executive function data, which does not increase the risk to type I error to is to use a

general linear model to analyze all of these potential relationships at the same time. This type of model

allows a determination of the extent the different executive function measures can predict ultimatum

game acceptance rates and if these relationships change as a function of the condition, i.e., Correct

Feedback vs. No Feedback, the participants were in. A weakly significant positive relationship is found

between performance on a verbal working memory task, i.e., N-back, and ultimatum game acceptance

rates, one-back task, F(1,54) = 4.123, p < 0.05, two-back task F(1,54) = 0.202, not significant, and

three-back F(1,54) = 0.271, not significant. There is no significant relationship between switching cost

and ultimatum game acceptance rates, F(1,54) = 0.060, not significant. There is no significant

relationship between SSRTs and ultimatum game acceptance rates, F(1,54) = 1.119, not significant.

There is no significant relationship between ultimatum game acceptance rates and condition, i.e.,

Correct Feedback vs. No Feedback, F(1,54) = 1.697, not significant. There is no significant interactions

predicting ultimatum game acceptance rates between condition and BADS score, F(1,54) = 0.801, not

significant, visual digit span performance, F(1,54) = 0.881, N-back, F (1,54) = 0.198, not significant,

SSRT F(1,54) = 1.527, not significant, three-back F(1,54) = 0.111, not significant, or switching cost F

(1,54) = 0.172, not significant. There were trends towards interactions between condition and two-back

F(1,54) = 2.163, p = 0.15 and one-back F(1,54) = 2.568, p = 0.12.

acceptance rates. The No Feedback condition has a significant positive relationship, r(23) = 0.44, p <

0.05, and the Correct Feedback conditions has a positive trend, r(23) = 0.35, p = 0.09. Fisher’s r to z

transformation comparisons show that there is a possible difference, i.e., trend, between the order

effects correlation in Experiment 1 and the No Feedback condition, z = 1.59, p = 0.09. Fisher’s r to z

transformation comparisons show that there is a hint of a difference, i.e., slight trend, between the order

effects correlation in Experiment 1 and the Correct Feedback condition, z = 1.26, p = 0.20 (Figure-14).

Figure-14. Order effects for the Experiment 3 conditions and Experiment 1. This figure shows the
relative trial order best fit lines for ultimatum game acceptance rates in Experiment 1 (Green), No
Feedback condition (Blue) and Correct Feedback condition (Red).

It is possible that different ultimatum game paradigms prime participant’s willingness to accept

unfair offers. If this is so we should be able to see this difference by comparing the decisions

participants make on the very first trial. To test this possibility a chi square was done comparing the

percentage of acceptances over number of participants on the first trial between the control participants

in Experiment 1 (60%) with both the Correct Feedback condition (64%) chi-square (1) = 0.027, not

significant, and No Feedback condition (67%) chi-square (1) = 0.102, not significant.

Experiment 3 – Discussion

The pattern of correct responses in the visual digit span seems to indicate that the participants

were really trying and they did not simply ignore this part of the experiment. It was a valid concern that

the participants would not put a lot of effort into the memory task because they were not directly

compensated, i.e., they were not getting more money for trying their best. The significant decrease in

correct responses as a function of number of dots the participant had to remember also seems to

demonstrate that the visual digit span task provided a range of visual working memory difficulty for the

participants in this study.

Contrary to the hypothesis that memory tasks would promote or disrupt ultimatum game

acceptance rates, there was no difference in ultimatum game acceptance rates in either easy or hard

working memory trials for this within-subjects manipulation. At the two-dot level the visual digit span

task is very easy, as demonstrated by the high performance rates, but unlike the similar dot recognition

task in Experiment 1 the ultimatum game acceptance rate did not increase. This suggests that the

priming effect observed in Experiment 1 will only present in between-subject paradigms. The

possibility, as suggested by Camerer (2009), that this memory task would overload the DLPFC causing

a decrease in rational decision-making i.e. System 2, was also not supported. We do not find these

relationships in spite of the fact that there is evidence that the working memory system is being

overloaded, i.e., the correct trials decrease as a function of task difficulty. These results suggest that the

rational (System 2) is neither being overloaded or primed in this experiment.

It is possible that the lack of overloading or priming may be the result of a difference in the way

participants calibrate acceptable offers. The within-subject memory task, i.e., memory difficulty varies

on each trial, may induce a different fairness threshold than the Easy Working Memory task used in

Experiment 1. In other words, participants who experienced only easy memory trials set their threshold

for fairness in one place, and the participants receiving all levels of memory difficulty set their

threshold in a different place. It is possible fairness thresholds may have been set differently because in

Experiment 1 the participants received less negative feedback initially then the participants in this

experiment. The greater negative feedback or the experience of doing a harder task may have resulted

in different ultimatum game acceptance thresholds. This explanation might also explain why there are

no differences in ultimatum game acceptance rates between the Hard Working Memory and Control

conditions in Experiment 1 & 2.

Evidence that there is a difference in threshold calibration is observed in the comparison of the

trial order effects and the trend in initial unfairness thresholds ascertained for Experiment 1 and

Experiment 3. There were no trial order effects observed in Experiment 1 but there are indications that

participants became more rational, i.e., accepted more offers, as Experiment 3 progressed. This

suggests that participants in Experiment 1 had a consistent unfairness threshold where as the

participants in Experiment 3 had a changing unfairness threshold. These different types of thresholds

may have been due to the different mood profiles elicited by each paradigm, which may have changed

ultimatum game acceptance rates over the course of the experimental session or set these thresholds

differently initially (Harle & Sanfey, 2007). For example, the initial difficulty of the Experiment 3

memory task may have frustrated participants resulting in relatively low ultimatum game acceptance

rates at the start of the session. However, as the session progressed participants may have habituated to

this difficulty therefore decreasing this mood effect.

Despite the fact that there was no evidence for priming or overloading of the System 2, it was

found that an independent measurement of working memory ability, i.e., N-back, did have a positive

correlation with ultimatum game acceptance rates. It was also found that there was a strong positive

trend in the correlation between performance on the visual digit span task and ultimatum game

acceptance rates in the Correct Feedback condition but not for the No Feedback condition. This

suggests that verbal working memory ability in this experiment is a moderate predictor of rational

ultimatum game decision-making (Cohen, 1988). However, it was not found that other measures of

executive functioning, i.e. switching (switching cost) and inhibition (SSRT), are good predictors of

ultimatum game acceptance rates. This suggests that the relationship between executive functioning

and rational ultimatum game decision-making is limited to verbal working memory, i.e., updating

executive function. There were also no interactions between the presence of feedback and any of the

executive functions for predicting ultimatum game acceptance rates. This suggests the emotional

priming due to the presence of feedback in the memory condition does not significantly change the

relationship between independent measures of updating executive function and System 2 decision-

making.

Furthermore, using group mean centering to try to model how each level of difficulty feels to

each participant, i.e., creating a perceived memory difficulty measure, does not change the relationships

between ultimatum game acceptance rate and performance on the visual digit span. There is no

significant relationships between any measure of memory task difficulty and ultimatum game

acceptance rates for either the Correct Feedback condition or the No Feedback condition. This suggests

that participants experience of difficulty of the memory task did not change ultimatum game

acceptance rates.

It should be noted that if less complex statistical analysis, similar to the analysis used in

Experiment 1 & 2, there seems to be indications of more complex relationships between these factors.

For example, there is a strong tread for a positive relationship between visual digit span performance

and ultimatum game acceptance rates, which seems to decrease when there is no memory task

feedback, i.e., No Feedback condition. This indicates that given a much larger data set an interaction

may have been found demonstrating that there is a small relationship between visual digit span

performance and ultimatum game acceptance rates but this relationship would be dependent on whether

feedback was given to the participants. However, it is hard to justify further investigation to find such a

small relationship and the current data supporting this conclusion doesn’t stand up to strict statistical

scrutiny.

These results can be reconciled with fMRI data which find that switching and inhibition tasks

activate areas of the brain that do not show activation during ultimatum game decisions. Switching

executive functioning is associated with medial prefrontal cortex (MPFC), which is not highly activated

during ultimatum game decisions (Sanfey et al., 2003; Sylvester et al., 2003). Inhibition executive

functioning is associated with activation in medial prefrontal cortex, right inferior prefrontal cortex,

and the left caudate none of which are highly activated during ultimatum game decisions (Rubia et al.,

1999; Sanfey et al., 2003). If brain activation allows us to see the important cognitive processes

involved in these executive functions, is not surprising that behavioral measures of these executive

functions also show no relationship with ultimatum game decisions.

Experiment 4 – Incongruent Feedback Manipulation Study

Rationale

This experiment was a result of an incorrectly coded ultimatum game task. In this task

participants saw visual digit span feedback from the previous trial instead of the trial they had just

completed. Instead of scraping this data altogether this data was scrutinized to see if it could tell us

anything about the ultimatum game decision-making process. It was realized that the behavior

produced by this task would produce a negative mood prime if the participants realized they were being

treated unfairly. In effect this condition treats the participants unfairly by taking direct control of the

memory feedback away from them.

Experiment 4 – Methods

Participants

Seventeen participants (8 Female and 9 Male) were recruited from the University of Arizona

online subject pool. They were paid 10% of the money they earned during their experience in addition

to the 3 class credits they received (1.5 hours worth).

Procedures

Visual digit span memory task. The visual digit span test as described in Experiment 3 was

used. However, in this Incongruent Feedback condition participants were told whether they were

correct or incorrect for the previous trial. This meant that participants who were good at the visual digit

span saw many correct feedback trials and participants who were relatively bad at the visual digit span

saw many incorrect feedback trials. However, the participants did not have direct control over what

kind of feedback they received because the feedback for that trial was not dependent on their

performance on that trial.

Ultimatum Game Task. The participants received the same task and instructions as described in

Experiment 3.

Experiment 4 Results

An one-way generalized linear model using a binary logistic distribution demonstrated that as

the memory difficulty, i.e., number of dots, increased the amount of memory correct decreased, Wald

Chi-Square (4) = 69.963, p < 0.05 (Figure-15).

Figure-15. Memory performance vs. memory difficulty for the Incongruent Feedback condition. This
figure shows that the participants got fewer memory trials correct as the number of dots they had to
remember increased. Error bars are ± SEM.

There is a significant negative correlation between a participants performance on the visual digit

span task and a participants overall ultimatum game acceptance rate, r(15) = -0.656, p < 0.05. See

Figure-16 for a scatterplot showing a comparison of different conditions in the strength of the

relationship between overall visual digit span performance and overall ultimatum game acceptance

rate.

Figure-16. Correlations between % correct and % accepted vs. conditions. This figure shows a
dissociation between the relationship of performance in the visual digit span task and overall ultimatum
game acceptance rates for different types of feedback. The red line shows the best fit line (r = 0.313)
for participants who had correct feedback. The purple line shows the best fit line (r = 0.06) for
participants who had no feedback. The blue line shows the best fit line (r = -0.656) for participants who
had incongruent feedback.

As in the other experiments each participant had a beta value calculated for a measure of a

participants individual relationship between memory task performance and ultimatum game acceptance

rates. A one sampled t-test analysis of these beta values demonstrates no significant relationship t(13) =

0.540, not significant.

Like the Experiment 2 correlations were done showing the relationships between some

executive function measures and ultimatum game acceptance rates. However, the data for the Stop

Task, N-back and Global-Local task were lost due to complications with the computer programs

processing this data incorrectly. Ultimatum game acceptance rates did not significantly correlate with

self perceived lack of executive function (BADS) r(11) = 0.176, not significant and also did not

significantly correlate with visual digit span performance, r(12) = -0.154, not significant.

As in Experiment 3 a general linear model was analyzed to see if the measures of executive

functioning or the type of feedback were significant predictors of ultimatum game acceptance rates.

There was a significant main effect for Feedback F(2,83) = 3.09, p < 0.05. There were no main effects

for either performance on the visual digit span, F(1,83) = 0.03, not significant, or BADS score F(1,83)

= 0.06, not significant. There was no interaction between Feedback and the BADS F(2,83) = 1.42, not

significant, but there was a trend for the interaction between feedback and visual digit span

performance F(2,83) = 2.23, p = 0.11. There were large difference between the condition, i.e., No

Feedback, Correct Feedback and, Incongruent Feedback, sizes so group comparisons are subject to

greater scrutiny. However, it should also be noted that parametric test are robust to violations of the

assumptions and as a result we can probability trust these results (Bradley, 1978). A graph of the

average ultimatum game acceptance rates can be seen in Figure-17.

As in Experiment 3 a concern that has yet to be addressed is the possibility that participants may

change their willingness to accept ultimatum game offers over the course of the study. To investigate
this possibility a correlation was done between order of decisions and of average ultimatum game

acceptance rates. Their was a very small non-significant positive relationship between trial order and

ultimatum game acceptance rates, r(23) = 0.276, not significant. Fisher’s r to z transformation

comparisons show that there is no significant difference between the order effects correlation in

Experiment 1 and Experiment 4, z = 1.0, p = 0.317 or between Experiment 3 and Experiment 4, z =

0.28.

It is possible that different ultimatum game paradigms prime participant’s willingness to accept
unfair offers. If this is so we should be able to see this difference by comparing the decisions
participants make on the very first trial. To test this possibility a chi square was done comparing the
percentage of acceptances over number of participants on the first trial between the control participants

in Experiment 1 (60%), with Incongruent Feedback condition (50%), chi-square (1) = 0.115.

Figure-17. Ultimatum game acceptance rates vs. all feedback conditions. This figure shows the average
ultimatum game acceptance rates for each money offer level separated by the Correct Feedback, No
Feedback and Incongruent Feedback conditions. Error bars show ± SEM.

A Fisher’s r-z transformation showed a significant difference between the correlations for

ultimatum game acceptance rates and performance on the visual digit span in the Correct Feedback

condition and Incongruent Feedback condition z = 3.51, p < 0.05.

Experiment 4 Discussion

Like Experiment 3 it was demonstrated that correct memory trials decreased as the visual digit

span got more difficult. This finding is important because it demonstrates that participants did not give

up on this memory task even after becoming aware of their inability to directly control the memory

feedback. Participants seem to be highly motivated to try their best on the visual digit span task.

Not giving participant direct control over the feedback given for the visual digit span, seems to

induce a System 1 emotional prime caused by their experienced unfairness. This emotional prime

results in lower acceptance rates for the participants most likely to discover they were being treated

unfairly. So an unfairness emotional prime, i.e., treating participants unfairly, seems to be able to

override the tendency for participants with superior working memory to make rational ultimatum game

decisions and push them to make irrational ultimatum game decisions. Harle and Sanfey (2007)

showed a similar effect when they decreased acceptance rates by priming unpleasant moods with sad

movie clips.

Experiment 5 – Priming vs. Loading Study

Rationale

This experiment sought to find the differences in decision-making, if any, of working memory

primes and loads during an ultimatum game trial. If a visual working memory load effectively

positively primes the volitional System 2 as was observed in Experiment 1 then presumably there

should be similar ultimatum game acceptance rates for either a working memory load or a comparable

working memory prime. However, if the working memory load is needed to change the decision-

making process this experiment would elucidate the various decision effects of this loading. It would

also be instructive to see how loading affects the relationships between ultimatum game acceptance

rates and executive functions because it would allow us to know if the relationships that we have found

in the other experiments are dependent on the working memory load during the ultimatum game.

Experiment 5 Methods

Participants

Forty participants (15 Males and 25 Females) were recruited from the University of Arizona

online subject pool. Twenty participants were randomly assigned to either the Priming conditions or the

Loading condition. The participants were paid 10% of the money they earned during their experience in

addition to the 3 experimental participation credits they received (1.5 hours worth).

Loading Working Memory Task. In the Loading condition each trial began with a randomized

presentation of 2-6 dots in a 3 x 3 grid (see Table-4 for timeline). The working memory presentation

was followed by an ultimatum game trial as described in Experiment 3. These participants were then

tested on their memory by entering in the pattern seen in the working memory presentation on a 3 x 3

key grid (Figure-6). Each trial was then followed by 3 trials of the color matching task, each 5 seconds

long.

Priming Working Memory Task. In the Priming condition each trial began with a randomized

presentation of 2-6 dots in a 3 x 3 grid (see Table-4 for timeline). The working memory presentation

was followed by 3 trials of the color matching task. The participant was then tested on their memory by

entering in the pattern seen in the working memory presentation on a 3 x 3 grid on the computer

keyboard (Figure-6). After the memory task they then moved onto an ultimatum game trial as described

in Experiment 3.

Table-4.

Condition Timelines

Condition WM Presentation Loaded task WM Test Unloaded task

Loading condition Dot Presentation → UG Trial → Dot Test → Motor Task

Priming condition Dot Presentation → Motor Task → Dot Test → UG Trial

Table-4. Priming vs. loading timeline. In the Loading condition participants followed the timeline as
described in Figure 3 which will be followed by 3 trials (5 seconds) of the color matching motor task.
In the Priming condition the dot presentation was be followed by 3 trials of the color matching motor
task, the dot test and than an ultimatum game trial.

Simple Motor Task. This task required typing keys on the keyboard that correspond to a random

presentation of red, yellow, or blue squares presented on the screen. In effect this was very much like

the control trials in a Stroop task (Stroop, 1935). Each trial consisted of 3 color matchings (5 seconds

each). The square disappeared once the participant entered a response. This simple motor task was

intended as a control cognitive task with little working memory or emotional involvement but of the

same approximate length as an ultimatum game trial.

Ultimatum Game Task. The participants received the same task and instructions as in

Experiment 3 with the exception that the training trials were now followed by the motor task, i.e.,

Loading condition, or be preceded by the motor task, i.e., Priming condition.

Experiment 5 Results

An one-way generalized linear model using a binary logistic distribution demonstrated that as

the memory difficulty increased the amount of correct memory trials in the visual digit span task

decreased, Wald Chi-Square (4) = 169.637, p < 0.05 (Figure-18). There was also a main effect

demonstrating that participants in the Loading condition had significantly better memory for the visual

digit span then the participants in the Priming condition, Wald Chi-Square (1) = 9.449, p < 0.05

(Figure-17). There was no main effect for offer level, Wald Chi-Square (4) = 2.033, not significant.

There were no significant two-way interactions between memory difficulty and offer level, Wald Chi-

Square (16) = 9.045, not significant, memory difficulty and condition, Wald Chi-Square (4) = 3.306,

not significant, or condition and offer level, Wald Chi-Square (4) = 6.919, p = 0.14.

Figure-18. Memory performance vs. memory difficulty separated by Priming and Loading conditions.
This figure shows as the memory task gets harder participants get more trials incorrect separated by
experimental condition. Error bars are ±SEM.

A generalized linear model analysis was done using a binary logistic distribution to determine if
memory difficulty level, i.e., number of dots required to remember during that trial, amount of money

offered, if the participant got the visual digit span performance, or if having active working memory,

i.e., Loading condition, could predict whether the participants accepted the ultimatum game offer.

There was a significant main effect showing the amount of money offered increased acceptances, Wald

Chi-Square (4) = 111.954, p < 0.05. There was no main effects for memory difficulty level or visual

digit span performance, Wald Chi-Square (4) = 0.173, not significant and Wald Chi-Square (1) = 0.003,

not significant respectively. There was no significant difference between the acceptances of participants

in the Priming condition or participants in the Loading condition Wald Chi-Square (1) = 0.985, not

significant. There were also no significant interactions between memory difficulty and offer level, Wald

Chi-Square (16) = 4.022, not significant, memory difficulty and visual digit span performance Wald

Chi-Square (4) = 1.762, not significant, memory difficulty and cognitive load Wald Chi-Square (4) =

3.237, not significant, or visual digit span performance and offer level Wald Chi-Square (4) = 6.001,

not significant. However there was a significant interaction between visual digit span performance, i.e.,

correct or incorrect, and cognitive load Wald Chi-Square (1) = 9.019, p < 0.05, showing participants

who were both in the Loading condition and were going to get their memory trial incorrect had a higher

acceptance rate. See Figure-19 for a visualization of the interaction between memory performance and

condition and Figure-20 to see the Loading condition’s participants ultimatum game acceptance rates

split by memory correct and memory difficulty.

Figure-19. Bar graph of memory performance and condition vs. ultimatum game acceptance rate. This
figure demonstrates a higher acceptance rate for people in the Loading condition who were about to get
the memory trial wrong. Error bars are ±SEM.

Figure-20. Memory performance and memory difficulty vs. ultimatum game acceptance rates. This
figure shows the interaction between memory correct and cognitive load seems to peak at 4 dots
memory difficulty level and drop for easier or harder memory levels.

To investigate the patterns in the data within each of the conditions in this experiment separate

analysis was done on the data from each condition. For this analysis a generalized linear model analysis

was used a binary logistic distribution to determine if memory difficulty level, i.e., number of dots

required to remember during that trial, amount of money offered, if the participant got the visual digit

span performance could predict whether the participants accepted the offer or not. There was a

significant main effect showing the amount of money offered increased acceptances within the Priming

condition, Wald Chi-Square (4) = 54.537, p < 0.05 and Loading condition Wald Chi-Square (4) =

51.579, p < 0.05. There were no main effects for memory difficulty level within the Loading condition

Wald Chi-Square (4) = 0.464, not significant or Priming condition Wald Chi-Square (4) = 0.000, not

significant. There were also no main effects for visual digit span performance within the Loading

condition Wald Chi-Square (4) = 0.000, not significant or Priming condition Wald Chi-Square (4) =

0.000, not significant. There were no significant interactions between memory difficulty and offer level

for either the Priming condition, Wald Chi-Square (4) = 4.268, not significant, or Loading condition

Wald Chi-Square (4) = 4.105, not significant. There were no significant interactions between memory

difficulty and visual digit span performance for either the Priming condition, Wald Chi-Square (4) =

0.320, not significant, or the Loading condition Wald Chi-Square (4) = 5.894, not significant. Finally,

there were no significant interactions between visual digit span performance and offer level for either

the Priming condition, Wald Chi-Square (4) = 6.843, not significant, or the Loading condition Wald

Chi-Square (4) = 2.993, not significant. For a visualization of these ultimatum game acceptance rate of

the participants relationships in relation to memory difficulty and offer amount see Figure-21a for the

Loading condition and Figure-21b for the Priming condition.

Load (A)

Prime (B)

Figure-21. Memory difficulty and offer level vs. ultimatum game acceptance rates. This figure shows
ultimatum game acceptance rates split up by memory difficulty (Dots) and offer level for both
Loading(A) and Priming (B) conditions.

The generalized linear model analysis above assumes that the difficulty of remembering 2 to 6

dots is the same for all participants. To take in to account the difference in perceived difficulty the

difference between a participants overall average visual digit span performance and the average correct

of each memory level was calculated. This technique is described by Field (2009), p. 740 as “group

mean centering” of the data. This perceived memory correct measure was used as a predictor with offer

level to see if they can predict ultimatum game acceptances for that memory level. For the Loading

condition a very weak negative trend was found between perceived memory difficulty, i.e., percent of

correct trials for the five trials with a given dot memory requirement, and the ultimatum game

acceptance rate for that level of memory difficulty, r(18) = -0.256, not significant. The Priming

condition did not have a significant correlation between memory difficulty and ultimatum game

acceptance rates r(18) = 0.11, not significant. It should also be noted this type of analysis inflates the

degrees of freedom because each participant produces five non-independent data points so the

measured degrees of freedom has been corrected to reflect this violation of the assumption of data

independence.

As with the other experiments a single sample t-test was done on the beta values which

represent an individuals relationship between memory difficulty and ultimatum game acceptance rate.

The beta values did not significantly differ from zero for either the condition, t(19) = 0.086, not

significant or the Priming condition, t(19) = -0.525, not significant. This suggests that there is no

relationship between perceived memory difficulty and acceptance rates in this experiment.

To prepare the executive function measures to be analyzed some data had to be culled because

of participant non-compliance. Eleven data points in the Priming condition and three in the Loading

condition were treated as missing in the SSRT data. This was an extremely large defect rate. To ensure

that the people who did not follow instructions on the stop signal task didn’t come from a different

population an independent samples t-test (not assuming equal variances because of the large sample

size difference) was done comparing the ultimatum game acceptance rates for the non-compliers and

compliers. This test showed no significant difference between these groups t(24.155) = -0.133, not

significant. See Figure-22 for a bar graph showing these groups to be almost identical in respect to their

ultimatum game acceptance rates.

Figure-22. Bar graph of SSRT defectors and compliers vs. ultimatum game acceptance rates. This
figure shows the relative ultimatum game acceptance rates for people who had readable SSRT’s or not.
Error bars ±SEM.

As in Experiment 1 & 2 correlations were done showing the relationships between the

Executive Function measures and ultimatum game acceptance rates. Ultimatum game acceptance rates

were not found to be correlated with the executive function measure of updating ability, i.e., Correct

trials in a N-back task, in the Priming condition, one-back task r(18) = -0.420, p = 0.07, two-back task

r(18) = 0.038, not significant, and three-back task r(18) = 0.253, not significant. Ultimatum game

acceptance rates were not found to be correlated with the executive function measure of updating

ability, i.e., Correct trials in a N-back task, in the Loading condition, one-back task r(18) = -0.248, not

significant, two-back task r(18) = -0.008, not significant, and three-back task r(18) = 0.113, not

significant. There is a negative correlation in the relationship between performance on the visual digit

span task and ultimatum game acceptance rates, r(18) = -0.520, p < 0.05 for the Loading condition and

no significant correlation for the Priming condition r(18) = 0.250, not significant. There were no

significant correlations between the executive function measure of switching (switching cost) and

ultimatum game acceptance rates for the Loading condition, r(18) = 0.057, not significant or the

Priming condition r(18) = -0.052, not significant. There were also no correlations between an executive

function measure of inhibition (SSRT) and ultimatum game acceptance rates for the Loading condition,

r(15) = 0.091, not significant or the Priming condition, r(18) = 0.292, not significant. Finally there were

no significant correlations between the BADS score and ultimatum game acceptance rates for either the

Loading condition r(18) = 0.314, not significant or Priming condition r(18) = -0.056, not significant.

As mentioned above it is inherently dangerous to do so many tests without regard to the

increasing alpha each test represents. A more conservative way to analyze this executive function data,

which does not increase the risk to type I error to is to use a general linear model to analyze all of these

potential relationships at the same time. This type of model allows a determination of the extent the

different executive function measures can predict ultimatum game acceptance rates and if these

relationships change as a function of the condition, i.e., Priming vs. Loading conditions, the

participants were in. There are some significant relationships between performance on a verbal working

memory task, i.e., N-back, and ultimatum game acceptance rates, one-back F(1,27) = 6.682, p < 0.05,

not significant, two-back F(1,27) = 2.102, not significant and three-back F(1,27) = 2.715, not

significant. There is no significant relationship between switching cost and ultimatum game acceptance

rates, F(1,27) = 0.124, not significant. There is a significant relationship between ultimatum game

acceptance rate and visual digit span performance F(1,27) = 9.696, p < 0.05. There is no significant

relationship between SSRTs and ultimatum game acceptance rates, F(1,11) = 2.184, not significant.

There is no significant relationship between BADS score and ultimatum game acceptance rates, F(1,27)

= 0.775, not significant. There is a significant interaction predicting ultimatum game acceptance rates

between condition and visual digit span performance. There are no significant interactions predicting

ultimatum game acceptance rates between condition and BADS score, F(1,30) = 0.865, not significant,,

N-back, F(1,30) = 0.500, not significant, SSRT F(1,14) = 0.017, not significant, or switching cost

F(1,30) = 0.033, not significant.

As in Experiment 3 a concern that has yet to be addressed is the possibility that participants may
change their willingness to accept ultimatum game offers over the course of the study. To investigate
this possibility a correlation was done between order of decisions and of average ultimatum game

acceptance rates. Their was a large significant positive relationship between trial order and ultimatum

game acceptance rates for the Loading condition, r(23) = 0.722, p < 0.001, but only a minor trend for

the Priming condition, r(23) = 0.295, p = 0.15. Fisher’s r to z transformation comparisons show that

there is a significant difference between the order effects correlation in Experiment 1 and Experiment 4

for the Loading condition, z = 2.7, p < 0.05 but not for the Loading condition. There was also a robust

trend between Experiment 3 and Experiment 4 for the Loading condition, z = 1.81, p = 0.07, but not for

the Priming condition, z = 0.2, not significant.

percentage of acceptances over number of participants on the first trial between the control subjects in

Experiment 1 (60%) with both the Loading condition (55%), chi-square (1) = 0.035, not significant,

and the Priming condition (50%), chi-square (1) = 0.147, not significant.

Experiment 5 Discussion

It has been demonstrated that the visual digit span task did indeed increase in difficulty as the

memory task difficulty increased. Additionally, the participants in the Loading condition did better on

the visual digit span task than the participants in the Priming condition. This result was unexpected and

suggests that the color matching task required even more cognitive resources than the ultimatum game.

This could be a result of the color matching task being more intrusive than was expected.

The major focus of this study was to find differences in ultimatum game acceptance rates

between the Priming and Loading conditions. There was no main effect for this comparison, however

there was a significant interaction showing that in trials where the participants in the Loading condition

got the memory test incorrect they were more likely to accept an ultimatum game offer. The increase in

acceptance rate seems to happen more for memory difficulty levels that are of medium difficulty which

suggests that this effect is strongest when the memory tasks are neither too easy or too hard. For those

trials where the participants are getting the memory task incorrect and accepted the ultimatum game

offers it seems that the memory task in effect primes the rational decision of accepting unfair offers.

However, this finding is contrary to the hypothesis that working memory overloading will displace the

rational, i.e., System 2, decision maker.

The relationships between the executive function measures and ultimatum game acceptance

rates in the Priming or Loading conditions shed further light on the decision processes for the

ultimatum game. In the Loading condition minor relationships were found between indicators of

working memory and ultimatum game acceptance rates however, these relationships were in the

opposite direction of those in the comparable condition in Experiment 3. There was also a minor

relationship found between the BADS measure and ultimatum game acceptance rates suggesting that

participants in the Loading condition who had higher self assessed executive function had lower

ultimatum game acceptance rates. However, it should be noted that these relationships were too small

to reach significance when they are subjected to statistical scrutiny. Like the other experiments no

correlation between ultimatum game acceptance rate and measures of either switching or inhibition

executive functions were found. Additionally, the relationships that are found seem to be dependent on

the active loading of the working memory system during the ultimatum game trials. Few executive

function correlations for the participants in the Priming condition are observed. The one exception is a

trend for a negative relationship between ultimatum game acceptance rate and 1-Back performance,

which is similar to the results of the unfairness primed participants in Experiment 4.

The intended control task, i.e., Loading condition, for the Priming condition also lacked the

relationships observed in the other experiments putting into question whether finding of the differences

observed between the Priming condition and the other experiments can be trusted. Both the Priming

and Loading conditions added a color matching task to the paradigm used in the other

experiments.

It is

therefore possible that this added task may also explain the discrepancies found between the Priming

condition and the other experiments. If the correlation between percentage of memory trials correct on

the visual digit span task, and percentage of accepted ultimatum game offers for the Loading condition

is observed a strong negative correlation is found. This suggests this condition elicits similar behavior

patterns to Experiment 2, i.e., Incongruent Feedback condition, which also had a strong negative

relationship between these two factors. The addition of the color matching task may also be an

explanation for an unusually high number of Stop Task defects because it lengthened the ultimatum

game experimental session possibly decreasing participants motivation to participate in a task that

demands a fair amount of attention. Additionally, careful rechecking of the computer programs used in

the Loading condition revealed no reason why these participants would behave differently besides the

addition of the color matching task. At the very least the results of the Loading condition demonstrates

the fragility of the link between measures of working memory and ultimatum game acceptance rates

because a subtle addition of a simple task reversed the direction of this relationship. It might also

suggest that negative emotional priming is extremely easy for ultimatum game decisions.

As in Experiment 3 and 4 participants had a relatively low initial fairness threshold that

increased as the experiment progressed. In fact the Priming condition showed the greatest positive

relationship between average acceptance rates and order of decision. These findings suggest that this

pattern of ultimatum game responses is due to the within-subject paradigm which is used in all of these

experiments.

Overall Results

To compare the relative ultimatum game acceptance rates over all conditions a repeated

measures ANOVA was done to examine each participants acceptance rates for each offer level. This

analysis demonstrated there was a significant effect of offer level, F(4,216) = 213, p < 0.05. However,

there was only a trend for the comparison of the different conditions, F(7,216) = 1.755, p = 0.098 and

for the interaction of condition and offer level, F(28,216) = 1.312, p = 0.131. See Figure-23 for a graph

comparing these conditions.

Figure-23. Ultimatum game acceptance rates compared for all conditions. shows the average
acceptance rates for each offer level in each condition. Error bars ±SEM.

Overall Discussion

The argument for a two-system model for ultimatum game decisions began with the observation

that both the volitional rational brain areas, i.e., DLPFC, and emotional brain areas, i.e., anterior insula,

are active during these decisions (Sanfey et al., 2003). However, it should be noted that only the

activation of anterior insula varied with the ultimatum game acceptance rates in this study suggesting a

greater role in these decisions then other brain areas because of its concomitant variation. In accordance

with this theory Harle and Sanfey (2007) demonstrated priming negative emotions, i.e., System 1, can

also predictably decrease ultimatum game acceptance rates. Both of these studies provide clear

evidence for the importance of emotions in ultimatum game decisions. However, it remained unclear

how important rational deliberations that were presumed to activate the DLPFC were in making

ultimatum game decisions. To help answer this question both Van‘t Wout et al. (2005) and Knoch et al.

(2006) demonstrated that TMS on the DLPFC can increased ultimatum game acceptance rates making

it clear that this brain area is involved in these decisions and that there is a role for volitional rational

systems when describing ultimatum game decisions. Furthermore, during Experiment 1 it was found

that lightly loading the DLPFC with a task shown to activate this area of the brain also increases

ultimatum game acceptance rates. These findings invite a System 1 vs. System 2 explanation to

ultimatum game responder decisions and suggested that behavioral tasks or primes might be able to

enhance or disrupt volitional rational decisions as measured by ultimatum game acceptances.

One of the primary goals of this study was to discover the plausibility of using visual working

memory tasks, which have been shown to activate the DLPFC, to enhance and disrupt rational System

2 decisions, i.e., ultimatum game acceptances, in a within-subject paradigm. The experiments that were

conducted in this study demonstrate that it is unlikely that within-subject manipulations of memory can

be used to promote or disrupt rational and volitional System 2 decisions. However, it was also

demonstrated that the visual digit span task that was developed for this purpose did provide participants

with a task that challenged the spectrum of visual working memory abilities.

Another important goal of these studies was to determine whether a relationship exists between

System 2 motivated decisions and the executive functions. Experiment 3 and Experiment 2 suggest that

there is a moderate positive relationship between working memory ability and ultimatum game

acceptance rates, i.e., rational System 2 decisions. These finding indicate an important role for updating

executive function, i.e., working memory, in rational System 2 decisions. It should be noted that the

strength of this relationship is comparable to the relationships that are found between the different

executive functions with each other and thus it suggests that System 2 incorporates components of the

working memory system during ultimatum game decisions. However, no relationship was found

between ultimatum game acceptance rates and measures of either the switching, or inhibition executive

functions, at least not when a within-subject ultimatum game paradigm was used.

The positive relationship between working memory ability and ultimatum game acceptance

rates but not the other executive functions also seems to mirror the pattern of brain activation observed

in ultimatum game responders (Sanfey et al., 2003). Like the behavioral observations there is

congruence between the activation of areas of the brain associated with working memory and

activations responding to ultimatum game offers (Sakai, Rowe, & Passingham, 2002; Sanfey et al.,

2003). Additionally, few relationships are found between ultimatum game acceptance rates and

switching and inhibition executive functions during either neurophysiological, i.e., fMRI activations or

behavioral investigations. These findings suggest that fMRI activation predicted the patterns of

behavioral relationships found in this study.

Important exceptions are found to the general congruence of neurological and behavioral

relationships during the ultimatum game. One such exception is seen in the DLPFC, which has been

found to activate during many executive function tasks, i.e., switching and inhibition. However, we did

not find relationships between these executive functions and ultimatum game decisions. It should also

be noted that there are also brain areas that are associated with verbal working memory tasks, i.e., N-

back, which are not activated during the ultimatum game. These areas include: the ventral lateral

prefrontal cortex, lateral premotor, and dorsal cingulate cortex (Owen, McMillan, Laird, & Bullmore,

2005). This suggests that the lack of co-activation cannot always be taken as evidence of a lack of a

relationship. So fMRI activations give imperfect clues about the relationship between ultimatum game

acceptance rates and executive functions.

This pattern of executive function relationships with ultimatum game acceptance rate also seem

to parallel the findings of correlations between intelligence tests and the different executive functions.

Friedman et al. (2006) found that intelligence, like the ultimatum game acceptance rates measured in

this study, correlated with the executive function of updating, i.e., working memory, but not with the

executive functions of switching or inhibition. The similarity of the patterns for intelligence and

ultimatum game acceptance rates suggests that both of these measures may be tapping the same

cognitive system(s). If this hypothesis is true, than it should be expected that measures of intelligence

would correlate with ultimatum game acceptance rates. The best way to test this possibility would be to

do another study and correlate measures of intelligence and ultimatum game acceptance rates.

The similar but incongruent relationships between neurological imaging and behavioral

correlations in this study may also suggest that while there may be general systems that correspond to a

System 1 or System 2, participants utilize subsets of these systems during different kinds of decisions.

For example, when making decisions about whether to accept an unfair offer System 2 components

located in the DLPFC may be recruited but other areas associated with System 2, like the brain areas

associated with switching and inhibition, may be relatively inert. So there may be many two-system

circuits which are utilized during different types of decisions, like the ones used in moral decision-

making or delay discounting. However, this does not preclude the general distinction that can be made

between emotional and volitional brain areas.

An alternative explanation for the results in this study could indicate that the measures of

executive function that were used in this study were somehow deficient in their ability to capture the

general executive function system as described by Miyake et al. (2000). In support of this possibility

few of the executive function measures used in this study significantly correlated with each other,

which leaves open the possibility that an important part of the general executive function system was

not captured.

The general findings described above make the assumption that emotional involvement, i.e.,

System 1, was not a factor in the ultimatum game decision processes we measured. However, it should

be noted that there was often not sufficient control of possible emotional manipulations that were

embedded in the ultimatum game paradigms that were used. The No Feedback condition in Experiment

3 fixed the most blatant of these possible emotional manipulations by eliminating the visual digit span

feedback. The elimination of memory task feedback resulted in the relationship between acceptance

rates and independent measures of verbal working memory, i.e. 1-back, to drop, i.e., from r = 0.406 in

the Correct Feedback condition to r = 0.206 in the No Feedback condition. These findings suggest that

the memory feedback participants are given may be a very slight promoter of the observed positive

relationship between ultimatum game acceptance rates and performance on the 1-back. However, there

continued to be a relationship between working memory and ultimatum game acceptance rates

suggesting this connection is not entirely dependent on the memory feedback.

A strong positive trend was found between performance on the visual digit span task and

ultimatum game acceptance rates but only when participants were given true feedback on this memory

task. This trend disappears when no feedback is given suggesting that to the relationship between visual

digit span task and ultimatum game acceptance rates is dependent on the working memory feedback.

Furthermore, this trend seems to reverse when incongruent, i.e., unfair, feedback is given, i.e.,

Experiment 4, or when a simple color matching task is added, Experiment 5. This suggests that the

feedback given for the visual digit span memory task may be an important influence for the observed

trend in the relationship between visual digit span performance and ultimatum game acceptance rates.

It should be noted that this is the same relationship that was found in Experiment 2 so any conclusions

based on the relationship observed in Experiment 2 should be made with caution.

Another possible promoter of the relationship between the working memory system and

ultimatum game acceptance rate observed in Experiment 3 is the loading of the working memory

system during the ultimatum game trials. It has been shown that both the ultimatum game and visual

working memory tasks activate the DLPFC, suggesting the possibility that loading the working

memory system will promote, i.e. Experiment 1, or disrupt the ultimatum game decision-making

process (Sakai, Rowe, & Passingham, 2002; Sanfey et al., 2003). Experiment 5 manipulated the

requirement of actively loading the working memory system in order to test if taking the working

memory load off of the DLPFC would change the relationship between the executive function

measures and ultimatum game acceptance rates. In the Priming condition the relationships between

executive functions, i.e., N-backs, found in the other conditions disappeared or were reversed. This

suggests that indeed the active loading of the working memory system may be required to promote the

observed relationship between independent working memory measures, i.e., N-backs, and ultimatum

game acceptance rates i.e., Experiment 3. However, the intended control of this priming condition also

did not show these relationships. Instead both of the Experiment 5 conditions the patterns of decisions

patterns closely resembled those in Experiment 4 suggesting that these two experiments elicited a

similar mental state two each other and cannot be directly compared to the paradigms used in the other

experiments, i.e., Experiments 1, 2 & 3.

The complex nature of these findings suggest there are many variables determining a persons

willingness to accept unfair offers and how this willingness relates to other cognitive functions. The

general mood, i.e., priming of the anterior insula, of the person can decrease willingness to accept

unfair ultimatum game offers (Harle & Sanfey, 2007). Additionally, Experiment 1 demonstrated that

lightly loading the working memory system can increase peoples willingness to accept unfair

ultimatum game decisions. These two findings put together suggest a simple two-system model

because priming either emotions with movie clips or concentration on a working memory task changes

willingness to accept unfair ultimatum game offers.

This simple two-system model does not explain all of the results found in this study. We also

found differences in how willingness to accept ultimatum game offers changed over the course of the

experiment. In the between-subject experiments, i.e., Experiment 1 & 2, that used tasks, which held a

constant memory difficulty, participants either slightly decreased their willingness to accept ultimatum

game offers or it stayed the same as the experiment progressed. In contrast, during the within-subject

experiments, Experiments 3, 4, & 5, that used variable memory difficulty tasks, participants generally

increased their willingness to accept ultimatum game offers as the experiment progressed. In support of

the idea that within-subject designs create a distinct ultimatum game environment, Harle (Unpublished)

has also found that within-subject ultimatum game paradigms, that manipulate emotions, wash out

effects observed in similar between-subject designs. These findings suggest that the ultimatum game

strategy differs between between-subject and within-subject ultimatum game paradigms.

As mentioned above a possible framework that might help explain these convoluted results is a

cognitive structure with general System 1 and System 2 structures, both cognitive and anatomical, but

when challenged by different decision situations, i.e., decision paradigms, subsets of these systems are

recruited. For example, the finding that the different relationships between ultimatum game acceptance

rates and visual digit span performance is dependent on the kind of memory feedback that was given

during the paradigm may be due to recruitment of a slightly different subset of System 1 and System 2

areas. Paradigms that include loading of the working memory system during the ultimatum game have

not yet been done in fMRI so it is unclear where these subsets may be located but it is expected that

emotional areas would be relatively more recruited when ultimatum game acceptances are low and

volitional areas would be relatively more recruited when ultimatum game acceptances are high.

Findings like the fact memory can influence figure-ground perception also challenge the idea of

hierarchical, i.e., serially processed, and distinct cognitive system, suggesting a two-system circuit

model of decision-making may engender the flexibility that a strict and distinct two-system model

lacks.

This framework produces some implications for other cognitive theories. For example, one

interpretation of the Somatic Marker Hypothesis is that it demonstrates that emotional inputs can

support what was thought to be deliberative decision-making. However, as Sanfey et al. (2003b)

demonstrated, somatic markers reduce the tolerance to all high variance decisions and not just

disadvantageous choices. The ventromedial

prefrontal cortex may simply be a component of the general System 1 that may be conditionally

recruited during some decision paradigms, i.e. those involving tolerance to risk, but not others, like the

ultimatum game. Likewise, other two-system models like the facial expression systems or the high and

low road visual systems may be conditionally recruited during different decision-making paradigms.

It remains unclear if two-system theories that describe the cognitive processing of vision or

facial expression relate to the cognitive systems of decision-making. Its quite possible that these

systems would be recruited by either the volitional or emotional cognitive systems, but only when the

decision paradigm requires their input. However, the only way to sort out the interactions of these

cognitive systems is to find the relationships and circumstances that bind them together.

As the comparison of the different experimental condition in this study demonstrated, even

minor changes in the ultimatum game can produce different decision-making behavior. Therefore, the

use of within-subject manipulations of working memory as opposed to between-subject manipulations

seems to have significantly changed the decision-making process. Supporting this possibility there was

evidence suggesting there are important unfairness threshold differences between the between-subject

and the within-subject paradigms. In order to know if the findings of this study were not simply due to

artifacts of the within-subject paradigm the same memory manipulations would need to be done with a

between-subject design. It may be found that this differentiation is due to different two-system model

circuits being recruited.

Discovering the outline of the different cognitive systems may also help to explain preference

reversals. Preference reversals are situations where many participants change their decisions when they

are presented with the same options in different contexts. For example, Lichtenstein and Slovic (1971)

demonstrated that participants generally choose bets with a high probability of winning over bets with

large possible rewards. However, when these very same bets were evaluated on their individual worth,

the bets with large possible rewards were priced in general higher by participants than bets with a high

probability of winning. Choosing one type of bet and deciding the other bet is worth more demonstrates

a fundamental inconsistency in decision-making. It is possible that preference reversals like the one

described above may be explained by understanding the interaction of different cognitive systems. For

example, the relative involvement of System 1 and System 2 might explain why participants choose

one option in calculating value and another option when choosing between a selection of options.

Future studies should elucidate how the cognitive systems involved in decision-making interact to

produce the inconsistent decisions people can make.

97
Appendix A

BADS Self-Rating Dex Questionnaire
This questionnaire looks at some of the difficulties that people sometimes experience. Please
indicate how often you personally experience the following statements. Use the scale below and
write your answers in the spaces provided.

Never Occasionally Sometimes Often Very Often Frequently AlmostAlways

-3 -2 -1 0 +1 +2 +3

I have problems understanding what other people mean unless they keep things simple and
straightforward.

I act without thinking, doing the first thing that comes to mind.

I sometimes talk about events or details that never actually happened, but I believe did happen.

I have difficulty thinking ahead or planning for the future.

I sometimes get over-excited about things and can be a bit “over the top” at these times.

I get events mixed up with each other, and get confused about the correct order of events.

I have difficulty realizing the extent of my problems and am unrealistic about the future.

I am lethargic, or unenthusiastic about things.

I do or say embarrassing things when in the company of others.

I really want to do something one minute, but couldn’t care less about it the next.

I have difficulty showing emotion.

I lose my temper at the slightest thing.

I am unconcerned about how I should behave in certain situations.

I find it hard to stop repeating saying or doing things once I’ve started.

I tend to be very restless, and “can’t sit still” for any length of time.

I find it difficult to stop myself from doing something even if I know I shouldn’t.

I will say one thing, but will do something completely different.

I find it difficult to keep my mind on something, and am easily distracted.

I have trouble making decisions, or deciding what I want to do.

I am unaware of, or unconcerned about, how others feel about my behavior.

97
References

Baddeley, A. (1998). Working Memory. Life Sciences, 321, 167-173.

Band, G. P. H. (1997). Preparation, Adjustment, and Inhibition of Responses. Universiteit Van
Amsterdam: faculteit psychologie.

Bechara, A., & Damasio, A. R. (2005). The somatic marker hypothesis: A neural Theory of economic
decision. Games and Economic Behavior, 52, 336-372.

Bechara, A., Damasio, H., Damasio, A. R. (2000a). Emotion, Decision Making and the Orbitofrontal
Cortex. Cerebral Cortex, 10, 295-307.

Bechara, A., Tranel, D., & Damasio, H., (2000b). Characterization of the decision-making impairment
of patients with bilateral lesions of the ventromedial prefrontal cortex. Brain, 123, 2189-2202.

Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Different contributions of the human
amygdala and ventromedial prefrontal cortex to decision-making. Journal of Neuroscience,
19(13), 5473-5481.

Bechara, A., Damasio, H., Tranel, D., & Anderson, S. W. (1998). Dissociation of Working Memory
from Decision Making within the Human Prefrontal Cortex. The Journal of Neuroscience,
18(1), 428-437.

Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1997). Deciding advantageously before
knowing the advantageous strategy. Science, 275, 1293-1295.

Bazerman, M. H., Loewenstein, G. F., & White, S. B. (1992). Reversals of preference in allocation
decisions: Judging an alternative versus choosing among alternatives. Administrative Science
Quarterly, 37, 220-240.

Bernoulli, D. (1954) Exposition of a New Theory on the Measurement of Risk. Econometrica
(Sommer, L., Trans.). 22 (1), 22-36 (Original work published 1738).

Blackmore, S. (2004). Consciousness, New York : Oxford University Press.

Bodendurg, S., & Dopslaff, N. (2008). The Dysexecutive Questionnaire Advanced: Item and Test Score
Characteristics, 4-Factor Solution, and Severity Classification. The Journal of Nervous and
Mental Disease, 196(1), 75-78.

Bower, G. H., Gilligan, S. G., & Monteiro, K. P. (1981). Selectivity of Learning Caused by Affective
States. Journal of Experimental Psychology: General, 110 (4), 451-473.

Bradley, J. V. (1978). Robustness. British Journal of Mathematical and Statistical Psychology, 31, 144-
152.

97
Calder, A. J., Lawrence, A. D., & Young, A. W. (2001). Neuropsychology of fear and loathing. Nature

Reviews Neuroscience, 2, 352-363.

Camerer, C.F. (Guest). (2009, July 24). The Science of Decision-Making [Radio broadcast episode]. In
I. Flatow (Host/Producer), Talk of he Nation: Science Friday. New York, NY: National Public
Radio

Camerer, C.F. (2003). Behavioral Game Theory. Princeton University Press: Princeton.

Casey, B. J., Cohen, J. D., O’Crave, K., Davidson, R. J., Irwin, W., Nelson, C. A., Noll, D. C., Hu, X.,
Lowe, M. J., Rosen, B. R., Truwitt, C. L., & Turski, P. A. (1998). Reproducibility of fMRI
results across four institutions using a spatial working memory task. Neuroimage, 8, 249-461.

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel
distributed processing account of the Stroop effect. Psychological Review, 97, 332-361.

Courtney S. M., Petit L., Maisog J. M., Ungerleider, L.G., & Haxby, J. V. (1998). An area specialized
for spatial working memory in human frontal cortex. Science, 279, 1347-1351.

Damasio, A. R. (1994). Descartes’ Error. New York : G. P. Putnam’s Sons.

Damasio, A. R., Tranel, D., & Damasio, H. (1991). Somatic markers and the guidance of behavior:
Theory and preliminary testing. In: Levin, H.S., Eisenberg, H.M., Benton, A.L. (Eds.), Frontal
Lobe Function and Dysfunction. Oxford Univ. Press, New York, 217-229.

Derbyshire, S. W., Jones, A. K. P., & Gyulai, F. (1997). Pain processing during three levels of noxious
stimulation produces differential patterns of central activity. Pain, 73, 431-445.

Duchenne, G. B. A. (1855). De l’Électrisation Localisée et de son Application à la Physiologie, à la
Pathologie et à la Thérapeutique, Paris: J.B. Baillière.

Duncan, J., Emslie, H., Williams, P., Johnson, R., & Freer, C. (1996). Intelligence and the frontal lobe:
The organization of goal-directed behavior. Cognitive Psychology, 30, 257-303.

Ekman, P. (2003). Darwin, deception, and facial expression. Annals of the New York Academy of
Science, 1000, 205–221.

Eslinger, P. J., & Damasio, A. R., (1985). Severe disturbance of higher cognition after bilateral frontal
lobe ablation. Patient EVR. Neurology, 35, 1731-1741.

Eysenck, H. (1986). Decline and Fall of the Freudian Empire. Harmondsworth: Pelican.

Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of
Abnormal and Social Psychology, 58, 203-211.

97
Field, A. (2009). Discovering statistics using SPSS: Third Edition. Los Angeles: Sage.
Foder, J. A. (1983). The modularity of mind. Cambridge, M.A.: MIT Press.

Francesco, F. (2001). Jerry A. Fodor: Mente e Linguaggio. Rome: Editori Laterza.

Frede, M., & Striker, G. (1996). Rationality in Greek Thought. Clarendon Press: Oxford.

Freud, S. (1949). The Ego and the Id. London: The Hogarth Press Ltd.

Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., DeFries, J. C., & Hewitt, J. K. (2006). Not all
executive functions are related to intelligence. Psychological Science, 17(2),172-179.

Gay, P. (1998). Pleasure Wars (Bourgeois Experience, Victoria to Freud, Vol V). New York, NY: W. W.
Norton & Company.

Gazzaniga, M. S. (1967). The Split Brain in Man. Scientific American, 217(2), 24-29.

Gigerenzer, G., Todd, P. M., & The ABC Research Group (1999). Simple heuristics that make us smart.
Oxford U. K.: Oxford University Press.

Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An fMRI
Investigation of Emotional Engagement in Moral Judgment. Science, 293, 2105-2108.

Grether, D. M., & Plott, C. R. (1979). Economic theory of choice and the preference reversal
phenomenon. American Economic Review, 69, 623-628.

Guth, W., Schmittberger, R., & Schwarz, B. (1982). An Experimental Analysis of Ultimatum
Bargaining. Journal of Economic Behavior and Organization, 3, 376-388.

Harle, K., & Sanfey, A. G. (2007). Incidental sadness biases social economic decisions in the ultimatum
game. Emotion, 7(4), 876-881.

Hadas, Moses (ed.), Essential Works of Stoicism, New York : Bantam.

Hohmann, G. W. (1966). Some effects of spinal cord lesions on experienced emotional feelings.
Psychophysiology, 3, 143-156.

Hsee, C. K. (1996). The availability hypothesis: An explanation for preference reversals between joint
and separate evaluations of alternatives. Organizational Behavior and Human Decision
Processes, 66, 122-129.

Huber, J., Payne, J. W., & Puto, C. (1982). Adding asymmetrically dominated alternatives: violations of
regularity and the similarity hypothesis. The Journal of Consumer Research, 9(1), 90-98.

Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1990). Experimental tests of the endowment effect and

97
the coase theorem. The Journal of Political Economy, 98(6), 1325-1348.

Kahneman, D. (2003) A perspective on judgment and choice: Mapping bounded rationality. American
Psychologist, 58, 697-720.

Kirby, K. N., & Marakovic, N. N. (1995). Modeling myopic decisions: evidence for hyperbolic delay-
discounting within subjects and amounts. Organizational Behavior and Human Decisions
Processes, 64, 22-30.

Kirchner, W. K. (1958), Age differences in short-term retention of rapidly changing information.
Journal of Experimental Psychology, 55(4), 352-358.

Klein, J. (1965). A Commentary on Plato’s Meno. Chapel Hill NC: University of North Carolina Press.

Klüver, H. (1949). Visual functions after removal of the occipital lobes. Journal of Psychology, 11, 23-
45.

Knetsch, J. L.,& Sinden, J. A. (1984). Willingness to pay and compensation demanded: experimental
evidence of an unexpected disparity in measures of value. The Quarterly Journal of Economics,
99, 507-521.

Knoch, D., Pascual-Leone, A., Meyer, K., Treyer, V., & Fehr, E. (2006). Diminishing reciprocal
fairness by disrupting the right prefrontal cortex. Science, 314, 829-832.

Koenigs, M., Young, L., Adolphs, R., Tranel, D., Cushman, F., Hauser, M., & Damasio, A. (2007).
Damage to the prefrontal cortex increases utilitarian moral judgements, Nature, 446, 908-911.

Lavie, N. (2005). Distracted and confused?: Selective attention under load. TRENDS in Cognitive
Sciences. 9(2), 75-82.

Lavie, N. (2004). Load Theory of Selective Attention and Cognitive Control. Journal of Experimental
Psychology: General, 133(3), 339–354.

Le Doux, J. E. (1994). Emotion, memory and the Brain. Scientific American, 270(6), 32-39.

Lichtenstein, S., & Slovic. P. (1971). Reversals of preference between bids and choices in gambling
decisions. Journal of Experimental Psychology, 89, 46-55.

Luce, R. D., & Suppes, P. (1965). Preference, utility, and subjective probability in Handbook of
Mathematical Psychology, eds. R. D. Luce, R. R. Bush, and E. Galanter, (Vol. 111, pp. 249-
410). New York: Wiley.

Manoach, D. S., Schlaug, G., Siewert, B., Darby, D. G., Bly, B. M., Benfield, A., Edelman, R. R., &
Warach, S. (1997). Prefrontal cortex fMRI signal changes are correlated with working memory
load. Neuroreport, 8(2), 545-549.

Martin, R. (2004). The St. Petersburg Paradox. The Stanford Encyclopedia of Philosophy (Fall 2004
Edition). Ed. Edward N. Zalta. Stanford, California: Stanford University.

Martinez, A., Moses, P., Frank, L., Buxton, R., Wong, E., & Stiles, J. (1997). Hemispheric asymmetries
in global and local processing: evidence from fMRI. NeuroReport, 8, 1685-1689.

Meehl, P. E. (1990).Why Summaries of Research on Psychological Theories are Often Uninterpretable,
Psychological Reports, 66, 195-244.

Metcalfe, J., & Jacobs, W.J. (1996). A “hot-system/cool-system” view of memory under stress. PTSD
Research Quarterly, 7(2), 1-3.

McClure, S. M., Laibson, D. I., Loewenstein, G., & Cohen, J. D. (2004). Separate neural systems value
immediate and delayed monetary rewards. Science, 306, 503-7.

Miller, E., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of
Neuroscience, 24, 167-202.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. (2000). The
unity and diversity of executive functions and their contributions to complex “frontal lobe”
tasks: A latent variable analysis. Cognitive Psychology, 41, 49-100.

Navon, D. (1977). Forest before trees: The precedence of global features in visual perception.
Cognitive Psychology, 9, 353-383.

Nishino, S., Ripley, B., Overeem, S., Lammers, G. L., & Mignot, E. (2000). Hypocretin (orexin)
deficiency in human narcolepsy. The Lancet, 355, 39-40.

Owen, A. M., McMillan, K. M., Laird, A. R., & Bullmore, E. (2005). N-Back Working Memory
Paradigm: A Meta-Analysis of Normative Functional Neuroimaging Studies. Human Brain
Mapping, 25, 46-59.

Peterson, M. A., & Grant, E. S. (2003). Memory and Learning in Figure-Ground Perception In B. Ross
& D. Irwin (Eds.) Cognitive Vision: Psychology of Learning and Motivation, 42, 1-34.

Pillutla, M. M., & Murnighan, J. K. (1996). Unfairness, anger, and spite: emotional rejections of
ultimatum offers. Organizational Behavior and Human Decision Processes, 68, 208-224.

Rilling, J. K., Goldsmith, D. R., Glenn, A. L., Jairam, M. R., Elfenbein, H. A., Dagenais, J. E.,
Murdock, C. D., & Pagnoni, G. (2008). The neural correlates of the affective response to
unreciprocated cooperation. Neuropsychologia. 46, 1256-1266.

Rubia, K., Overmeyer, S., Taylor, E., Brammer, M., Williams, S. C. R., Simmons, A., & Bullmore, E. T.
(1999). Hypofrontality in Attention Deficit Hyperactivity Disorder During Higher-Order Motor

97
Control: A Study With Functional MRI. American Journal of Psychiatry, 156, 891-896.

Sakai, K., Rowe J. B., & Passingham, R. E. (2002). Active maintenance in prefrontal area 46 creates
distractor-resistant memory. Nature Neuroscience, 5, 479–484.

Sanfey, A. G., & Chang, L. J. (2008). Multiple Systems in Decision Making. Annals of the New York
Academy of Sciences, 1128, 53–62.

Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L. E., & Cohen, J. D. (2003a) The neural basis of
economic decision-making in the ultimatum game. Science, 300, 1755-1758.

Sanfey, A. G., Hastie, R., Colvin, M. K., & Grafman, J. (2003b) Phineas gauged: decision-making and
the human prefrontal cortex. Neuropsychologia. 41, 1218-1229.

Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-prime reference guide. Pittsburgh, PA:
Psychology Software Tools.

Scott, W. D. (1903). The psychology of advertising in theory and practice. Boston: Small, Maynard &
Company.

Schnall, S., Haidt, J., Clore, G. L., & Jordan, A. H. (2008). Disgust as Embodied Moral Judgment.
Personality and Social Psychology Bulletin, 34(8), 1096-1109.

Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions.
Journal of Neurology, Neurosurgery and Psychiatry. 20, 11-21.

Shiv, B., & Nowlis, S. (2004). Effects of distraction while tasting a food sample: The interplay of
informational and affective components in subsequent choice. Journal of Consumer Research,
31, 599-608.

Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review,
63(2), 129-138.

Simonson, I., & Tversky, A. (1992). Choice in Context: Tradeoff Contrast and Extremeness Aversion,
Journal of Marketing Research, 29(3), 281-295.

Simonson, I. (1989). Choice based on reasons: the case of attraction and compromise effects. The
Journal of Consumer Research. 16(2), 158-174.

Singer, T., Seymour, B., O’Doherty, J., Klaas. E. S., Dolan, J. D., & Frith, C. (2006). Empathic neural
responses are modulated by the perceived fairness of others. Nature, 439, 466-469.

Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science,
283(5408), 1657-61.

Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental

97
Psychology, 18, 643-622.

Sylvester, C. Y. C., Wager, T. D., Lacey, S. C., Hernandez, L., Nichols, T. E., Smith, E. E., & Jonides, J.
(2003). Switching attention and resolving interference: fMRI measures of executive functions.
Neuropsychologia, 41(3), 357-370.

Tranel, D., Bechara, A., Damasio, H., & Damasio, A. R., (1996). Fear conditioning after ventromedial
frontal lobe damage in humans. Society of Neuroscience Abstracts, 22, 1108.

Tversky, A., Sattath, S., & Slovic, P. (1988). Contingent weighting in judgment and choices.
Psychological Review, 95, 371-384.

Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: heuristics and biases. Science, 187,
1124-1131.

Van ’t Wout, M., Kahn, R. S., Sanfey, A. G., & Aleman, A. (2005). rTMS over the right dorsolateral
prefrontal cortex affects strategic decision making. NeuroReport, 16, 1849-1852.

Vince, M. A. (1948). The intermittency of control movements and the psychological refractory period.
British Journal of Psychology, 38, 149-157.

Von Neumann, J., & Morgenstern, O. (1944). Theory of games and economic behavior. Princeton,
N.J. :Princeton University Press.

Weiskrantz, L. (1986). Blindsight: A case study and Implications. Oxford: Oxford University Press.

Weiskrantz, L., Warrington, E. K., Sanders, M. D., & Marshall, J. (1974). Visual capacity in the
hemianopic field following a restricted occipital ablation. Brain, 97, 709-728.

Wells, R. E., & Lyengar, S. S. (2005). Positive illusions of preference consistency: When remaining
eluded by one’s preferences yields greater subjective well-being and decision outcomes.
Organizational Behavior and Human Decision Processes, 98(1), 66-87.

Wilson, B. A., Alderman, N., Burgess, P. W., Emslie, H., & Evans, J. J. (1996) Behavioural Assessment
of the Dysexecutive Syndrome. Bury St. Edmund: Thames Valley Test Company.

Xiao, E., & Houser, D. (2005). Emotion expression in human punishment behavior. Proceedings of the
National Academy of Sciences USA, 102, 7398-7401.

Turn in your highest-quality paper
Get a qualified writer to help you with

“ ESSAYGWIZARD ”

Get high-quality paper

NEW! AI matching with writer

Hire a Writer

Client Reviews

4.9

Sitejabber

4.6

Trustpilot

4.8

Our Guarantees

100% Confidentiality

Information about customers is confidential and never disclosed to third parties.

Original Writing

We complete all papers from scratch. You can get a plagiarism report.

Timely Delivery

No missed deadlines – 97% of assignments are completed in time.

Money Back

If you're confident that a writer didn't follow your order details, ask for a refund.

New to Your Trusted Assignment Help Service? Sign up & Save

Calculate the price of your order

Type of paper needed:

Pages:

You will get a personal manager and a discount.

Academic level:

We'll send you the first draft for approval by at

Total price:

$0.00

Power up Your Academic Success with the
Team of Professionals. We’ve Got Your Back.

Power up Your Study Success with Experts We’ve Got Your Back.

Order Now Order Now