A VIRTUAL REALITY-BASED EXERGAMES INTERVENTION TO IMPROVE THE 

EXECUTIVE FUNCTIONS OF ADULTS AGED 50 AND OLDER 

MOVING WITH PRESENCE:  

 

 
 
 
 
 

By 

 

Kuo-Ting Huang 

 

 

 

 

 

 

A DISSERTATION 

Submitted to 

Michigan State University 

in partial fulfillment of the requirements 

for the degree of 

Information and Media — Doctor of Philosophy 

2018 

ABSTRACT 

MOVING WITH PRESENCE:  

A VIRTUAL REALITY-BASED EXERGAMES INTERVENTION TO IMPROVE THE 

EXECUTIVE FUNCTIONS OF ADULTS AGED 50 AND OLDER 

 

By 

Kuo-Ting Huang 

Prior literature in exergame research suggests that exergaming could improve cognitive 

function in older adults, yet what types of exergames would contribute to cognitive improvement 

have not been fully identified. This dissertation seeks to investigate whether level of immersion 

and type of cognitive load would yield differential outcomes in executive functions in the context 

of exergaming. The process model of the spatial presence experiences and perceptual load theory 

serve as the theoretical framework and guide the research design.   

In a 2 (level of immersion: high vs. low) × 2 (type of task load: task-relevant vs. task-

irrelevant) between-subjects experimental design, participants were randomly assigned into one 

of four conditions and asked to play an exergame (Fruit Ninja) for eight sessions within four 

weeks. Forty-one participants aged over 50 (mean age = 63) finished a single bout of exergaming 

training and 33 participated in the long-term training. Cognitive improvements were assessed 

after a single bout, after two weeks, and after four weeks. 

The results of repeated-measures analysis of covariance (ANCOVAs) revealed a 

significant interaction effect of immersion × time for cognitive flexibility and inhibitory control, 

with an improvement in the high-immersion condition over the course of the four-week training. 

Furthermore, the feeling of spatial presence mediated the relationship between immersion and 

cognitive improvement. However, type of task load had no effect on any aspect of task 

performance over the course of the exergaming training. The results of intent-to-treat analyses, 

which were conducted to handle missing responses, were consistent with the initial analyses 

using raw data. 

These findings suggested that spatial presence, elicited by the immersive virtual reality, 

was involved in a cognitive process, which later led to an improved performance in inhibitory 

control and cognitive flexibility. This study also demonstrated the feasibility and potential for 

older adults to use virtual reality-based exergames as a potential tool for cognitive health. For the 

theoretical implications, this research extends previous research by showing how exergaming in 

virtual reality, which leads to the feeling of presence, could contribute to older adults’ cognitive 

improvement. This study also provided practical implications such that the design of exergames 

could emphasize the game features requiring spatial attention and orientation, which can serve as 

a novel strategy for preventing cognitive decline in older adults. 

 

 

 

 

 

 

 

 

 

 

 

 

Copyright by 
KUO-TING HUANG 
2018 

 

 

 

ACKNOWLEDGEMENTS 

 

This research would not have been completed without the financial support of the 

Dissertation Completion Fellowship, the Pearl J. Aldrich Endowment in Gerontology, or 

Department of Media and Information at Michigan State University.  

I would like to thank Dr. Wei Peng, the chairman of my committee, who has been 

supportive of my research and career goals. I am also grateful to my dissertation committee 

members, Dr. Shelia Cotten, Dr. Rabindra Ratan, and Dr. Hairong Li, who have provided me 

extensive guidance during this process. 

I am especially indebted to my family members and MSU friends, who always cheer me 

up when I have low self-efficacy. I would not be able to finish this dissertation without their 

emotional support.  

 

 

v 

TABLE OF CONTENTS 

 

LIST OF TABLES ___________________________________________________________ viii 
 
LIST OF FIGURES  ___________________________________________________________ ix 
 
CHAPTER 1: INTRODUCTION _________________________________________________  1 
 
CHAPTER 2: LITERATURE REVIEW  ___________________________________________  7 
2.1. The Core Concepts of Executive Functions____________________________________  8 
2.2. Executive Functions, Aging, and Cognitive Reserve  ____________________________  9 
2.3. Exergaming: A Combination of Physical and Mental Exercise  ___________________  13 
2.4. Exergames as Mediated Environments in Conjunction with Spatial Presence and 
Executive Functions ________________________________________________________  16 
2.5. Cognitive Load, Spatial Presence, and Executive Functions ______________________  22 

 
CHAPTER 3: METHODOLOGY  _______________________________________________  26 
3.1. Participants ____________________________________________________________  26 
3.2. Stimuli and Intervention Design  ___________________________________________  28 
3.3. Manipulation Check and Test for Confounding  _______________________________  30 
3.4. Outcome Variables______________________________________________________  31 
3.5. Control Variables _______________________________________________________  35 
3.6. Pilot Test of Manipulations _______________________________________________  35 
3.7. Process Evaluation ______________________________________________________  37 
3.8. Procedure _____________________________________________________________  39 
3.9. Analytic Techniques  ____________________________________________________  40 

 
CHAPTER 4: RESULTS ______________________________________________________  42 
4.1. Demographics  _________________________________________________________  42 
4.2. Manipulation Check and Confounding Testing ________________________________  43 
4.3. Overview: Effects of VR and Types of Task Load on Cognitive Improvement _______  45 
4.3.1. Overview __________________________________________________________  45 
4.3.2. Intention-To-Treat Analysis ___________________________________________  51 
4.4. The Baseline-Posttest Comparisons at Different Time Frames ____________________  53 
4.4.1. First Posttest: After the first Single-Bout _________________________________  53 
4.4.2. Second Posttest: After the Two-Week Training (four sessions) ________________  56 
4.4.3. Third Posttest: After Four-week Training (eight sessions) ____________________  59 
4.5. Effects of VR and cognitive load on Spatial Presence Experiences ________________  63 
4.6. Mediation Effects of Spatial Presence on Executive Functions  ___________________  65 

 
CHAPTER 5: DISCUSSION ___________________________________________________  67 
5.1. Overview of the Findings  ________________________________________________  68 
5.2. Summary of Findings on Different Time Frames ______________________________  70 
5.3. The Role of Level of Immersion on Cognitive Improvement _____________________  71 

vi 

5.4. Spatial Presence as a Mediator between Immersion and Cognitive Improvement _____  72 
5.5. Task-relevant vs. Task-irrelevant Load on Cognitive Improvement ________________  74 
5.6. Summary of Discussion __________________________________________________  78 
5.7. Limitations ____________________________________________________________  79 

 
CHAPTER 6: CONCLUSION __________________________________________________  82 
 
APPENDIX _________________________________________________________________  85 
 
REFERENCES ______________________________________________________________  92 
 

vii 

LIST OF TABLES 

  

Table 1: Process-Evaluation Plan ________________________________________________  38 
 
Table 2: Demographics and Baseline Executive Functions ____________________________  43 
 
Table 3: Manipulation Check of Level of Immersion  ________________________________  43 
 
Table 4: Confounding Testing of Types of Task Loads _______________________________  44 
 
Table 5: Repeated ANCOVAs of Outcome Measures Across Three Posttests  _____________  46 
 
Table 6: Estimated Marginal Means of Outcome Measures Across Three Posttests _________  47 
 
Table 7: ITT Analyses of Outcome Measures Across Three Posttests ____________________  52 
 
Table 8: Repeated ANCOVAs of Outcome Variables After the First Single-Bout __________  53 
 
Table 9: Estimated Marginal Means of Outcome Variables After the First Single-Bout ______  54 
 
Table 10: ITT Analyses of Outcome Measures After the First Single-Bout  _______________  55 
 
Table 11: Repeated ANCOVAs of Outcome Variables After the Two-Week Training  ______  57 
 
Table 12: Estimated Marginal Means of Outcome Variables After the Two-Week Training __  58 
 
Table 13: ITT Analyses of Outcome Measures After the Two-Week Training _____________  59 
 
Table 14: Repeated ANCOVAs of Outcome Variables After the Four-Week Training  ______  60 
 
Table 15: Estimated Marginal Means of Outcome Variables After the Four-Week Training __  61 
 
Table 16: ITT Analyses of Outcome Measures After the Four-Week Training _____________  62 
 
Table 17: ANCOVAs of Spatial Presence Experiences _______________________________  64 
 
Table 18: PROCESS Mediation Analysis__________________________________________  66 
 
Table 19: Results of The Hypotheses of The Current Study  ___________________________  69 
 
Table 20: The Effects of Manipulation at Different Time Frames _______________________  70 
  

 

 

viii 

LIST OF FIGURES  

 

Figure 1: Research Question and Hypotheses of the Current Research ___________________  25 
 
Figure 2: Fruit Ninja in Virtual Reality  ___________________________________________  28 
 
Figure 3: Stroop Test  _________________________________________________________  33 
 
Figure 4: Trail-making Task ____________________________________________________  33 
 
Figure 5: Digit Span Task ______________________________________________________  34 
 
Figure 6: Stroop Results after A Single Bout, Two Weeks, and Four Weeks ______________  48 
 
Figure 7: Trail-Making Results after A Single Bout, Two Weeks, and Four Weeks _________  49 
 
Figure 8: Backward Digit Results after A Single Bout, Two Weeks, and Four Weeks _______  50 
 
 

ix 

CHAPTER 1: INTRODUCTION 

 

The older population has grown dramatically since the beginning of the 21st century. 

There has been a 21% increase in the older population aged 65 or more, and one-seventh of the 

population currently belongs to the 65 or more age group (Centers for Disease Control and 

Prevention, 2013). Older adults are at a considerably higher risk for having problems related to 

the aging of the brain, and several executive abilities associated with reasoning decline starting 

around age 45 (Singh-Manoux et al., 2012). According to the World Health Organization 

(WHO), more than 20% of older adults suffer from a mental or neurological disorders, one of the 

most common being dementia (Yasamy, Dua, Harper, & Saxena, 2013), which refers to an 

impairment of cognitive functions.  

The cognitive problems of older adults need to be addressed due to the high costs of 

medical care and the negative social ramifications associated with dementia and other cognitive 

impairments. The economic cost of dementia in the United States ranged from $159 billion to 

$215 billion in 2010 and will increase to $511 billion by 2040 (Hurd, Martorell, Delavande, 

Mullen, & Langa, 2013). Most of these costs go toward institutional and home-based care, which 

accounts for 75% to 84% of dementia costs. Moreover, the annual costs of health care among 

adults 50 years and older is about $860 billion (Watson et al., 2016). Thus, it is important to 

promote healthy cognitive aging and to refine the prevention research in midlife and older 

adulthood.  

Both scientific literature (e.g., Diamond, 2015) and health-related organizations (e.g., 

World  Health Organization, 2015) have emphasized the positive impacts of physical activity on 

dementia and cognitive decline prevention. The WHO (2015) has listed several new treatments 

1 

for the prevention of cognitive decline and dementia; one of which is physical activity.  

Furthermore, recent longitudinal studies found a positive relationship between physical inactivity 

and a higher risk for cognitive decline in older adults, whereas regular physical activity can delay 

or even prevent chronic and cognitive diseases in adults age 50 and over (Diamond, 2015). The 

promotion of physical activity seems to provide protective factors against cognitive decline for 

adults 50 years or older.  

However, in the United States, 28% of adults ages 50 and older are physically inactive, 

which has been highly associated with mental illness, cognitive impairment, and the risk of 

premature death (Watson et al., 2016). Moreover, the Centers for Disease Control and Prevention 

(CDC) suggests that adults ages 50 and older should stay physically active because 80% of 

costly chronic conditions can be prevented or improved with physical activity (Centers for 

Disease Control and Prevention, 2014). The CDC emphasizes that adults age 50 and older should 

have more physical activity, which can reduce the risk of cognitive impairment and support 

healthy aging. 

Exergaming, which refers to using active video games as a form of exercise that requires 

players to use motion-based controllers or body movements to play, has the potential to combine 

the benefits of physical activity and the attractiveness of video games (Blondell, Hammersley-

Mather, & Veerman, 2014). The genres of exergames range from requiring arm and hand 

movements (e.g., boxing and baseball) to full body movements (e.g., dancing and jumping). As 

exergames exploded onto the market, researchers quickly began to study the effects of 

exergaming on executive functions in both children and older adults (Staiano & Calvert, 2011). 

In general, exergaming has been considered as a good alternative to, if not better than, traditional 

exercise because of its attractive combination of physical activity and gamification (Anderson-

2 

Hanley, Maloney, Barcelos, Striegnitz, & Kramer, 2017; Monteiro-Junior, Otero Vaghetti, 

Nascimento, Laks, & Deslandes, 2016).  

Similar to traditional exercise, exergames have been regarded as a potential tool to 

produce cognitive benefits and contribute to maintaining or improving executive functions (e.g., 

working memory and inhibitory control) in older adults (Staiano & Calvert, 2011). Researchers 

have identified two mechanisms that account for the relationship between exergaming and 

cognitive benefits: (1) physical exercise, which accounts for the increase in physiological arousal 

and attentional resources, and (2) mental exercise, which refers to the cognitive demands and 

mental effort during exergaming (Anderson-Hanley et al., 2017; Best, 2013). In other words, 

moving with thought during exergaming is key to improving cognitive functions in older adults. 

Regarding the impact of physical exercise during exergaming, systematic reviews on the 

effects of exergaming have found that physical activity has a consistent impact on cognitive 

improvement (Ogawa, You, & Leveille, 2016). However, the findings of previous research on 

the effects of traditional exercises on cognitive improvement are not always consistent 

(Diamond, 2015). A possible explanation might be that movement during exergaming is different 

from movement during traditional exercise, which does not always activate the cognitive 

process. The one key factor that separates traditional exercise, which has inconsistent results, and 

exergaming is the concept of moving with thought. Only when moving with thought will the 

cognitive process be activated (Diamond, 2015). Therefore, simple aerobic activities that require 

little thought (e.g., running, resistance training) will not lead to cognitive benefits. 

However, exercising with thought may not fully explain the difference between 

exergaming and traditional exercise in player’s cognitive improvement. For example, pure 

running during exergaming, an activity that is supposed to require little thought, still had a 

3 

positive impact on cognitive functions (e.g., Best, 2012). The different effects of physical 

activity between traditional forms of exercise and exergaming on cognitive improvement might 

result from a sense of presence. Compared to traditional forms of exercise, exergaming involves 

moving within a mediated environment, which requires use of players’ spatial navigation skills 

and selective attention to mentally simulate their locations (Monteiro-Junior et al., 2016). Spatial 

presence refers to the sense of being situated in a mediated environment and the perception of 

action possibilities in the space. Research indicates that spatial presence is another contributing 

factor that may help explain the cognitive benefits associated with being in mediated 

environments or virtual reality (VR) (Kober, Kurzmann, & Neuper, 2012). Another study applied 

a VR-based serious game, which consisted of daily life activities to train stroke patients, found a 

significant improvement in participants’ cognitive functions after a 4–6 week intervention 

compared to those of people who were on the waitlist for the study (Gamito et al., 2017). 

Therefore, playing an immersive VR game should involve the cognitive and psychological 

process of spatial presence, and the sense of spatial presence associated with VR and other 

mediated environments may play a major role during exergame-based interventions.  

Furthermore, neuroscientific techniques support the argument that the feelings of spatial 

presence elicited by immersive technologies are strongly associated with cognitive improvement. 

The brain circuits linked to executive functions become more active when media users navigate 

themselves within virtual environments (Monteiro-Junior et al., 2016). Studies using functional 

neuroimaging, a measure of the relationship between activities in certain brain regions and 

specific mental or cognitive functions, have found that executive functions (i.e., selective 

attention and task management) activated the dorsolateral prefrontal cortex (DLPFC) (Diamond, 

2015). The right-DLPFC is associated with visual-related cognitive functions, and the left-

4 

DLPFC is associated with the verbal-DLPFC (Lo Priore, Castelnuovo, Liccione, & Liccione, 

2003). In addition to the associated executive functions, the DLPFC also plays a major role 

during the experience of spatial presence (Gamito et al., 2015, 2017), so the activation of DLPFC 

is related to both the use of executive functions and the feeling of spatial presence.  

Based on the theoretical reasoning and neuroscientific evidence, the current research 

investigates the role of spatial presence in the process of exergaming. The core argument is that 

spatial presence plays a vital role in cognitive improvement via increasing players’ attention 

allocated to the mediated environment. The first goal of this study is to investigate the effects of 

immersion (i.e., virtual reality and 3D environments) as well as the role of spatial presence on 

executive functions in the process of exergaming. The model of the formation process of spatial 

presence experiences (Wirth et al., 2007) is used to explain the mechanism underlying the 

cognitive process of how exergaming influences executive functions.  

Recent studies have investigated the effects of cognitive loads on older adult’s cognitive 

improvement in the context of exergaming. However, the effects of the cognitive demands of 

exergaming were not fully explored. It is still unclear what types of exergames and what 

elements of exergames contribute to cognitive improvement. Some studies found that both the 

physical activity and the cognitive load had an effect on cognitive improvement (Best, 2013; 

Ogawa et al., 2016) but some studies only found that physical activity had an effect (Best, 2012, 

2013). The inconclusive findings may result from the different definitions of cognitive load used 

across studies on exergaming and cognitive improvement.  

Some studies defined cognitive demand or cognitive load as task-relevant load, which 

refers to solving a task that requires higher-order cognitive processes or more working memory 

and selective attention (e.g., Anderson-Hanley et al., 2017; Best, 2012). Some studies defined 

5 

cognitive demand or cognitive load as task-irrelevant load, which refers to solving an additional 

cognitively involving task in addition to doing the primary task (e.g., Radovanović, Jovičić, 

Marić, & Kostić, 2014). Both types of cognitive tasks require players’ mental effort, but in 

different ways. The second goal of this study is to investigate the impacts of the different types 

of cognitive load during exergaming and the possible mechanisms that may account for the 

improvement of older adults’ executive functions during exergaming. Perceptual load theory 

(Lavie, 1995, 2005) provides the current study with a general framework for classifying different 

types of task load based on previous research.  

This research addresses the following questions: Does playing an exergame in a more 

immersive mediated environment lead to more cognitive improvement than playing it in a less 

immersive environment? Does the level of spatial presence associated with exergaming have an 

impact on executive functions after exergaming? What roles do cognitive load and spatial 

presence perform during exergaming in relation to improving executive functions in older 

adults? Does the type of cognitive load influence executive functions differently during 

exergaming? In the literature review, the current study discusses these two theoretical 

frameworks (i.e., the model of spatial presence experience and perceptual load theory) to answer 

these questions.  

 

 

6 

CHAPTER 2: LITERATURE REVIEW 

 

Prior research has shown that physical activity can contribute to cognitive benefits in 

midlife and older adults through both physical exercise and mental exercise, and exergames 

include both factors (Anderson-Hanley et al., 2017). Before discussing the relationships among 

exercise, exergaming, and cognitive functions, the literature review begins with a brief overview 

of the conceptualization of executive functions as well as the characteristics of exergames. 

Current research argues that being in an immersive environment is a key contributor to 

mental exercise in the context of exergaming. To better understand the proposed relationship, 

this section reviews prior studies on the effects of (1) being in a mediated environment and (2) 

the effects of mental simulation of the mediated environment on cognitive improvement in the 

context of exergaming. Previous research suggests that mental simulation of a mediated 

environment has been regarded as a form of mental exercise during exergaming (Monteiro-

Junior et al., 2016). Also, being in a virtual reality (VR) and three-dimensional (3D) environment 

requires people to construct a mental simulation, which serves as a prerequisite of spatial 

presence experiences (Hofer, Wirth, Kuehne, Schramm, & Sacau, 2012; Schubert, 2009; Wirth et 

al., 2007). Therefore, the proposed study employs the process model of the formation of spatial 

presence experiences as the theoretical framework.  

Besides exploring the role of spatial presence in the relationship between exergaming and 

cognitive improvement, this study examines the role of cognitive load in exergames. Previous 

studies conceptualized and operationalized cognitive load differently, which may have led to 

inconclusive results regarding the impacts of exergaming (Ogawa et al., 2016). This chapter 

reviews the literature relating to the effects of two types of cognitive load (i.e., task-relevant vs. 

7 

task-irrelevant) on media users’ spatial presence and executive functions within the framework 

of perceptual load theory (Lavie, 1995, 2005). 

The aim of this literature review is to (1) gain a better understanding of the role of spatial 

presence in the relationship between being in mediated environments (immersion) and cognitive 

benefits and (2) examine how different types of cognitive loads in exergames may influence 

older adults’ executive functions. The first section starts with exploring what executive functions 

are and why they are important.  

2.1. The Core Concepts of Executive Functions 

Executive functions are defined as a series of top-down cognitive processes, including 

inhibitory control, working memory, and cognitive flexibility (Diamond, 2013). These three core 

executive functions are the foundations for higher-order cognitive functions, such as problem-

solving, planning, and reasoning. 

The first core executive function, inhibitory control, refers to our ability to control 

attention and behaviors, and to overcome impulses in order to do what is more suitable or needed 

(Diamond, 2013). Inhibitory control helps us react or pay attention to what we should focus on 

(e.g., specific or important tasks) instead of being affected by automatic responses (e.g., habits) 

to environmental stimuli. In other words, inhibitory control, which requires cognitive effort, 

prevents us from succumbing to responses to internal impulses or external distractions.  

Inhibitory control includes several aspects, including selective attention, cognitive 

inhibition, and self-control (Diamond, 2013, 2014). First, selective attention refers to the ability 

to selectively focus on a stimulus or thought over other stimuli based on our intention. In other 

words, selective attention enables us to focus on what we choose. Another aspect of inhibitory 

control is cognitive inhibition, which refers to the ability to suppress our thoughts, memories, 

8 

and other mental representations. Finally, self-control relates to the capacity to control our 

behaviors or emotions despite temptations or distractions. To sum up, these three aspects of 

inhibitory control help people stay focused on their goals and intentions by reducing 

environmental interferences. 

The second core executive function, working memory, refers to the ability to hold 

information in mind and use it. Working memory includes two content categories: nonverbal 

working memory and verbal working memory (Diamond, 2015). For example, calculating a 

math problem without writing it down on paper requires working memory. Also, processing new 

information to make plans for the future requires working memory. Different from short-term 

memory, working memory allows us not only to store new information but also to manipulate 

that information. In short, working memory makes it possible for us to integrate mental resources 

by organizing and rearranging new information in our mind.  

The other core executive function, cognitive flexibility, refers to the ability to “think out 

of the box” (Diamond, 2015, p. 147). For example, cognitive flexibility allows people to see 

things from another perspective or imagine events from another viewpoint. Cognitive flexibility 

is the ability to shift perspectives interpersonally or spatially. During the process of changing 

perspectives, we require the other two executive functions (i.e., inhibitory control and working 

memory) to switch from one perspective to another. Therefore, cognitive flexibility is highly 

dependent on the other two core executive functions. In the end, these three executive functions 

are essential for mental health as well as physical health (Diamond, 2015). 

2.2. Executive Functions, Aging, and Cognitive Reserve 

Over the past 100 years, researchers have consistently reported that age is one of the 

predictors of cognitive decline (Salthouse, 2009). For example, visual construction skills, which 

9 

refer to the ability to assemble individual items and objects into a whole, decline with aging 

(Singh-Manoux et al., 2012). Cognitive skills such as concept formation, abstract thinking, and 

information processing speed, also decline during aging (Singh-Manoux et al., 2012). Thus, 

cognitive decline is highly correlated with age. 

Much of the research on cognitive aging focuses on older populations, such as those aged 

60 (or 65) and higher. However, the focus on the cognitive aging of older populations has been 

widely challenged. Recent studies have started to find that cognitive decline may begin earlier 

than we expected (Finch, 2009; Salthouse, 2009). Multiple studies found that cognitive functions 

may start dropping in performance after midlife (i.e., age 45–49) (Singh-Manoux et al., 2012) or 

about 50 to 55 years old (Albert & Moss, 1988; Ronnlund et al., 2005). Some researchers have 

even explicitly pointed out that the implicit assumption of little cognitive decline occurring 

before age 60 limits researchers to examining the correlates instead of the causes of cognitive 

decline (Singh-Manoux et al., 2012). To be more specific, the impact of cognitive decline may 

start from middle age, and people who are less than 65 years old may still suffer from age-related 

cognitive decline. Therefore, simply focusing on older adults may ignore some early risk factors 

for adverse cognitive outcomes. 

Regarding executive functions, the brain regions associated with executive functions 

have been found to shrink with age. Inhibitory control, working memory, and cognitive 

flexibility are all susceptible to decline when the aging brain experiences physical and functional 

changes (Diamond, 2014). Regarding inhibition, the aging brain affects a person’s ability to 

inhibit automatic responses to a stimulus and ignore irrelevant aspects (Chao & Knight, 1997). 

Similar to inhibition, working memory and cognitive flexibility also decline with age due to age-

related changes in certain brain regions or a loss in brain tissue (i.e., prefrontal cortex; Gazzaley, 

10 

Cooney, Rissman, & D'esposito, 2005). Thus, aging is associated with declines in executive 

functions. 

Fortunately, it is possible to slow down age-related cognitive decline if people receive 

adequate cognitive training or intervention (Stern, 2012). The concept of cognitive reserve points 

the way toward slowing down the level of cognitive decline and maintaining cognitive functions 

as much as possible (Stern, 2012). According to the theoretical framework of cognitive reserve 

(Stern, 2012), two types of reserves account for age-related cognitive loss: brain reserve and 

cognitive reserve. Brain reserve refers to genetically determined characteristics (e.g., brain 

volume and the number of neurons in certain brain regions). Cognitive reserve focuses on the 

brain’s potential for plasticity and reorganization in neural processing. Specifically, brain reserve 

emphasizes the hardware characteristics such as brain volume and neuronal structural integrity, 

and cognitive reserve highlights the software aspects such as the cognitive functioning and 

plasticity of neural circuits (Cheng, 2016). In other words, the concept of cognitive reserve 

explains how to preserve cognitive functions from both a hardware and a software perspective.  

Previous cognitive interventions have supported the theoretical framework of cognitive 

reserve. For brain reserve, researchers have found that physical activities, such as exercise and 

dancing, improve older adults’ cognitive functions via activating certain brain regions associated 

with cognitive functions (Crescentini, Urgesi, Fabbro, & Eleopra, 2014; Valenzuela, Sachdev, 

Wen, Chen, & Brodaty, 2008). A systematic review summarizing the effects of physical activity 

on peripheral brain-derived neurotrophic factors in healthy adults found a clear link between 

habitual physical activity (i.e., acute and chronic aerobic exercise) and brain health (Huang, 

Larsen, Ried‐Larsen, Møller, & Andersen, 2014). This relationship between physical activity 

and brain health was much stronger in the studies using subjects aged 50 years old and above 

11 

rather than in the studies whose subjects were around 20 to 30 years old. Also, regular physical 

activity by people in midlife was associated with larger total brain volume in late life when 

compared to those with a sedentary lifestyle in midlife (Rovio et al., 2010). These findings show 

that brain reserve can be maintained by exercising regularly. 

Regarding cognitive reserve and plasticity, it is possible to use cognitive training 

interventions to minimize the cognitive decline associated with advancing aging (Harada, Love, 

& Triebel, 2013). A meta-analysis has found that speed process training had a consistent impact 

on older adults’ cognitive ability to perform daily living activities (Ball, Edwards, & Ross, 

2007).  Research also suggests that cognitively engaging activities have the potential to slow 

down the degeneration of the brain cells in certain regions (e.g., hippocampus) among older 

adults (Valenzuela et al., 2008). Furthermore, cognitive training interventions have been found to 

have long-term effects. The gains after various cognitive training interventions have been found 

to last for at least three months and up to two years after training (Ball et al., 2002; Brehmer, 

Westerberg, & Bäckman, 2012). In other words, interventions requiring cognitively engaging 

activities may help older adults preserve cognitive functions and promote healthy aging. 

Research suggests that playing complex games in conjunction with performing aerobic 

exercise may produce a combination effect that increases both brain reserve from the exercise 

and cognitive reserve from the cognitive training (Stern, 2012). Studies further indicate that 

physical activity requiring limited or little thought (e.g., strength training and treadmill running) 

do not contribute to executive functions, and therefore moving with thought should be used to 

improve executive functions (Diamond, 2015). Interventions and training programs that 

incorporate cognitive engagement into physical activities may better prevent or limit cognitive 

decline in older adults. 

12 

Exergaming, which combines physical exercise with interactive gaming and cognitively 

engaging features, has become regarded as an effective means to promote desired health 

outcomes. Empirical evidence has shown that exergaming is an effective tool for cognitive 

improvement in older adults (Anderson-Hanley et al., 2017). Qualitative research indicates that 

exergames provide older adults with the perception of cognitive benefits from playing the games, 

which subsequently fosters adherence to exercise and also increases the motivation for 

participation (Meneghini, Barbosa, de Mello, Bonetti, & Guimaraes, 2016).  That is to say, the 

effects of exergaming on cognitive benefits not only manifest in the participants’ cognitive task 

scores in the lab settings but also in the subjective perception of cognitive improvement in daily 

life. The next section introduces the mechanism underlying the relationships between exercises 

and executive functions.  

2.3. Exergaming: A Combination of Physical and Mental Exercise  

Previous research has shown that both inhibitory control and working memory affect 

cognitive flexibility which declines during aging (Jurado & Rosselli, 2007; Salthouse, Atkinson, 

& Berish, 2003). Fortunately, these executive functions are trainable and can be improved by 

repeated practice (Diamond, 2014). In particular, physical activity has been regarded as one of 

the most effective ways to enhance cognitive functions and decrease cognitive decline 

(Diamond, 2015). Empirical evidence also shows that doing aerobic exercises improves older 

adults’ executive functions, including processing speed, attention, and cognitive flexibility, even 

after controlling for age, gender, and education (Masley, Roetzheim, & Gualtieri, 2009). Similar 

conclusions have also been drawn in other studies (e.g., Hillman, Erickson, & Kramer, 2008; 

Kramer et al., 1999). 

13 

However, recent research has shown that simply doing physical exercise is not always an 

effective way to prevent cognitive decline in older adults. Two meta-analyses on the effects of 

physical exercise on executive functions found that many studies did not identify a significant 

improvement in cognitive benefits in older adults from doing simple aerobic exercises 

(Angevaren, Aufdemkampe, Verhaar, Aleman, & Vanhees, 2008; Smith et al., 2010). Another 

systematic review found that simple physical exercise (e.g., treadmill, running, and resistance 

training) had little or no effect on older adult’s executive functions (Diamond, 2015). The 

researchers concluded that cognitive improvement requires both physical activity and cognitive 

effort (Diamond, 2015). Therefore, simply doing traditional aerobic exercise may not be enough 

to improve the executive functions, and moving with thought is a key factor for cognitive 

improvement. 

Different from the results of traditional exercise on cognitive improvement, research has 

shown that older adults benefit from exergaming interventions across various genres of games. 

Exergames (or active video games), which are defined as video games involving motor skills and 

movements, have been positively linked to the improvement of executive functions in older 

adults (Staiano & Calvert, 2011). For example, a study on older adults’ executive functions using 

a non-commercial biking game found that several of the participants’ cognitive outcomes 

improved after the intervention, including executive functions (color trails difference, Stroop 

test, and digits backward test), visuospatial skills and memory, and verbal memory (Anderson-

Hanley et al., 2012). Another study found that playing a dancing exergame (Dance Dance 

Revolution) helped elderly females improve their inhibitory control (Chuang, Hung, Huang, 

Chang, & Hung, 2015). These findings show that exergaming can have a positive impact on 

older adult’s executive functions across different exergame genres. Moreover, compared to 

14 

traditional exercises, aerobic exergaming seems to have a stronger impact on older adults’ 

cognitive functions (Anderson-Hanley et al., 2012). 

Regarding the difference between exergaming and traditional forms of exercise, 

researchers argue that exergaming requires players to move with more mental effort. To be more 

specific, the reason why exergaming has a stronger potential to improve people’s executive 

functions is that exergaming is not only physically engaging but also cognitive engaging 

(Anderson-Hanley et al., 2017; Best, 2013; Monteiro-Junior et al., 2016). In the context of 

exergaming, players not only have to move in ways similar to traditional exercise but they also 

have to interact with the virtual environment and interpret the stimuli and cues from the 

environment (Anderson-Hanley et al., 2012; Monteiro-Junior et al., 2016). Therefore, when 

playing exergames, players have to be active both mentally and physically. 

Similar to the concept of cognitive reserve, two main mechanisms underlie the effects of 

exergaming on executive functions: (1) physical exercise—e.g., moving, jumping, or running—

which enlarges people’s attentional pool by increasing physiological arousal and motor control 

skills, and (2) mental exercise—e.g., memorizing the rules and movements of an exercise—

which involves higher-order cognition and primes players to use their executive functions 

(Anderson-Hanley et al., 2017; Barcelos et al., 2015; Best, 2012) . However, when discussing the 

effects of mental exercise, prior studies have mainly focused on the cognitive tasks in addition to 

physical tasks instead of the cognitive and psychological process of interacting with virtual 

environments. The current study argues that the media characteristics of exergames should also 

be examined instead of merely focusing on the effects of their additional cognitive demands. The 

following section discusses how media activate players thought within the theoretical framework 

of spatial presence experiences. 

15 

2.4. Exergames as Mediated Environments in Conjunction with Spatial Presence and 

Executive Functions  

Compared to traditional exercise, exergaming, requires players to not only move 

physically but also to simulate virtual environments cognitively (Anderson-Hanley et al., 2017). 

When moving through a mediated environment, such as a virtual reality or gaming environment, 

players also have to invest their cognitive resources to simulate and interact with the 

environment, which contribute to greater cognitive benefits (Anderson-Hanley et al., 2012; 

Monteiro-Junior et al., 2016). In other words, the demands of both the cognitive (i.e., interacting 

with the mediated environment) and the physical tasks (i.e., exercise movements) lead to the 

greatest improvements in executive functions.  

Regarding the impacts of media factors, research has pointed out the importance of 

presence during the cognitive process. Theoretically, presence has been defined as a key feature 

that relates to the central nervous system and helps human beings extend the perception of the 

sensory organs to better experience an external space (Riva, Waterworth, & Waterworth, 2004). 

The cognitive process of presence is also related to an automatic application of reasoning 

modules, which help humans understand physical and social causation better when interacting 

with mediated environments and objects (Lee, 2004). In other words, feeling presence is an 

innate function and is highly relevant to the reasoning functions of human beings. 

Clinical findings and neuroscientific evidence also indicate that more immersive 

environments, such as three-dimensional or VR-based settings, are associated with the use of 

cognitive functions, especially among older adults. Clinical findings found that, compared to 

traditional exercise, stationary cycling within a virtual environment led to greater improvement 

in executive functions (Anderson-Hanley et al., 2012). Another study examining the 

16 

effectiveness of VR-based serious games for stroke patients found that VR-based cognitive 

training enhanced the results of neuropsychological rehabilitation (Gamito et al., 2017). 

Biological mechanisms related to interacting with virtual environments could also improve 

people’s executive functions.  

Several brain areas that are responsible for cognitive functions become more active when 

people are in virtual environments (Maguire et al., 1998; Monteiro-Junior et al., 2016). 

Furthermore, interacting with virtual environments allows players to interpret the stimuli from 

the media, and this process activates specific brain circuits (i.e., hippocampus and frontal cortex) 

associated with cognitive functions (Maguire et al., 1998; Monteiro-Junior et al., 2016). In other 

words, players are more likely to become mentally stimulated in mediated environments when 

the media transmits more stimuli and cues. Therefore, the amount of stimuli and cues provided 

by media environments, which can be conceptualized as immersion (Slater & Wilbur, 1997), 

may be a key predictor of cognitive improvement. Based on the above theoretical frameworks 

and empirical findings, the first hypothesis of the current research is proposed: 

H1: In the context of an exergaming intervention, individuals in the high immersion 

condition will have better cognitive improvement in terms of executive functions as 

indicated in (a) inhibitory control, (b) cognitive flexibility, and (c) working memory than 

those in the low immersion condition. 

Prior research on exergame interventions indicates that being in a mediated environment 

requires people to simulate the environment mentally, which leads to cognitive improvement in 

older adults (Anderson-Hanley et al., 2012). However, few empirical studies have directly 

investigated media characteristics, such as the amount of stimuli from media technologies (i.e., 

immersion), on the mental simulation of mediated environments in the context of exergaming. A 

17 

better understanding of the relationship between immersion and cognitive improvement requires 

investigation of the role of immersion in the processes of mentally simulating mediated 

environments. The current study uses the model of the formation process of spatial presence 

experiences as the theoretical framework to illustrate how the immersion levels of media 

facilitates people’s mental simulation of the mediated environments, which leads to a sense of 

presence. 

According to the process model of the formation of spatial presence experiences (Hofer 

et al., 2012; Schubert, 2009; Wirth et al., 2007), the construction of a media user’s mental 

simulation of a mediated environment requires their mental and attentional resources to process 

the spatial cues from the mediated environment. The more spatial cues from the mediated 

environments, the more attentional resources media users will allocate to the media. For 

example, compared to non-VR settings, VR environments provide richer spatial cues that will 

attract users’ attention and allow them to build a better mental representation of the mediated 

environment. However, if users have fewer spatial cues from the mediated environment, it is less 

likely they will actively engage in constructing a mental representation of the environment. Thus, 

when the users allocate their mental effort and attentional resource to the virtual reality, they will 

actively build a spatial situational model of the mediated environment.  

Forming a mental representation and simulation of a mediated environment is one part of 

the mental exercise during exergaming. Exergame players not only create mental representations 

of the environments but they also have to actively engage in executive controls, such as 

inhibitory control and working memory, to relocate themselves into the virtual environment 

(Anderson-Hanley et al., 2017; Monteiro-Junior et al., 2016). In other words, the mental exercise 

of exergaming, which is very similar to the perceptual processing of spatial presence experiences 

18 

(e.g., Wirth et al., 2007), involves two processes: the construction of mental representations of 

mediated environments, and feeling relocated into virtual environments. 

Application of the concept of a spatial situational model to the context of exergames 

suggests people will focus more attentional resources on the mediated environment when the 

media provide more spatial cues. Therefore, attention on the mediated environment should 

enhance the likelihood of the construction of a mental model of the mediated environment. In 

other words, if people play exergames in an environment with richer spatial cues, they are more 

likely to engage in mental simulations and construct spatial mental models compared to those in 

the less immersive environments (Hofer et al., 2012; Schubert, 2009; Wirth et al., 2007). Also, 

when people are more engaged in mental exercise, which refers to creating a mental 

representation of the environment in a context of virtual reality, they are more likely to improve 

their executive functions after exergaming (Anderson-Hanley et al., 2012). 

Media users experience the feeling of spatial presence after going through a two-step 

process (Hofer et al., 2012; Schubert, 2009; Wirth et al., 2007). The first step is to create a 

mental representation of a mediated environment. During this process, individuals use their 

mental resources and allocate their attention to process the spatial cues from the mediated 

environment. When media users allocate their mental and attentional resources to the virtual 

reality, they can construct a spatial situational model of the mediated environment. During the 

second step, media users try to confirm or reject the mental representation of the mediated 

environment based on their knowledge and experience. If users’ mental representation of the 

mediated environment is mapped onto real-life experiences, media users suspend their disbelief 

(i.e., they reject distracting information) and accept the spatial situational model as their primary 

frame of reference. After this occurs, the feeling of spatial presence emerges  

19 

How does immersion influence the formation process of spatial presence? In the early 

1990s, presence scholars posited two contributors of presence: vividness, such as sensory breadth 

and depth, and interactivity, such as speed, range, and mapping (Steuer, 1992). A recent meta-

analysis of over 100 experiments on technological immersion and presence found that immersive 

features have a medium effect size on the spatial presence (Cummings & Bailenson, 2016) . 

According to the two-step model of spatial presence experiences (Hofer et al., 2012; Wirth et al., 

2007), an immersive technology (e.g., VR) will allow media users to allocate more attentional 

resources and enable individuals to construct better mental representations of mediated 

environments and stronger presence experiences. Therefore, a more immersive mediated 

environment, which contains more stimuli, will draw more of the user’s attention and cognitive 

resources to build a better mental representation of the mediated environment than will a less 

immersive mediated environment. 

Experimental and biological evidence also supports the argument that immersion, which 

leads to a greater attentional allocation to the mediated environment, is a contributor to spatial 

presence experiences. For example, previous research found that the quality of visual immersion 

in video games had an impact on players’ construction of spatial situational models (Hofer, Sele, 

& Wirth, 2013). Another lab experiment on the technological features of video games also found 

that of immersion had an impact on users’ spatial presence experiences (McMahan, Bowman, 

Zielinski, & Brady, 2012). Research using electroencephalography further suggests that 

participants invest more attentional resources in response to a more visually immersive 

experience when playing video games or doing spatial navigation tasks (Havranek, Langer, 

Cheetham, & Jancke, 2012; Kober et al., 2012).  

20 

Based on the theoretical framework of the model of spatial presence experiences (Hofer 

et al., 2012; Schubert, 2009; Wirth et al., 2007) and empirical evidence on the effects of 

immersion on presence experiences, this study proposes the following hypothesis: 

H2: In the context of an exergaming intervention, individuals in the high immersion 

condition will have a stronger sense of spatial presence than those in the low immersion 

condition. 

Previous research indirectly supports that spatial presence is positively associated with 

cognitive improvement and suggests that future research should directly measure the impact of 

cognitive simulation during exergame interventions (Monteiro-Junior et al., 2016).  A 

randomized clinical trial involving an exergaming intervention found that participants’ mental 

effort while engaging with virtual-reality stationary biking, which also included being in a 

mediated environment, significantly improved their cognitive functions after the intervention 

compared to those of people who used a traditional stationary bike (Anderson-Hanley et al., 

2012). To be more specific, the mental exercise associated with the exergame was linked to 

spatial navigation, prediction of obstacles, and decision making. Therefore, when playing an 

exergame, players require additional attentional resources, which are highly relevant to executive 

functions.  

Biological evidence also supports the argument that the spatial presence elicited by 

immersive technologies is associated with cognitive improvement. Prior studies found that, 

compared to simply doing similar traditional exercises, virtual reality exercises can help increase 

institutionalized older adults’ functionality of the brain circuits linked to cognition, including the 

hippocampus, caudate nuclei, frontal and parietal cortex, and cerebellum (Monteiro-Junior et al., 

2016). Those brain circuits are associated with working memories, spatial navigation, problem-

21 

solving, and decision making. Moreover, research using functional neuroimaging also found that 

executive functions (i.e., selective attention and task management) activate the dorsolateral 

prefrontal cortex (DLPFC) (Diamond, 2015). The DLPFC was found to play an important role in 

the control of the experience of spatial presence (Gamito et al., 2017), and the right-DLPFC is 

specifically associated with visual-related cognitive functions (Lo Priore et al., 2003). Therefore, 

the activation of the DLPFC is strongly related to both executive functions and spatial presence. 

Based on the biological and empirical findings, the current study proposes the following 

hypothesis: 

H3: When immersion has a main effect on cognitive functions, spatial presence will 

mediate the impacts of immersion on older adults’ cognitive improvement in (a) 

inhibitory control, (b) cognitive flexibility, and (c) working memory. 

2.5. Cognitive Load, Spatial Presence, and Executive Functions 

As previous sections mentioned, cognitively engaging tasks have been regarded as one of 

the major contributors to cognitive improvement in the context of exergaming (Barcelos et al., 

2015; Best, 2013). However, the types and characteristics of tasks in exergames that contribute 

to older adults’ cognitive improvement are still unknown (Ogawa et al., 2016).  Furthermore, 

prior research on media information processing found that not all task loads are equal regarding 

their demand on users’ cognitive resources (Wang & Duff, 2016). Therefore, to better understand 

the effects of different types of cognitive loads while exergaming, it is important to distinguish 

the effects of different task loads on executive functions.  

Regarding the role of cognitively engaging tasks in the process of exergaming, scholars 

have been paying attention to the impacts of cognitive load and mental effort during exergaming. 

Nevertheless, the definitions and manipulations of cognitive load are not consistent across 

22 

studies. For example, some studies focused on task-relevant (gaming relevant) mental effort  

(e.g., Anderson-Hanley et al., 2017; Best, 2012) while others focused on task-irrelevant (non-

gaming relevant) tasks (e.g., Radovanović et al., 2014). The different conceptualization and 

operationalization of cognitive load in exergaming research may have contributed to the 

inconsistent findings regarding the effects of cognitive tasks on executive functions after 

exergaming (Ogawa et al., 2016). This study uses perceptual load theory (Lavie, 1995, 2005) as a 

framework to investigate the mechanisms underlying the effects of cognitive load on executive 

functions in the context of exergaming. 

According to the perceptual load theory (Lavie, 1995, 2005), the ability to remain 

focused on a task and ignore irrelevant distractors is strongly associated with people’s cognitive 

functions. Moreover, when dealing with task-relevant and task-irrelevant load, people will 

experience two types of selective attention mechanisms associated with the increase of 

perceptual load: early and late selection. In the early selection condition, if task-relevant stimuli 

require a high perceptual load, people will not be able to process information about task-

irrelevant information (i.e., distractors or interferences). Furthermore, during the early perceptual 

selection process, people will invest their entire cognitive load into the main task, which requires 

high cognitive load (Lavie, 1995, 2005). People with a high task-relevant load will allocate most 

of their attention and cognitive resources to process task-relevant information. In the late 

selection condition, if task-relevant stimuli require a low perceptual load, people’s cognitive 

resources will spill over to task-irrelevant information until the cognitive capacity is exhausted 

(Lavie, 1995, 2005). Therefore, when dealing with high task-irrelevant load in the low task-

relevant load situation, people will have to invest more cognitive efforts, such as inhibitory 

23 

control, to keep the task-unrelated distractors from interfering with the task-relevant task. They 

will process both the task-relevant and task-irrelevant information.  

Perceptual load theory (Lavie, 1995, 2005) suggests that when people are in a high task-

irrelevant load condition, their attentional and cognitive resources will be used to prevent 

distractors. When players are forced to manage distractions during the process of exergaming, 

they will have to split their attentional and cognitive resources between both types of tasks, 

which requires the use of their cognitive functions. In this sense, dealing with a high level of 

high task-unrelated distractors in the low task-relevant load condition will improve cognitive 

performance.  

However, the process model of spatial presence experiences (Wirth et al., 2007) predicts 

that people who cannot devote their full attentional resources to the mediated environment will 

be less likely to experience spatial presence. In the context of exergaming, when people play 

exergames in a high task-relevant load condition, such as dealing with a hard task within the 

game, they will spend more attentional and cognitive resources on the assigned task in the 

mediated environment, which will lead to a stronger level of spatial presence and improved 

performance in executive functions compared to those who have a lower level of spatial presence 

due to low task-relevant load. This assumption is also consistent with previous empirical findings 

on task-relevant load and cognitive improvement (Anderson-Hanley et al., 2017). Moreover, 

empirical evidence also shows that task-related stimulus should boost neural pathways by 

recruiting mental resources to process the information, which will also activate executive 

functions (Mansouri, Tanaka, & Buckley, 2009). Players have to actively and intensely process 

information from the mediated world, which makes people devote themselves more to the media, 

foster a sense of presence, and then activate larger cognitive resources. Therefore, people who 

24 

experience high task-irrelevant load will be less likely to experience spatial presence and its 

subsequent cognitive benefits.   

Based on the theoretical framework of perceptual load theory and empricial evidence, the 

current study proposes one more hypothesis. Furthermore, one research question is proposed to 

investigate whether task-relevant and task-irrelevant load have different impacts on cognitive 

improvement. Figure 1 shows the proposed model and the research question and hypotheses of 

the current research. 

H4: In the context of an exergaming intervention, individuals in the high task-relevant 

load condition will report a stronger sense of spatial presence compared to people in the high 

task-irrelevant condition.   

RQ1: In the context of an exergaming intervention, is there a difference between 

individuals in the high task-relevant load condition and those in the high task-irrelevant 

condition in terms of participants’ performance in (a) inhibitory control, (b) cognitive flexibility, 

and (c) working memory?  

 

Figure 1: Research Question and Hypotheses of the Current Research 

 

25 

CHAPTER 3: METHODOLOGY 

 

The goal of the current research is to examine the impacts of level of visual immersion, 

types of cognitive load, and feelings of spatial presence on older adults’ executive functions 

during an exergame-based intervention. A four-week exergame-based intervention, which  

consisted of eight 20-minute exergame sessions (two sessions per week), was designed to test the 

hypotheses and research questions of the current research. Each individual session length was 20 

minutes, which is similar to session length in previous research on physical activity and 

executive functions (e.g., Chu, Chen, Hung, Wang, & Chang, 2015; Kirk, MacMillan, Rice, & 

Carmichael, 2013). The intervention experiment was a 2 (task-relevant load vs. task-irrelevant 

load) × 2 (low immersion vs. high immersion) between-subject factorial design, and participants 

were randomly assigned to one of the four groups. 

3.1. Participants  

Exergaming had a medium effect size (d= 0.50) for executive functions across studies 

(Anderson-Hanley et al., 2012). Twenty-eight subjects were randomly assigned into the four 

conditions with repeated measures (four times), which reached a statistical power of 0.85 and 

passed the sufficient statistic power of 0.80 (Cunningham & McCrum-Gardner, 2007). Most of 

the studies examining the effect of exergaming on cognitive function had a small sample size (n= 

8 to 10 per condition; e.g., Anderson-Hanley et al., 2017). The current research recruited a total 

of forty-one participants (eight in the pilot test and thirty-three in the 4-week training) who were 

at least 50 years old.  

Recruitment was conducted mainly through the community SONA recruiting pool, a paid 

recruitment system managed by the College of Communication and Art Sciences at Michigan 

26 

State University. During the time of data collection, there were 405 total individuals in the 

recruiting pool: 230 individuals between 51 and 55 years old, 66 individuals between 56 and 60 

years old, 87 individuals between 61 and 70 years old, and 22 individuals more than 70 years 

old. Besides recruiting thirty-two participants from the Community SONA recruiting pool, nine 

participants were recruited by snowball sampling in the Lansing and East Lansing areas.  

Regarding the inclusion criteria, this study recruited subjects who were at least 50 years 

old. As previous chapters stated, there are two reasons for using the age of at least 50 as an 

inclusion criterion: First, peoples’ cognitive ability may start declining at around 50 years old 

(Albert & Heaton, 1988). Second, adults aged 50–64 are one of the target populations for 

preventive care services, but 28% of this population are physically inactive (Watson et al., 2016). 

For the exclusion criteria, people who have cognitive disabilities, such as Alzheimer’s disease or 

Parkinson’s disease, were excluded. Also, people who have difficulty performing physical 

activities were excluded from recruitment.  

 All participants were encouraged to complete the 4-week exergaming intervention, 

including a total of eight sessions of exergaming (two sessions per week). Participants played an 

exergame for 20 minutes and were instructed to do cognitive measures in the first, fourth, and 

eighth sessions. For the incentives, participants received 30 US dollars for completing the first 

week of sessions (15 dollars for each session in the first week), 35 US dollars for the second 

week (after the second session of the week), 40 US dollars for the third week, and 45 US dollars 

for the last week (after the second session of the week). Each participant received up to 150 US 

dollars total as an incentive. 

 

 

27 

3.2. Stimuli and Intervention Design 

The stimulus for the study was Fruit Ninja, which is an active video game in which 

players use arm and hand movements to swing virtual swords to slice fruit (see Figure 2). This 

game is designed to promote players’ physical activity via slicing, juggling, and skewering fruit. 

Previous research also used a similar virtual reality exergame, which requires players’ 

movements to match the programmed gestures or movements, to improve older adults’ cognitive 

reaction (Pachoulakis, Papadopoulos, & Ieee, 2016). Therefore, Fruit Ninja was selected as the 

stimulus for the current study after a pilot test of manipulations (please see Section 3.6: Pilot 

Test of Manipulations for more details).  

This game, which can be played in both VR (Fruit Ninja VR) and non-VR (Fruit Ninja 

Kinect) environments, shares the same rules in both conditions. Both versions have several types 

of difficulties and complexities and require varying levels of mental effort and cognitive load 

when playing the game. For example, in the Zen mode, players can just slice fruit without 

thinking about any rules, but players in the Survival mode have to avoid accidentally slicing 

bombs from flying cannons. The various Fruit Ninja game modes provide this study with an 

excellent manipulation of in-game task load.  

 

Figure 2: Fruit Ninja in Virtual Reality 

28 

 

The intervention included two manipulations—visual immersion (high vs. low) and 

cognitive load (intrinsic: task-relevant vs. extrinsic: task-irrelevant). Regarding the levels of 

visual immersion, there were two conditions in the study—virtual reality (VR) and 2D screen 

conditions. In the VR (Fruit Ninja VR) condition, participants used a head-mounted display 

(Oculus Rift) to play the exergame. In the non-VR (Fruit Ninja Kinect) condition, participants 

used a 27-inch LCD monitor to play the exergame. When playing the 2D Fruit Ninja Kinect 

version, participants were instructed to stand in front of the television (1.5–2 m away) and use 

the motion-tracking Kinect controller to play the game. In the Fruit Ninja VR version, 

participants were asked to play the game with a VR head-mounted display and motion-tracking 

Oculus Touch controllers.  

Regarding the manipulation of cognitive demand, there were two types of cognitive load 

in the intervention: high task-relevant load and high task-irrelevant load. For the high task-

relevant load condition, participants were asked to play the game at an advanced level, which 

required them to invest more task-relevant load on slicing the fruit in more complicated ways. 

People in the high task-relevant load were asked to play the Survival mode, which required 

higher in-game load.  

For the high task-irrelevant load condition, this study used one of the most cited 

manipulations of the task-irrelevant load: distractors or interference (Lavie, Hirst, de Fockert, & 

Viding, 2004). In the task-irrelevant load condition, participants were asked to remember a 7-

digit (7±2) number as a distractor when they played the exergame at a basic level (i.e., Zen 

mode). Previous studies suggest that older adults showing no deficits in short-term memory can 

hold about 7 ± 2 digits if they rehearse the numbers in their mind (Glisky, 2007). The exact digits 

of the number were tested in the pilot study by using the forward digit span test.  

29 

The only difference between high task-relevant load and high task-irrelevant load 

conditions was whether the complexity of the rules was task-relevant or not. The frequencies of 

arm and hand movements between two modes were similar. Therefore, the level of energy 

expenditure should be similar across conditions. This study used heart rate sensors (Fitbit Charge 

2 Activity wristbands) to control for the level of energy expenditure when analyzing the data. To 

make sure people in the high task-relevant and task-irrelevant conditions had similar levels of 

mental effort, the current study adjusted the length of the numbers based on participants’ self-

report of mental effort (see Section 3.6: Pilot Test of Manipulations for more details). 

Participants were randomly assigned to one of the four conditions based on the results of an 

online random number generator before they arrived at the lab for the first time. All participants 

were asked to play either Fruit Ninja VR or Fruit Kinect with motion-based controllers.  

3.3. Manipulation Check and Test for Confounding 

There were two manipulations in this study: (1) levels of visual immersion and (2) types 

of task load. To make sure the manipulations worked as expected, a manipulation check and a 

test for confounding were included in the study. For levels of visual immersion, participants were 

asked to rate five questions: (1) “how good was the quality of the visual immersion in the 

game?”, (2) “how much did the visual display quality interfere or distract you from performing 

assigned tasks or required activities?”, (3) the extent to which physical reality was shut out, (4) 

the breadth of range of sensory modalities accommodated, and (5) the degree of richness and 

information content of the displays. All responses were recorded using a 7-point Likert-like 

scale. All questions were adopted based on the definitions of technological immersion from 

previous research (Cummings & Bailenson, 2016; Schubert, Friedmann, & Regenbrecht, 2001; 

Witmer & Singer, 1998). 

30 

For types of task load, the current study used a three-item measure for mental effort 

adopted from previous studies on cognitive load and perceived mental effort (Leppink, Paas, Van 

der Vleuten, Van Gog, & Van Merriënboer, 2013) as the test for confounding. First of all, to 

check whether participants in two types of task-load conditions invested a similar amount of 

mental effort, participants were asked “How much mental effort did I invest during the gameplay 

that just finished?” Also, two questions were used to check the strength of task-relevant and task-

irrelevant load. Participants in both conditions were asked “How difficult was the game I just 

finished?” and those in the task-relevant load condition were asked an additional question: “How 

difficult was the memory task that I just finished?” These items cover different types and 

dimensions of mental effort. Participants rated all responses to questions from 1 (very, very low 

mental effort) to 9 (very, very high mental effort).  

People in the high task-relevant load condition were expected to score higher on “How 

difficult was the game that I just finished?” than people in the task-irrelevant load condition. 

Also, the scores of participants in the task-relevant load condition on “How difficult was the 

game that I just finished?” should also be similar to the scores of participants in the task-

irrelevant load condition on “How difficult was the memory task I just finished?” People in both 

conditions should have similar scores on “How much mental effort did I invest in the gameplay 

that just finished?” 

3.4. Outcome Variables  

Spatial presence. The current study argues that spatial presence mediates the relationship 

between the improvement of executive functions and (1) visual immersion and (2) cognitive 

load. For the measures of spatial presence experiences, participants were asked to rate their 

subjective experience of spatial presence during the gameplay. The measures are adopted from 

31 

the MEC Spatial Presence Questionnaires (MEC-SPQ; Vorderer et al., 2004). The MEC-SPQ 

scales have been empirically tested across diverse media settings (Hartmann et al., 2014) and 

theoretically tested via structural equation modeling (Hofer et al., 2012). The variables of the 

spatial presence experiences include attention allocation, spatial situational model, self-location, 

and possible action. Attention allocation is accessed using eight items (e.g., “I devoted my whole 

attention to the gaming environment”). The formation of a spatial situation model is assessed 

using eight items (e.g., “I was able to imagine the arrangement of the space very well”). The 

feeling of spatial presence includes two scales: self-location (e.g., “I had the feeling that I was in 

the middle of the action rather than merely observing”) and possible action (e.g., “I had the 

impression that I could act in the environment of the presentation”). All the items (see Survey 

Instrument) were rated on a 7-point Likert-like scale (1 = strongly disagree and 7 = strongly 

agree). 

Executive functions. A systematic review found that studies examining the impacts of 

exergaming on older adults’ cognitive benefits mainly measured participants’ executive 

functions via subjects’ reaction times before and after the exergaming interventions (Ogawa et 

al., 2016). In previous studies, measures of executive functions can be divided into three 

categories: shifting (requiring cognitive flexibility), inhibition, and working memory 

(Eggenberger, Wolf, Schumann, & de Bruin, 2016). Consistent with previous studies, the 

measures for executive functions in this study include the Stroop test, trail-making task, and digit 

span. (Anderson-Hanley et al., 2017; Eggenberger et al., 2016). 

The Stroop test (Color Word Stroop with Keyboard Responding) is an evaluation of 

inhibitory control, which requires participants to state a color under several conditions while 

suppressing habitual responses related to the conditions. The Stroop test is used to measure 

32 

executive control by response inhibition. A 40-item version of the Stroop test was administered, 

which is consistent with previous studies (e.g., Barcelos et al., 2015). This test involves the use 

of a screen paired with a computer running the Inquisit 5.0 software and a keyboard with four 

highlighted colored buttons (G, R, B, & Y) which are matched with the test colors (green, red, 

blue, yellow; see Figure 3).  

The Trail-making task is a well-established and empirically-tested measure (Kayama et 

al., 2014). The trail-making task is used to evaluate deficits in executive functions with a special 

focus on cognitive flexibility. The trail making task consists of two tasks, a visual scanning task 

and a cognitive flexibility task. Participants have to scan the locations of twenty-five circles and 

then connect them in numerical order. The scores are calculated as the difference between the 

pretest and posttest (see Figure 4).  

 

 

 

 

Figure 3: Stroop Test 

Figure 4: Trail-making Task 

Digit Span is an evaluation of working memory. Participants first listen to a set of 

numbers; they then repeat the numbers in a forward order (digital span forward), listen to a set of 

numbers again, and then repeat the same set of numbers in a reverse order (digital span 

33 

backward). The digit span score is computed by using the number of correct trails backward 

divided by the number of correct trails forward. A higher score indicates better cognitive 

function. Similar to the Stroop test, this test involves the use of a screen paired with a computer 

running the Inquisit 5.0 software and a keyboard with numbers from 0–9 (see Figure 5). 

 

 

Figure 5: Digit Span Task 

 

All the outcome measures were administered at baseline, after a single bout, after the 

second week, and after the 4-week intervention. Randomized reaction time measures were used 

to minimize practice effects. The results before and after a single bout were used to assess 

whether there was an immediate impact of exergaming. The outcome measures after the second 

week and the 4-week intervention helped the researcher track the effects over the course of the 

intervention.  

These three cognitive measures, which have been commonly used to measure executive 

functions, are sensitive for the detection of significant effects of experimental manipulations and 

have good validity across neuropsychological studies (Sánchez-Cubillo et al., 2009). Previous 

34 

research has shown that these measures have good to excellent test-retest reliability between two 

or more data points (Calamia, Markon, & Tranel, 2013), which indicates a lower possibility of a 

practice effect. Furthermore, previous studies measured these three tests together when 

measuring older adults’ cognitive functions, and those researchers did not report any task fatigue 

(Anderson-Hanley et al., 2017; Barcelos et al., 2015). Following previous studies, the total 

length of the outcome measures data collection was around 10 minutes (3–5 minutes for 

executive function tests and 5 minutes for surveys of spatial presence and subjective cognitive 

functions). The current study randomized these three tasks to avoid potential confounding 

factors. 

3.5. Control Variables 

Control variables include participants’ demographic information (e.g., gender, age, 

race/ethnicity, and education level) and their experiences with virtual reality headsets (e.g., 

Oculus Rift, HTC Vive, and PlayStation VR). Concerning their VR experience, participants 

answered questions such as “Have you ever used VR technologies before (e.g., Oculus Rift, HTC 

Vive, PlayStation VR, and other smartphone VR headsets)?” Participants rated their experience 

from 1 (never) to 7 (very often). This study used heart rate sensors (Fitbit Charge 2 Activity 

wristbands) to measure participants’ energy expenditure as a control variable for the strength of 

physical activity. Participants’ gameplay scores were controlled in the data analysis. 

3.6. Pilot Test of Manipulations 

The purpose of the pilot test was to make sure both manipulations worked as expected. A 

total of eight participants (two participants per condition) were included in the pilot test. After 

playing the Fruit Ninja games in either the VR or the non-VR version, participants were asked to 

rate the quality of visual immersion. Ideally, people who played Fruit Ninja VR were expected to 

35 

report higher scores on the quality of visual immersion than those who played Fruit Ninja Kinect 

(non-VR version). Both Fisher’s exact test (Pett, 2015) and the Mann–Whitney U test (Nachar, 

2008), which are designed for small sample sizes, were employed in the current study to 

determine whether there was a significant difference between two conditions. If participants who 

played Fruit Ninja VR did not report a higher level of immersion than people playing Fruit Ninja 

Kinect, this study would have used one of the other two virtual reality-based exergames, 

VirZoom Arcade or Holodance.  

The Virzoom Arcade was one of the alternatives to the stimulus. The game requires 

players to use the pedals of the VirZoom bike to control their speed, complete missions, and 

compete with others on racetracks and battlegrounds. In the VR condition, people experience a 

360-degree immersive view. For example, VR players are able to look back at opponents in a 

racing game or look around in a flying game. The other option, Holodance, is an exergame 

designed for players to dance in a virtual environment. Holodance is intended to promote 

players’ physical activities via integrating dancing through a series of music videos. The themes 

of the Holodance game involve the player combating monsters such as dragons. Each song has 

several types of difficulties and complexities that require varying levels of mental effort and 

cognitive load when playing the game. In the VR Holodance game, participants are expected to 

hit virtual objects while synchronizing the movements to music. If Fruit Ninja VR had not 

provided participants with higher immersion than Fruit Ninja Kinect in the pilot study, there 

would have been another run of the pilot study using VirZoom Arcade or Holodance as the 

stimulus for the current research. 

For the manipulations of types of cognitive load, this study also measured participants’ 

perceived mental effort. Ideally, participants in the high task-relevant load condition were 

36 

expected to report the same level of mental effort as those in the high task-irrelevant load 

condition. If participants had a significant difference in their ratings of mental effort across 

conditions, the difficulty of the task-irrelevant condition was adjusted by decreasing or 

increasing the number of digits (from 7 to 5 or 9) to be remembered.  

The pilot data were collected during the first week of November 2017. A total of eight 

participants finished the pilot study. The results of Mann–Whitney U tests showed that 

participants in the VR condition reported a higher level of presence compared to those in the 

non-VR condition at the significance level (Z = –2.274, p = .023). People in both types of task-

load conditions reported a similar level of total mental effort during the gameplay (task-relevant 

load: 6.57 vs. task-irrelevant load: 6.67), and also reported investing a similar amount of mental 

effort into the assigned tasks between task-relevant load (higher difficulty game:4.29 out of 9) 

and task-irrelevant load (memory task: 4.67 out of 9) conditions. Therefore, both manipulations 

were successful in the pilot test. 

3.7. Process Evaluation 

Building from previous research on health promotion programs (e.g., Moore et al., 2015; 

Saunders, Evans, & Joshi, 2005) and exercise interventions for older adults (e.g., Ellard, Taylor, 

Parsons, & Thorogood, 2011), I developed a process-evaluation plan for determining whether the 

proposed intervention and program activities were implemented as intended. I focused on the 

following five elements: fidelity, dose delivered and received, reach, recruitment, and context. 

The elements and tools of my process-evaluation plan are listed in Table 1. 

37 

Table 1: Process-Evaluation Plan  

Component  Questions 

Measures (Sources) 

Timing of Collection  Data Analysis 

Fidelity 

To what extent is the intervention 
implemented as planned and 
consistently with the theoretical 
framework? 

Manipulation checks 
(subjects’ self-reported 
scores) 

Each training session  Compare the mean 
difference between 
groups. 

Dose 

Do participants receive the 
intended units of the intervention? 

Time spent on exergaming 
(experimenter’s logs) 

How often do the subjects come 
to the training sessions? 

How many subjects participate in 
at least half of possible sessions? 

Reach 

Does the demographic 
information of the subjects reflect 
the population characteristics in 
the Lansing and East Lansing 
areas?  

Recruitment  What are the barriers to recruiting 

subjects? 

Physical activity level (heart 
rate monitors and 
accelerometers) 

Attendance rate & frequency 
of coming to the training 
sessions (experimenter’s logs) 

Demographic information 
(subjects’ self-reports) 

 

Each training session  Calculate the average 

scores based on 
percentage of 
intended units 
included 

In the beginning of 
the first training 
session 

Compare the 
distribution of sample 
to the population in 
the target areas 

Experimenter(s) will 
document all the recruitment 
activities (experimenter’s 
logs) 

During recruitment 

Find the factors 
influencing 
recruitment from 
experimenter’s notes. 

Context 

What are factors influencing the 
exergaming intervention? 

1. Subjects’ demographic 
Information & familiarity 
with VR (self-reported).  

1. In the beginning of 
the first training 
session  

List the contextual 
factors as covariates 
in the data analysis. 

2. Enjoyment (self-reported) 
& scores of the gameplays 
(experimenter’s logs) 

2. After each training 
session 

38 

3.8. Procedure 

All participants were encouraged to complete the 4-week intervention, which consisted of 

eight sessions of exergaming. All participants were randomly assigned to one of the four 

experimental conditions when they signed up the study. After they signed up for the study, they 

were asked to finish a pre-screen survey, which included questions about their demographic 

information and previous experience with VR headsets. 

When they came to the lab, participants were asked to read the consent form first. After 

they agreed to participate in the study, they did a series of executive function tasks (Stroop test, 

Trail-making task, and Digit Span) before starting the first exergaming session. These results 

served as the baselines for participants’ scores on executive functions.  

Before participants started playing the exergame for a 20-minute session, they were asked 

to read the instructions for playing the game (please see Appendix) and then put on a fitness 

wristband, which measured their energy expenditure during the gameplay. If the participants 

were assigned to the high immersion (VR) conditions, they were asked to put on a head-mounted 

display and adjust the head-mounted display to the degree they feel comfortable before playing 

the game. They were told that they could ask to stop playing and then be dropped out of the 

study if they had motion sickness during the gameplay. Participants assigned to the low 

immersion conditions were asked to stand a proper distance (1-1.5 meters away from the screen) 

from the LCD monitor to play the game.  

Participants in the high task-irrelevant load condition were instructed to (1) memorize a 

7-digit number (the number of digits vary based on the results of the pilot study) through the 20-

minute gameplay session and (2) report the number after the gameplay. Participants in the high 

task-relevant load condition were instructed to play in the high-complexity mode (i.e., Survival 

 

39 

mode). Participants were asked to complete two 20-minute exergaming sessions a week for four 

consecutive weeks. For the measures and surveys, participants were asked to do three executive 

function tasks as well as answer questions about their spatial presence experiences and subjective 

perception of their cognitive functions before the first exergaming session. Participants then did 

the cognitive tasks and surveys after their first, fourth, and last sessions. After completing the 

first session participants could schedule a time for the next session if they decided to continue to 

participate. The total length of each session (including the gameplay and measures) was around 

35 minutes.  

In the other seven sessions, participants completed the same amount of exergaming (20 

minutes) each time. Regarding incentives for participation, subjects received $15 after their first 

experimental session. After finishing the second session of the first week, participants received 

another $15. From the second to the fourth weeks, participants received incentives after 

completing every two sessions ($35 for week 2, $40 for week 3, and $45 for week 4). After four 

weeks, participants who completed the eight exergaming sessions also had cognitive measures 

conducted at four time frames: one pretest and three posttests.  

3.9. Analytic Techniques 

Three analytic techniques were implemented in the present study. First, descriptive 

statistics were used to estimate each variable of interest and gain a better contextual 

understanding of the data. Second, a series of repeated-measures, mixed-linear, two-way 

ANCOVA (analysis of covariance) were used to examine two manipulations by time interaction 

effects. Age and sex were included as covariates in statistical analyses, given their potential 

effects reported in previous empirical evidence (Huntley, Gould, Liu, Smith, & Howard, 2015; 

Lam et al., 2013). Last, to provide a more conservative way to interpret the data, a series of 

 

40 

repeated ANCOVAs were conducted to address the issues of missing responses by using 

Intention-To-Treat (ITT) analysis. ITT analysis has been widely used to deal with 

noncompliance and missing outcomes in research using randomized controlled trials (Gupta, 

2011). Previous studies (e.g., Peng, 2009) suggest three approaches to analyze missing values: 

(1) substituting the mean, (2) carrying the pretest value forward, and (3) carrying the posttest 

value forward. The current study provides results using these three approaches.  

Regarding hypothesis testing, a series of repeated ANCOVA were the primary analytic 

tool used to examine the differences across the different conditions. In testing H1 and RQ1, the 

level of immersion and types of task load were the independent variables, and executive 

functions were the dependent variable, while age and gender were used as covariates. In testing 

H2 and H4, the level of immersion and types of task loads were independent variables, and 

spatial presence was the dependent variable, while age and gender were used as covariates.  

After testing H1, H2, H4, and RQ1, the current study then tested H3, which showed a 

mediation effect of spatial presence on immersion and executive functions by using PROCESS 

mediation analysis (Hayes, 2013). PROCESS has been used to examine the potential mediating 

role of spatial presence between the independent and dependent variables instead of examining 

the entire proposed model simultaneously, as in previous studies. Descriptive statistics, 

ANCOVAs, paired t-tests, and PROCESS tests were conducted via SPSS 25.0. 

 

 

 

41 

CHAPTER 4: RESULTS 

 

4.1. Demographics 

There was a total of 41 participants, including those who only participated in a single-

bout session (n=8) and those who finished all training sessions (n=33). Thirty-two participants 

were recruited from a paid-community research pool and nine were recruited through snowball 

sampling (e.g., other participants’ invitations). No significant difference in demographics was 

found between the two recruiting pools. The average time of playing video games per day was 

less than 30 minutes, and 60% of participants did not play video games at all. None of the 

participants had played either version of Fruit Ninja before the exergaming training. 

All the participants were invited to participate in the longer-term training after finishing 

the first single bout of exergame training. Eight participants were recruited for the single-bout 

session and served as the pilot test of the current study. Of the remaining 33 participants who 

intended to participate in longer-term training, one dropped out after the fourth session due to 

personal reasons. Overall, 32 participants finished all the training sessions and adhered to the 

minimum amount of training sessions (twice a week, 15–20 minutes per visit) over four weeks. 

The adherence rate (97%) of the current study was higher than those reported in similar studies, 

which normally ranged from 19 – 75% (Barcelos et al., 2015). 

Before participants started their first training session, they were asked to complete three 

cognitive measures, which served as the baselines of their cognitive performance. The results of 

ANCOVAs showed that there was no significant difference among the four groups in all three 

cognitive measures in the pretest (see Table 2).  

 

 

 

42 

Table 2: Demographics and Baseline Executive Functions 

VRTR (N=10)  VRTI (N=10)  NVRTR (N=10)  NVRTI (N=11)  AN
OV
A 
p 

SD 

SD 

SD 

SD 

M 

M 

M 

M 

 

 
Demographics 
Age 
Sex (Male) 
Education 
Active Level 
Adherent (%) 

58.80 
40% 
17.40 
4.40 
80% 

5.75 
 
3.31 
1.43 
 

62.90 
40% 
16.80 
4.90 
90% 

7.31 
 
2.78 
1.45 
 

62.5  10.97 
20% 
 
2.27 
16.50 
1.42 
3.70 
80% 
 

61.93 
10% 
16.01 
4.28 
73% 

116.29 

6.99  118.67 

-.20 
-.15 
6.30 

121.50  13.12  114.11  11.12 

Heart Rate 
 
Baseline Executive Function 
Trials Making 
Stroop AC 
Digit Span 
VRTR= VR + Task-Relevant Load 
VRTI= VR + Task-Irrelevant Load 
NVRTR= Non-VR + Task-Relevant Load 
NVRTI= Non-VR + Task-Irrelevant Load 
Active Level was measured on a 7-point scale (1: not active at all to 7: very active) 
Both Trail Making and Stroop AC used standardized scores 

.00 
.16 
5.54 

.86 
.96 
1.69 

.86 
1.31 
1.45 

-.04 
-.14 
5.10 

.21 
.12 
5.20 

.55 
.73 
1.34 

8.21 
 
2.59 
1.42 
 

9.71 

.59 
.30 
.98 
.33 
.82 

.14 

.73 
1.00 
1.97 

.67 
.85 
.36 

 

4.2. Manipulation Check and Confounding Testing 

 

There were two tests for manipulations in this study: a manipulation check (level of 

immersion) and a test for confounding (type of task load). Regarding level of immersion, 

participants in the VR condition reported a higher level of immersive quality compared to those 

in the non-VR condition across three posttests at a significance level. Therefore, the 

manipulation of immersion worked as expected (see Table 3).  

 

Table 3: Manipulation Check of Level of Immersion 

 
 
First Manipulation Check  
Second Manipulation Check 
Last Manipulation Check  

VR 

Non-VR 

 

Mean (N)  SD  Mean (N) 
6.27 (20) 
5.17 (21) 
5.31 (16) 
6.25 (17) 
6.21 (17) 
5.46 (15) 

.62 
.69 
.76 

SD  T-Test 
1.12  3.90 
.95  3.26 
1.06  2.26 

P 
.000*** 
.002** 
.027* 

 

43 

Regarding type of task load, the manipulation was to create two types of task load and 

induce a similar level of mental effort between two conditions. The confounding testing question 

asked participants about their mental effort during the gameplay and perceived difficulty of the 

game. Participants in both conditions (i.e., task-relevant and task-irrelevant load) reported a 

similar amount of mental effort across the three posttests, and those in the task-relevant load 

condition reported a higher level of perceived difficulty of the game than those in the task-

irrelevant load condition. Therefore, the results of these two questions suggested that the difficult 

game mode resulted in a similar amount of mental effort as the combination of the easy game 

mode and the assigned game-irrelevant task. In both condition, participants invested more mental 

effort and performed the assigned tasks better at time 2 and time 3 than they did in the time 1.  

To compare the perceived difficulty of assigned tasks (a more difficulty game mode vs. 

an additional memory task), participants in the task-irrelevant load condition were asked to 

answer another question about the perceived difficulty of the memory task. After comparing the 

scores for the perceived difficulty of the game in the task-relevant condition and the scores for 

the perceived difficulty of the assigned memory task in the task-irrelevant load condition, the 

results showed that participants in both conditions reported a similar amount of perceived 

difficulty of the tasks. Therefore, the amount of total mental effort during game play and 

perceived difficulty of the assigned tasks were similar between two conditions (see Table 4). 

 
Table 4: Confounding Testing of Types of Task Loads 

 
First Posttest 
How much mental effort you invested 
during the gameplay you just finished? 
How difficult the game that you just 
finished was?  
1How difficult the memory task that 
you just finished was1? 

Task-Relevant 
Task-Irrelevant 
Mean(N)  SD  Mean(N)  SD 
6.20 (20)  1.85 

6.90 (21)  1.64 

T-Test   P 
 
1.11 

 
.27 

4.90 (20)  2.10 

3.05 (21)  2.60 

2.50 

.02* 

4.90 (20)  2.10 

5.86 (21)  1.32 

1.76 

.08 

 

44 

Table 4: Confounding Testing of Types of Task Loads (cont’d) 

5.88 (16)  2.21 

3.94 (17)  2.63 

2.55 

.01* 

 
  .69 

5.88 (16)  2.21 

5.12 (17)  2.06 

 
7.25 (16)  1.57 

 

 
7.59 (17)  1.26 

 

Second Posttest 
How much mental effort you invested 
during the gameplay you just finished? 
How difficult the game that you just 
finished was?  
1How difficult the memory task that 
you just finished was? 
Last posttest 
How much mental effort you invested 
during the gameplay you just finished? 
How difficult the game that you just 
finished was? 
1How difficult the memory task that 
you just finished was? 
1Participants in the task-relevant load condition did not have this question, so the results of t-
test is the comparison between difficulty of games vs. difficulty of memory tasks across two 
groups. 

 
7.53 (16)  2.03 

 
7.75 (16)  1.29 

6.75 (16)  1.77 

3.73 (16)  2.34 

6.75 (16)  1.77 

6.13 (16)  2.26 

 
  .35 

.00*
** 
.39 

4.64 

  .85 

 

 

 
.49 

 
.71 

1.02 

.31 

 

4.3. Overview: Effects of VR and Types of Task Load on Cognitive Improvement  

The first hypothesis and research question investigated the impact of level of immersion 

and type of task load on cognitive improvement, including Stroop, trail-making, and backward 

digit span. A series of repeated ANCOVAs were conducted to examine whether there were 

differences in cognitive improvement among the four conditions across three posttests as an 

overview of the effects of the interventions.  

4.3.1. Overview 

The exergaming training included eight sessions, and each session was between 15–20 

minutes. Before participants started playing the exergames in the first session, they were asked to 

complete a pretest survey and three cognitive measures (i.e., Stroop, trail-making and backward 

digit span). These cognitive measures served as their baseline cognitive performance. Over the 

course of the four-week training, they were asked to do posttests (a survey and three cognitive 

measures) in three different time frames: after the first single-bout session, after two weeks, and 

 

45 

after four weeks. Forty-one participants completed the first session, 33 completed four sessions, 

and 32 completed all eight training sessions, resulting in a final sample of 32 for data analysis. 

Three sets of repeated measures ANCOVAs were conducted for the executive function measures 

while controlling for age, gender, and the baseline results of the executive function measures (see 

Table 5 & 6). 

 

Table 5: Repeated ANCOVAs of Outcome Measures Across Three Posttests 

Stroop AC (N=28) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Trail Making (N=28) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Backward Digit Span (N=30) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 

 

 
 
 
 
 
 
 
 

 

df 
2 
2 
2 
2 
2 
2 
2 
42 
df 
2 
2 
2 
2 
2 
2 
2 
42 
df 
2 
2 
2 
2 
2 
2 
2 

 

 

F 

2.38 
1.22 
2.00 
1.46 
4.54 
1.37 
3.17 

F 
.84 
.69 
.70 
.09 
3.39 
1.68 
4.42 

F 

1.37 
1.76 
.898 
.552 
1.24 
.484 
3.92 

 

 

P 
.105 
.305 
.148 
.243 
.016* 
.263 
.052+ 

P 
.439 
.506 
.500 
.911 
.043* 
.198 
.018* 

P 
.263 
.184 
.414 
.580 
.299 
.620 
.027* 

Partial Eta2 
.102 
.055 
.087 
.065 
.178 
.062 
.131 

 
Partial Eta2 
.038 
.032 
.032 
.004 
.139 
.074 
.174 

 
Partial Eta2 
.056 
.071 
.038 
.023 
.051 
.021 
.146 

 

 

 

46 

Table 6: Estimated Marginal Means of Outcome Measures Across Three Posttests 

S.D. 
.351 
.342 
.333 
.289 
.282 
.274 
.329 
.320 
.312 
.396 
.385 
.375 

Mean 
-.164 
-.090 
.717 
.163 
-.426 
.077 
.142 
-.520 
-.505 
.145 
.795 
-.253 

Stroop AC (n=28) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-relevant load (Time 3) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
VR + Task-irrelevant load (Time 3) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-relevant load (Time 3) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 3) 
*Covariates appearing in the model are: Age = 62.29, Gender = .32 
Trail Making (N=28) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-relevant load (Time 3) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
VR + Task-irrelevant load (Time 3) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-relevant load (Time 3) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 3) 
*Covariates appearing in the model are: Age = 62.50, Gender = .28 
Backward Digit Span (N=30) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 

Mean 
.151 
.358 
-.516 
.011 
-.378 
.046 
.317 
.029 
.622 
-.181 
-.162 
.114 

S.D. 
.220 
.191 
.203 
.184 
.159 
.170 
.228 
.197 
.210 
.236 
.204 
.218 

S.D. 
.436 
.501 

Mean 
5.645 
6.369 

VR + Task-relevant load (Time 3) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
VR + Task-irrelevant load (Time 3) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-relevant load (Time 3) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 3) 
*Covariates appearing in the model are: Age = 61.93, Gender = .27. 

6.205 
6.305 
5.817 
6.369 
6.180 
5.941 
6.573 
5.472 
6.685 
7.338 

.479 
.387 
.445 
.426 
.440 
.506 
.484 
.492 
.565 
.541 

 

 

47 

 

ULCI 
.565 
.621 
1.408 
.764 
.160 
.647 
.826 
.146 
.143 
.968 
1.596 
.527 

ULCI 
.608 
.754 
-.094 
.393 
-.047 
.399 
.791 
.440 
1.059 
.309 
.262 
.567 

ULCI 
6.547 
7.406 

7.196 
7.106 
6.737 
7.250 
7.091 
6.987 
7.574 
6.489 
7.854 
8.457 

LLCI 
-.894 
-.801 
.025 
-.439 
-1.012 
-.494 
-.542 
-1.186 
-1.153 
-.677 
-.006 
-1.032 

LLCI 
-.307 
-.039 
-.939 
-.371 
-.709 
-.307 
-.156 
-.381 
.185 
-.672 
-.587 
-.339 

LLCI 
4.743 
5.333 

5.213 
5.504 
4.896 
5.488 
5.269 
4.894 
5.572 
4.455 
5.516 
6.220 

 

Regarding the Stroop results (n=28), four participants did not finish one of the posttests 

successfully and were excluded from data analysis. A significant interaction effect of level of 

immersion × time was found on Stroop at the significance level. [F(2, 42) = 4.53; p = .016; ηp

2 

= .178]. The interaction effect suggested that people in the VR condition had a more improved 

performance in Stroop than those in the non-VR condition over the course of the training. 

Furthermore, a three-way interaction of level of immersion × type of task load × time was found 

on Stroop at a marginal significance level. The results suggested that the interaction effect of VR 

and task-relevant load on Stroop varied across three posttests, with those in the VR+ task-

relevant load condition showing a more improved performance than those in the other three 

conditions across different time frames [F(2, 42) = 3.16; p = .052; ηp

2= .131] (See Figure 6). 

 

0.8

0.6

0.4

0.2

0

-0.2

-0.4

-0.6

-0.8

 

Stroop

Baseline

Single Bout

Two Weeks

Four Weeks

VRTR

VRTI

NVRTR

NVRTI

Figure 6: Stroop Results after A Single Bout, Two Weeks, and Four Weeks 

 

Regarding the trail-making results (n=28), four participants did not finish one of the 

posttests successfully and were excluded from data analysis. Similarly, a significant interaction 

effect of level of immersion × time was found on the trail-making (n=28), which showed that 

 

48 

participants in the VR condition had a more improved performance than those in the non-VR 

condition at the significance level [F(2, 42) = 3.39; p = .043; ηp

2= .139]. A three-way interaction 

of immersion × task-relevant load × time was also found on the trail-making. The results 

revealed that the impact of the interaction VR + task-irrelevant load varied across three posttests, 

with those in the VR + task-relevant load condition showing a more improved performance 

among the four experiment conditions over the course of the training [F(2, 42) = 4.42; p = .018; 

ηp

2= .174] (See Figure 7). 

 

Trail Making

-.8000
-.6000
-.4000
-.2000
.0000
.2000
.4000
.6000
.8000

Baseline

Single Bout

Two Weeks

Four Weeks

VRTR

VRTI

NVRTR

NVRTI

Figure 7: Trail-Making Results after A Single Bout, Two Weeks, and Four Weeks 

 

Regarding the backward digit span (n=30), two participants did not complete one of the 

posttests and were excluded from data analysis. The results revealed a three-way interaction of 

level of immersion × type of task load × time at the significance level [F(2, 42) = 3.92; p = .027; 

ηp

2= .146]. The findings indicated that the interaction effects of immersion and task-irrelevant 

load varied across three posttests, with those in the non-VR + task-irrelevant load condition 

 

49 

showing a more improved performance in backward digit span compared with the other three 

groups (See Figure 8) 

Backward Digit Span

 

7.5

7

6.5

6

5.5

5

4.5

Baseline

Single Bout

Two Weeks

Four Weeks

VRTR

VRTI

NVRTR

NVRTI

Figure 8: Backward Digit Results after A Single Bout, Two Weeks, and Four Weeks 

 

In sum, level of immersion had an interaction effect with time in two of the three 

cognitive measures. Therefore, H1a and H1b were supported. Regarding the research question, 

there was no main effect of type of task load on participants’ cognitive performance. Type of 

task load and level of immersion had interaction effects on all three outcome variables over the 

course of the exergaming training. Specifically, the three-way interactions revealed that 

participants in the VR + task-relevant load condition had a more improved performance across 

times on the Stroop and trail-making results compared to other groups, and those in the non-VR 

+ task-irrelevant load condition had a more improved performance across times on backward 

digit span results compared to other groups.  

 

 

 

50 

4.3.2. Intention-To-Treat Analysis 

To provide a more conservative way to interpret the data, a series of repeated ANCOVAs 

were conducted to address the issues of dropping-out cases and missing responses by using 

Intention-To-Treat (ITT) analysis. ITT analysis has been widely used to deal with 

noncompliance and missing outcomes in research using randomized controlled trails (Gupta, 

2011). Previous studies (e.g., Peng, 2009) suggested three different approaches to analyze 

missing values: 1) mean substitution, 2) pretest value carried forward, and 3) posttest value 

carried forward. Therefore, the current study also provided results using these three approaches.  

A total of 33 participants who decided to participate in the four-week exergaming 

training were included in the analyses. Among the 33 participants, four had missing values for 

one of their posttests and one dropped out after two weeks. After applying these three different 

approaches for analyzing missing values, most of the results were still consistent with the 

previous section’s results using raw data (See Table 7). The main effect of level of immersion on 

Stroop and interaction effects on Stroop, trail-making, and backward digit span were either 

significant or marginally significant. However, the main effect of level of immersion on the trail-

making test was no longer significant despite showing the same pattern as the previous analyses.  

The results of ITT analyses showed that the impact of level of immersion and interaction 

effects of two manipulations were still significant or marginally significant even using stricter, 

conservative approaches to analyze the data. Therefore, it is safe to argue that immersion and its 

interaction with the task-relevant load had positive impacts on participants’ performance in 

Stroop and trail-making at either a marginal significance or significance level.  

To sum up, the ITT analyses showed that the patterns of main effect of level of 

immersion and interaction effects of level of immersion and type of task load were still the same 

 

51 

despite the effect sizes being smaller. However, ITT has also been criticized for being more 

likely to have type II errors due to its conservativeness (Gupta, 2011). Therefore, ITT analyses 

served as references for the main analyses of the current study instead of as standards for 

answering research questions and supporting or rejecting the hypotheses. 

 

Table 7: ITT Analyses of Outcome Measures Across Three Posttests 

 

Stroop AC (N=33) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 

Trail Making (N=33) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 

Backward Digit Span (N=33) 
Time 
Time × Baseline 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 

 

Mean  

substitution 

Pretest  
Carried 

Posttest 
Carried 

Partial 
Eta2  F 
.042 
.055 
.036 
.041 
.129 
.057 
.088 

1.60 
1.60 
1.44 
  .94 
3.88* 
1.59 
4.55* 

Partial 
Eta2  F 
.048 
.026 
.040 
.005 
.074 
.085 
.167 

1.51 
1.03 
1.18 
  .55 
1.06 
1.08 
5.20** 

Partial 
Eta2  F 
.026 
.020 
.017 
.023 
.050 
.014 
.090 

1.27 
2.71+ 
  .68 
  .52 
1.43 
  .94 
5.00** 

Partial 
Eta2  F 
.058 
.058 
.052 
.035 
.130 
.058 
.149 

1.02 
1.45 
  .86 
1.25 
3.85* 
1.71 
3.42* 

Partial 
Eta2  F 
.055 
.038 
.043 
.021 
.039 
.040 
.178 

1.02 
  .48 
  .80 
  .02 
2.15 
2.57 
3.79* 

Partial 
Eta2  F 
.047 
.094 
.025 
.020 
.055 
.035 
.161 

1.00 
  .92 
  .77 
  .51 
  .65 
  .41 
3.44* 

Partia
l Eta2 
.038 
.053 
.032 
.046 
.129 
.062 
.116 

Partia
l Eta2 
.038 
.018 
.030 
.001 
.076 
.090 
.127 

Partia
l Eta2 
.037 
.034 
.029 
.019 
.024 
.016 
.040 

F 
1.15 
1.52 
  .97 
1.10 
3.86* 
1.56 
2.52+ 

F 
1.30 
  .69 
1.08 
  .12 
2.09 
2.43 
5.20** 

F 
  .70 
  .54 
  .44 
  .61 
1.36 
  .38 
2.59+ 

52 

4.4. The Baseline-Posttest Comparisons at Different Time Frames 

Besides conducting ITT analysis, this study also conducted repeated analyses to compare 

three posttests with the baseline separately to better understand whether there was a difference 

between the baseline measures and posttests at different times — right after a single bout of 

exergaming, after two weeks, and after four weeks. These post-hoc analyses included results 

using both raw data and ITT analyses. 

4.4.1. First Posttest: After the first Single-Bout 

A total of 41 individuals participated in the single-bout training. A series of repeated 

measures ANCOVAs were conducted for each of three cognitive measures while controlling for 

age, gender, and the baseline results of the executive function measures (see Tables 8 & 9). 

 
Table 8: Repeated ANCOVAs of Outcome Variables After the First Single-Bout 
Stroop AC (n=39) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Trail Making (n=37) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Backward Digit Span (n=41) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 

SS 
.005 
.016 
.509 
.018 
.355 
.015 
25.787 
SS 
.002 
.009 
.018 
.246 
2.297 
.012 
13.753 
SS 
2.527 
.370 
3.327 
.178 
.019 
5.492 
42.342 

F 
.006 
.020 
.651 
.023 
.454 
.020 
 
F 
.004 
.021 
.041 
.555 
5.178 
.027 
 
F 
2.089 
.306 
2.750 
.147 
.015 
4.540 
 

df 
1 
1 
1 
1 
1 
1 
33 
df 
1 
1 
1 
1 
1 
1 
31 
df 
1 
1 
1 
1 
1 
1 
 

.936 
.889 
.426 
.881 
.505 
.889 
 

P  Partial Eta2 
.000 
.001 
.019 
.001 
.014 
.001 
 
P  Partial Eta2 
.000 
.001 
.001 
.018 
.143 
.001 
 
P  Partial Eta2 
.056 
.009 
.073 
.004 
.000 
.115 
 

.953 
.886 
.840 
.462 
.030* 
.869 
 

.157 
.584 
.106 
.704 
.902 
.040* 
 

 

53 

Table 9: Estimated Marginal Means of Outcome Variables After the First Single-Bout 

 

ULCI 
.430 
.656 
.543 
.646 
.826 
.716 
.892 
.775 

ULCI 
.339 
.351 
.978 
.293 
.412 
.728 
.601 
.124 

 
ULCI 
7.033 
6.616 
6.168 
6.605 
6.231 
6.924 
6.702 
6.163 

S.D. 
.357 
.365 
.323 
.330 
.325 
.331 
.332 
.339 

S.D. 
.239 
.196 
.273 
.224 
.228 
.186 
.238 
.195 

LLCI 
-1.023 
-.828 
-.773 
-.698 
-.495 
-.632 
-.458 
-.604 

Mean 
-.296 
-.086 
-.115 
-.026 
.165 
.042 
.217 
.086 

Stroop AC (n=39) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
*Covariates appearing in the model are: Age = 61.6923, Gender = .2821. 
*The means were standardized. The higher scores mean the better reaction. 
Trail Making (N=37) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
*Covariates appearing in the model are: Age = 61.7027, Gender = .2162. 
*The means were standardized. The higher scores mean the slower reaction. 
Digit Span (n=41) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
*Covariates appearing in the model are: Age = 61.9268, Gender = .2683. 
* The higher scores mean the better reaction.  

Mean 
6.005 
5.878 
5.171 
5.889 
5.226 
6.203 
5.726 
5.463 

Mean 
-.150 
-.049 
.422 
-.162 
-.052 
.347 
.115 
-.275 

S.D. 
.507 
.363 
.491 
.352 
.495 
.355 
.481 
.345 

LLCI 
-.638 
-.449 
-.135 
-.618 
-.517 
-.033 
-.372 
-.673 

LLCI 
4.977 
5.141 
4.174 
5.174 
4.220 
5.481 
4.749 
4.762 

 

 

 

 

 

Regarding the Stroop results, two participants did not finish the first posttest successfully. 

Therefore, they were excluded from data analysis (n=39). The results of the repeated ANCOVAs 

did not reveal any significant main or interaction effects of level of immersion and type of task 

load on Stroop.   

 

54 

Regarding the results of the trail-making test, four participants who did not finish the test 

were excluded from the data analysis (n=37). A significant interaction effect of type of task load 

× time interaction was found on the trail-making, which showed that participants in the task-

irrelevant load condition had a more improved performance than those in the task-relevant 

condition across two time points [F(1,31) = 5.18; p = .030; ηp

2= .143].  

Regarding participants’ performance on backward digit span (n=41), there was a 

significant three-way interaction effect of level of immersion × type of task load × time on their 

first posttest. Specifically, the three-way interaction revealed that participants in the VR + task-

irrelevant load condition improved their performance in the backward digit span the most among 

the four experimental conditions. The unexpected interaction relationship will be further 

illustrated in the discussion section [F(1,35) = 4.54; p = .040; ηp

2 = .115].  

 
Table 10: ITT Analyses of Outcome Measures After the First Single-Bout 

 

Mean  

substitution 

Pretest  
Carried 

Posttest  
Carried 

Stroop AC (N=41) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Loads 

Trail Making (N=41) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Loads 

F 
  .44 
  .47 
  .33 
  .00 
  .20 
  .53 

F 
  .01 
  .03 
  .00 
  .50 
4.56* 
  .14 

Partial 
Eta2 
.012 
.013 
.009 
.000 
.006 
.015 

Partial 
Eta2 
.000 
.001 
.000 
.014 
.115 
.004 

F 
  .07 
  .06 
  .49 
  .01 
  .40 
  .25 

F 
  .01 
  .03 
  .02 
  .49 
4.61* 
  .17 

Partial 
Eta2 
.002 
.002 
.014 
.000 
.011 
.007 

Partial 
Eta2 
.000 
.001 
.001 
.014 
.116 
.005 

F 
  .45 
  .49 
  .30 
  .01 
  .18 
  .61 

F 
  .13 
  .16 
  .01 
  .16 
3.42+ 
  .52 

Partial 
Eta2 
.013 
.014 
.008 
.000 
.005 
.017 

Partial 
Eta2 
.004 
.005 
.000 
.004 
.089 
.015 

 

55 

In general, level of immersion did not have any effect on outcome measures right after 

the first exergame training. Regarding the effects of type of task load, the task-irrelevant load 

had an impact on participants’ performance in trail-making. Furthermore, the task-irrelevant load 

also positively affected the backward digit span results for those in the VR condition and the 

task-relevant load negatively affected backward digit span for those in the non-VR condition. 

The results of ITT analyses were consistent with the initial analyses (See Table 10).   

4.4.2. Second Posttest: After the Two-Week Training (four sessions) 

Of the 41 participants who finished the first exergame training, eight decided not to 

participate in the longer-term training. A series of repeated ANCOVAs were conducted to 

investigate the impacts of two manipulations (See Tables 11 & 12). For the Stroop task results 

after the two-week training, two outliers whose scores were below or exceeding three standard 

deviations (Khng & Lee, 2014) were excluded from the data analysis (N=31). There was a three-

way interaction effect of level of immersion × type of task load × time [F(1,25) = 5.07; p = .033; 

ηp

2 = .175]. To be more specific, participants in the VR condition + task-relevant load and those 

in the non-VR condition + task-irrelevant load had the most improved performance in the second 

Stroop test among all experimental conditions. However, no effect of immersion and type of 

cognitive load was found on Stroop. 

Regarding the trail-making test, three participants were excluded from the data analysis 

due to incomplete results (n=30). Similar to the findings of trail-making in the first posttest, the 

results of the second posttest also revealed an interaction effect of type of task load × time. To be 

more specific, participants in the task-irrelevant condition showed a more improved performance 

than those in the task-relevant condition [F(1,24) = 8.55; p = .055; ηp

2= .151].  

 

56 

For the results of backward digit span, three participants did not complete the tests, and 

were thus excluded from the data analysis (n=30). The result showed a significant three-way 

interaction effect of level of immersion × type of task load × time on participants’ backward digit 

span scores in the second posttest [F(1,24) = 5.72; p = .026, ηp

2= .214]. Participants in the non-

VR + task-irrelevant load conditions had improved their performance more between the baseline 

and posttest measures than in the other three conditions. No effect of level of immersion and type 

of task load was found on backward digit span. 

 

Table 11: Repeated ANCOVAs of Outcome Variables After the Two-Week Training  
Stroop AC (n=31) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Trail Making (n=30) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Backward Digit Span (n=30) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 

P  Partial Eta2 
.098 
.088 
.021 
.001 
.004 
.175 
 
P  Partial Eta2 
.005 
.003 
.002 
.016 
.151 
.003 
 
P  Partial Eta2 
.252 
8.129 
.007 
.010 
.093 
.214 
 

SS 
3.015 
2.674 
.607 
.028 
.102 
5.878 
27.664 
SS 
.056 
.036 
.023 
.165 
1.824 
.028 
10.267 
SS 
6.704 
8.129 
.131 
.209 
2.042 
5.426 
19.924 

F 
2.616 
2.320 
.527 
.025 
.089 
5.099 
 
F 
.126 
.081 
.051 
.370 
4.086 
.064 
 
F 
7.066 
8.568 
.139 
.221 
2.152 
5.719 
 

df 
1 
1 
1 
1 
1 
1 
25 
df 
1 
1 
1 
1 
1 
1 
24 
df 
1 
1 
1 
1 
1 
1 
24 

.119 
.141 
.475 
.877 
.768 
.033* 
 

.726 
.779 
.824 
.549 
.055+ 
.803 
 

.015* 
.008** 
.713 
.643 
.157 
.026* 
 

 

 

 

 

57 

Table 12: Estimated Marginal Means of Outcome Variables After the Two-Week Training 

S.D. 
.411 
.376 
.341 
.312 
.422 
.386 
.373 
.341 

Mean 
-.459 
.006 
.194 
-.459 
.157 
-.574 
.162 
.893 

Mean 
.010 
.211 
.201 
-.354 
.282 
.282 
.135 
-.210 

Stroop AC (n=31) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Covariates appearing in the model are: Age = 62.9333, Gender = .3000. 
Trail Making (N=30) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Covariates appearing in the model are: Age = 62.1563, Gender = .2500. 
Digit Span (n=30) 
VR + Task-relevant load (Time 1) 
VR + Task-relevant load (Time 2) 
VR + Task-irrelevant load (Time 1) 
VR + Task-irrelevant load (Time 2) 
Non-VR + Task-relevant load (Time 1) 
Non-VR + Task-relevant load (Time 2) 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
Covariates appearing in the model are: Age = 62.2963, Gender = .2963. 

Mean 
5.832 
6.873 
5.174 
5.616 
5.598 
4.938 
5.558 
7.157 

S.D. 
.296 
.196 
.268 
.177 
-.635 
-.635 
.319 
.211 

S.D. 
.603 
.483 
.541 
.433 
.767 
.614 
.754 
.604 

LLCI 
-1.308 
-.769 
-.510 
-1.103 
-.714 
-1.370 
-.609 
.189 
 

 

LLCI 
-.598 
-.191 
-.349 
-.717 
.525 
.525 
-.520 
-.643 

 

 

LLCI 
4.579 
5.869 
4.049 
4.715 
4.003 
3.661 
3.990 
5.901 

 

 

ULCI 
.389 
.782 
.898 
.184 
1.029 
.223 
.933 
1.598 

ULCI 
.619 
.613 
.751 
.010 
.282 
.282 
.791 
.223 

ULCI 
7.085 
7.876 
6.299 
6.517 
7.193 
6.216 
7.126 
8.413 
 

 

Overall, the results of cognitive training after two weeks (four sessions) did not show any 

effect of level of immersion on executive functions. However, level of immersion started to have 

an interaction effect with the task-relevant load on cognitive performance while the impact of 

task-irrelevant load persisted in certain conditions. Task-irrelevant load still had an impact on the 

trail-making, but only at the marginal significance level. The results of ITT analyses were 

consistent with the initial analyses (See Table 13). 

 

 

58 

Table 13: ITT Analyses of Outcome Measures After the Two-Week Training 
 

Mean 

Stroop AC (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 
Trail Making (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 
Backward Digit Span (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 

 

substitution 
F 

Partial 
Eta2 
.052 
.050 
.010 
.014 
.001 
.124 

Partial 
Eta2 
.004 
.003 
.004 
.024 
.101 
.011 

Partial 
Eta2 
.202 
.235 
.008 
.002 
.056 
.155 

1.47 
1.42 
  .26 
  .38 
  .03 
3.84+ 

F 

  .11 
  .06 
  .30 
  .84 
3.04+ 
  .59 

F 

5.59* 
6.78* 
  .17 
  .05 
1.30 
4.09+ 

Pretest 
Carried 

F 

1.96 
1.78 
  .46 
  .30 
  .01 
4.56* 

F 

  .22 
  .15 
  .05 
  .13 
4.12+ 
  .86 

F 

5.43* 
6.58* 
  .17 
  .05 
1.24 
3.89+ 

Partial 
Eta2 
.069 
.062 
.017 
.011 
.000 
.144 

Partial 
Eta2 
.009 
.006 
.002 
.005 
.137 
.003 

Partial 
Eta2 
.198 
.230 
.008 
.002 
.053 
.150 

Posttest 
Carried 

F 

Partial 
Eta2 

2.14 
1.81 
  .70 
  .44 
  .11 
4.14* 

.073 
.063 
.025 
.016 
.004 
.140 

F 

Partial 
Eta2 

  .11 
  .07 
  .00 
  .52 
3.12+ 
  .11 

.004 
.003 
.000 
.019 
.132 
.004 

F 

Partial 
Eta2 

5.84* 
7.32* 
  .07 
  .03 
1.50 
4.58* 

.210 
.250 
.003 
.002 
.064 
.172 

4.4.3. Third Posttest: After Four-week Training (eight sessions) 

One participant dropped out of the exergame training after the forth session, and 32 

remained for the rest of the training sessions. A series of repeated ANCOVAs were conducted to 

investigate the impact of two manipulations (See Tables 14 & 15). In the last posttest, level of 

immersion × time interaction had a significant impact on the Stroop, which revealed that 

participants in the VR condition had a more improved performance than those in the non-VR 

condition [F(1,23) = 5.17; p = .034; ηp

2= .197]. Furthermore, a three-way interaction of level of 

 

59 

immersion × type of task load × time was found. The results showed that VR and task-relevant 

load varied between two times, with those in the VR + task-relevant load condition having a 

more improved performance in Stroop compared to other conditions [F(1,23) = 4.35; p = .049; 

ηp

2= .172]. 

Regarding the results of participants’ trail-making performance, an interaction effect of 

level of immersion × time was found in the last posttest, which showed that participants in the 

VR condition had a more improved performance than those in the non-VR condition [F(1,22) = 

4.88; p = .038; ηp

2= .181]. Different from the first and second posttests, type of task load no 

longer had an impact on participants’ trail-making performance. 

 

df 

P 

.009* 
.013* 
.078+ 
.034* 
.578 
.049* 
 

F 
8.146 
7.378 
3.433 
5.166 
.320 
4.350 
 

SS 
5.600 
5.072 
2.360 
3.551 
.220 
2.990 
14.436 

Table 14: Repeated ANCOVAs of Outcome Variables After the Four-Week Training 
Stroop AC (n=29) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Trail Making (N=28) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 
Digit Span (n=32) 
Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task Load 
Error (Time) 

SS 
.067 
.062 
.166 
1.687 
.158 
.312 
7.612 
SS 
.778 
.560 
.028 
2.856 
1.081 
.021 
32.772 

P 
.665 
.677 
.496 
.038* 
.507 
.352 
 

F 
.192 
.178 
.480 
4.876 
.456 
.903 
 

F 
.617 
.444 
.022 
2.266 
.858 
.017 
 

1 
1 
1 
1 
1 
1 
23 

1 
1 
1 
1 
1 
1 
22 

1 
1 
1 
1 
1 
1 
26 

P 
.439 
.511 
.882 
.144 
.363 
.897 
 

Partial Eta2 
.279 
.260 
.141 
.197 
.015 
.172 
 

Partial Eta2 
.009 
.008 
.021 
.181 
.020 
.039 
 

Partial Eta2 
.023 
.017 
.001 
.080 
.032 
.001 
 

df 

df 

 

 

60 

 

 

 

Mean 

S.D. 

Mean 

S.D. 

LLCI 

ULCI 

.418 
.339 
.342 
.277 
.476 
.386 
.432 
.350 

-.565 
.664 
.174 
.143 
.446 
-.466 
.094 
-.093 

LLCI 
-1.436 
-.041 
-.538 
-.434 
-.544 
-1.269 
-.805 
-.822 

Table 15: Estimated Marginal Means of Outcome Variables After the Four-Week Training 
Stroop AC (n=29) 
.305 
VR + Task-relevant load (Time 1) 
1.369 
VR + Task-relevant load (Time 2) 
.886 
VR + Task-irrelevant load (Time 1) 
.720 
VR + Task-irrelevant load (Time 2) 
1.437 
Non-VR + Task-relevant load (Time 1) 
.336 
Non-VR + Task-relevant load (Time 2) 
.993 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
.635 
Covariates appearing in the model are evaluated at the following values: Age = 63.4074, 
Gender = .3333. 
Trail Making (N=28) 
.620 
VR + Task-relevant load (Time 1) 
-.155 
VR + Task-relevant load (Time 2) 
.756 
VR + Task-irrelevant load (Time 1) 
.399 
VR + Task-irrelevant load (Time 2) 
.521 
Non-VR + Task-relevant load (Time 1) 
.841 
Non-VR + Task-relevant load (Time 2) 
.519 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
.397 
Covariates appearing in the model are evaluated at the following values: Age = 62.0313, 
Gender = .2813. 
Digit Span (n=32) 
7.084 
VR + Task-relevant load (Time 1) 
7.464 
VR + Task-relevant load (Time 2) 
6.409 
VR + Task-irrelevant load (Time 1) 
7.767 
VR + Task-irrelevant load (Time 2) 
7.193 
Non-VR + Task-relevant load (Time 1) 
6.216 
Non-VR + Task-relevant load (Time 2) 
6.356 
Non-VR + Task-irrelevant load (Time 1) 
Non-VR + Task-irrelevant load (Time 2) 
8.172 
Covariates appearing in the model are evaluated at the following values: Age = 62.6250, 
Gender = .2813. 

5.879 
6.396 
5.264 
6.752 
5.598 
4.938 
5.076 
7.038 

4.675 
5.327 
4.120 
5.737 
4.003 
3.661 
3.797 
5.903 

.007 
-.535 
.199 
.054 
-.070 
.475 
-.123 
-.001 

-.606 
-.915 
-.359 
-.292 
-.660 
.109 
-.766 
-.399 

.586 
.520 
.557 
.494 
.767 
.614 
.623 
.552 

.298 
.185 
.271 
.168 
.287 
.178 
.313 
.194 

S.D. 

LLCI 

ULCI 

 

 
Mean 

 

 

 

 

 

 

 

 

ULCI 

 

 

 

 

Regarding backward digit span, neither level of immersion nor type of task load had an 

impact on participants’ performance after four weeks of exergaming training. Furthermore, the 

interaction effects no longer persisted in the last posttest, which was different from the results of 

the first and second posttests. 

 

61 

To sum up, the level of immersion showed an impact on two out of three cognitive 

measures (i.e., Stroop and trail-making) in the posttest while the effect of task-irrelevant load 

was no longer significant. The results of ITT analyses were consistent with the initial analyses 

(See Table 16).  

 

Table 16: ITT Analyses of Outcome Measures After the Four-Week Training 

 

Mean 

Pretest 
Carried 

Posttest 
Carried 

F 

8.32** 
7.46* 
4.03+ 
5.07* 
  .23 
4.26+ 

F 

  .13 
  .13 
  .14 
5.20* 
  .01 
2.26 

F 

  .03 
  .57 
  .33 
  .60 
1.04 
  .32 

Partial 
Eta2 
.274 
.253 
.155 
.187 
.010 
.162 

Partial 
Eta2 
.000 
.000 
.001 
.163 
.000 
.076 

Partial 
Eta2 
.001 
.021 
.012 
.022 
.037 
.001 

F 

8.39** 
7.45* 
3.96+ 
5.21* 
  .26 
4.36* 

F 

  .00 
  .00 
  .00 
4.30* 
  .05 
2.79 

F 

  .26 
1.06 
  .27 
  .20 
  .45 
  .00 

Partial 
Eta2 
.276 
.255 
.153 
.192 
.012 
.166 

Partial 
Eta2 
.000 
.000 
.000 
.137 
.002 
.094 

Partial 
Eta2 
.010 
.038 
.011 
.008 
.016 
.000 

Stroop AC (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 
Trail Making (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 
Backward Digit Span (N=33) 

Time 
Time × Age 
Time × Gender 
Time × Immersion 
Time × Types of Task Load 
Time × Immersion × Types of Task 
Load 

 

 

 

 

substitution 
F 

8.08** 
7.19* 
4.17+ 
4.70* 
  .17 
3.96+ 

F 

  .00 
  .00 
  .00 
4.41* 
  .04 
2.75 

F 

  .05 
  .19 
  .54 
1.91 
2.04 
  .29 

Partial 
Eta2 
.269 
.246 
.159 
.176 
.008 
.152 

Partial 
Eta2 
.000 
.000 
.000 
.140 
.002 
.092 

Partial 
Eta2 
.002 
.008 
.022 
.074 
.078 
.012 

62 

4.5. Effects of VR and cognitive load on Spatial Presence Experiences 

To test the second and fourth hypotheses of the current study, a series of repeated 

ANCOVAs and ANCOVAs were conducted to examine the impact of level of immersion and 

type of task load on participants’ spatial presence experiences across four weeks. Both 

independent variables are categorical, and the dependent variable is a scale of continuous data. 

Therefore, repeated ANCOVA and ANCOVA would be a suitable tool for interpreting the 

results. Age and gender were included as covariates in the analyses (see Table 17).  

First of all, a repeated measure ANCOVA was conducted and there was no interaction 

effect of time and two manipulations for the within-subjects effects. For the between-subjects 

effects, the results showed that level of immersion had an impact on participants’ spatial 

presence experiences [F(1,26) = 186.17; p = .002; ηp

2= .302].. In other words, the impacts of 

immersion were consistent across three posttests, and participants in the VR condition reported a 

higher level of spatial presence than those in the non-VR condition. 

Next, a series of ANCOVAs were conducted to investigate the impacts of level of 

immersion and type of task load on spatial presence after a single bout (the first posttest), after 

two weeks (the second posttest), and after four weeks (the last posttest). In the first posttest, there 

was a main effect of immersion on participants’ spatial presence immediately after a single bout 

of exergame training [F(1,35) = 21.24; p = .000; ηp

2= .378]. However, no effect of type of task 

load was found in the first posttest.  

Similar to the first posttest, level of immersion had an effect on spatial presence in the 

second [F(1,27) = 16.33; p = .000; ηp

2 = .377] and last posttests [F(1,26) = 6.38; p = .018; ηp

2 

= .197]. However, type of task load still had no impact on spatial presence. 

 

63 

Overall, the results suggested that participants in the VR condition reported a higher level 

of spatial presence experiences than those in the non-VR condition. Type of task load was not a 

predictor of participants’ spatial presence experiences. Therefore, the second hypothesis was 

supported but the fourth hypothesis was not. 

 

 

SS 

SS 

5.290 
1.415 
186.171 
5.850 
37.988 
429.540 

3.875 
.082 
115.662 
2.982 
14.853 
190.586 
 

Table 17: ANCOVAs of Spatial Presence Experiences 
Across Three Posttests (N=32) 
Age 
Gender 
Immersion 
Types of Task Load 
Immersion × Types of Task Load 
Error  
Adjusted R2=.250 
The First posttest (N=41) 
Age 
Gender 
Immersion 
Types of Task Load 
Immersion × Types of Task Load 
Error  
Adjusted R2=.350 
The Second posttest (N=33) 
Age 
Gender 
Immersion 
Types of Task Load 
Immersion × Types of Task Load 
Error  
Adjusted R2=.289 
 The Third posttest (N=32) 
Age 
Gender 
Immersion 
Types of Task Load 
Immersion × Types of Task Load 
Error  
Adjusted R2=.109 

SS 
2.807 
2.899 
48.181 
6.393 
14.396 
196.266 
 

5.413 
.006 
92.504 
3.095 
5.371 
152.932 

df 
1 
1 
1 
1 
1 
26 
 

df 
1 
1 
1 
1 
1 
35 
 

df 
1 
1 
1 
1 
1 
27 
 
df 
1 
1 
1 
1 
1 
26 
 

SS 

 

 

 

 

64 

F 
.320 
.086 
11.269 
.354 
2.299 
16.521 
 

F 
.712 
.015 
21.241 
.548 
2.728 
 
 

F 
.956 
.001 
16.331 
.546 
.948 
 

 

F 
.372 
.384 
6.383 
.847 
1.907 
 
 

P 
.576 
.772 
.002** 
.557 
.141 
 
 

P 
.405 
.903 
.000*** 
.464 
.108 
 
 

P 
.337 
.974 
.000*** 
.466 
.339 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  P 
 
 
 
 
 
 
 

.547 
.541 
.018* 
.366 
.179 
 
 

Partial Eta2 
.012 
.003 
.302 
.013 
.081 
 
 
Partial Eta2 
.020 
.000 
.378 
.015 
.072 
 
 
Partial Eta2 
.034 
.000 
.377 
.020 
.034 
 

Partial Eta2 
.014 
.015 
.197 
.032 
.068 
 
 

 

 

4.6. Mediation Effects of Spatial Presence on Executive Functions 

To test the mediating role of spatial presence experiences in the relationship between two 

manipulations and cognitive improvement, a series of PROCESS mediation analyses (Hayes, 

2017; Model 4) were employed in each of the posttests separately. Age, gender, and baseline 

cognitive measures were entered as covariates and bootstrapping methods were used (5000 

samples). Regarding the first posttest, there was no main effect of immersion on cognitive 

outcomes. Therefore, no PROCESS mediation was conducted (see Table 18).  

A mediation effect was found in the second posttest, which was conducted after the two-

week training. The mediation model revealed a conditional mediation effect, which showed that 

playing the exergame in virtual reality with a task-relevant load influenced participants’ spatial 

presence experiences (b = 3.446, p = .001), which then influenced the speed of doing the Stroop 

test (b = .277, p = .049). In the task-relevant load condition, the indirect effect from the VR 

condition to spatial presence experiences and then to the Stroop was significant in the second 

posttest ([95% CI .002, .55]). 

Two more mediation relationships were found in the last posttest, which was conducted 

after all the training sessions had been completed. The first mediation model showed that level of 

immersion positively influenced participants’ spatial presence experiences (b = 2.463, p = .01), 

which then led to better Stroop results (b = .315, p = .007). The indirect effect from immersion to 

spatial presence experiences and then to Stroop results was significant (95% CI [.08, .55]). The 

second mediation relationship in the last posttest revealed that level of immersion positively 

influenced participants’ spatial presence experiences (b = 3.30, p = .005), which then led to a 

better performance in trail-making in the last posttest (b = -.124, p = .04). The indirect effect 

from VR to the trail-making test via spatial presence was significant (95% CI [-.24, -.01]). 

 

65 

To sum up, spatial presence experiences conditionally mediated the effect of level of 

immersion on Stroop in the second posttest and fully mediated the effect of level of immersion 

on Stroop and trail-making in the last posttest. In other words, when level of immersion had an 

impact on cognitive improvement, spatial presence experience was a mediator. There was no 

mediated effect of immersion on participants’ working memory. Therefore, H3a and H3b were 

supported. 

 

Table 18: PROCESS Mediation Analysis 

Stroop at the second posttest  
(N=17) 
Immersion → Stroop 
Immersion → Spatial Presence 
Direct Effect 
Indirect Effect 
Stroop at the third posttest  
(N=32) 
Immersion → Stroop 
Immersion → Spatial Presence 
Direct Effect 
Indirect Effect 
Trail Making at the third posttest 
(N=32) 
Immersion → Trail Making 
Immersion → Spatial Presence 
Direct Effect 
Indirect Effect 

t 
4.94 
-2.43 
-1.34 
2.11 

P 

.000*** 
.025 
.191 
.048* 

LLCI  ULCI 
4.89 
-.21 
.33 
.55 

2.00 
-2.76 
-.155 
.01 

β 
3.45 
-1.49 
-.61 
.28 

β 
2.46 
3.08 
.72 
.32 

se 
.70 
.61 
.45 
.13 

se 
.86 
.85 
.35 
.11 

t 
2.87 
3.64 
2.07 
2.88 

P 

.008** 
.001** 
.055+ 
.011* 

β 
se 
-.53 
.23 
3.30  1.06 
-.17 
.27 
.06 
-.12 

t 

-2.27 
3.12 
-1.04 
-2.20 

P 
.034* 
.005** 
.310 
.040* 

LLCI  ULCI 
4.22 
4.82 
1.46 
.55 

.71 
1.35 
-.02 
.08 

LLCI  ULCI 
.86 
.51 
.40 
-0.01 

-2.55 
1.09 
-.74 
-.24 

LLCI=lower limit confidence interval; LLCI= upper limit confidence interval 

 

66 

 

 

CHAPTER 5: DISCUSSION 

 

The current study aimed to better understand the mechanisms of cognitive improvement 

in exergaming by investigating the impacts of immersion and type of task load on cognitive 

improvement. Prior studies have suggested that exergaming with cognitively challenging tasks 

had a strong potential to improve people’s executive functions (Anderson-Hanley et al., 2017; 

Best, 2013; Monteiro-Junior et al., 2016). Furthermore, exergames require players to move 

physically and simulate virtual environments cognitively at the same time (Anderson-Hanley et 

al., 2017), which is consistent with the moving with thought argument for cognitive improvement 

(Diamond, 2015).  

Based on prior research, the current study argues that the process of feeling located in an 

immersive virtual environment can be treated as a cognitive or mental exercise of exergaming. 

Thus, the study designed an exergaming intervention to investigate the impacts of level of 

immersion and two types of cognitive load on cognitive improvement.   

There were three main questions that this study aimed to answer. First, does exergaming 

in a more immersive environment lead to more cognitive improvement than exergaming in a less 

immersive environment? Second, does being in a more immersive mediated environment 

influence participants’ cognitive improvement of exergaming? If so, does the feeling of spatial 

presence, a brain process elicited by level of immersion, have an impact on the relationship 

between level of immersion and cognitive improvement? Third, this study further investigated 

whether type of task load had an impact on spatial presence and cognitive improvement in the 

context of exergaming. 

 

67 

This chapter will first provide an overview of the findings and results based on different 

time frames. Following this overview is an interpretation of these results based on previous 

research. Limitations and future research directions are included at the end. 

5.1. Overview of the Findings 

Table 19 includes the results of the current study’s hypotheses. The first hypothesis posits 

that level of immersion will have an impact on participants’ cognitive performance. The results, 

including those of all three posttests, suggested that level of immersion positively influenced 

participants’ cognitive improvement in inhibitory control (Stroop) and cognitive flexibility (trail-

making), despite having no impact on working memory (backward digit span). Therefore, H1a 

and H1b were supported.  

The second hypothesis proposes a positive relationship between level of immersion and 

participants’ spatial presence experience. The results revealed a between-subjects effect of level 

of immersion on spatial presence across all three posttests, which suggested that exergaming in 

VR positively influences participants’ feelings of presence. Therefore, the second hypothesis was 

also supported.  

The third hypothesis posits that spatial presence will mediate the effect of immersion on 

participants’ cognitive improvement when level of immersion has an impact on their cognitive 

performance. In the first posttest, no main or mediation effect was found. In the second posttest, 

the results of mediation analyses revealed that spatial presence conditionally mediated the effects 

of immersion on participants’ inhibitory control (Stroop). In the last posttest, the impact of level 

of immersion on inhibitory control (Stroop) and cognitive flexibility (trail making) were fully 

mediated by spatial presence. However, no mediated effect of immersion on participants’ 

working memory (backward digit span) was found. Therefore, H3a and H3b were supported. 

 

68 

The fourth hypothesis posits that type of task load has an impact on participants’ spatial 

presence experiences. However, the results suggested that there was no difference between task-

relevant and task-irrelevant load in users’ spatial presence experiences in all three posttests. 

Thus, the fourth hypothesis was not supported. 

 

Table 19: Results of The Hypotheses of The Current Study 
 
H1 

Hypotheses 
In the context of an exergaming intervention, individuals in 
the high immersion condition will have a better cognitive 
improvement as indicated in a) inhibitory control, b) 
cognitive flexibility, and c) working memory than those in 
the low immersion condition. 
In the context of an exergaming intervention, individuals in 
the high immersion condition will have a stronger sense of 
spatial presence than those in the low immersion condition. 
H3: When immersion has a main effect on cognitive 
functions, spatial presence will mediate the impacts of 
immersion on older adults’ cognitive improvement in a) 
inhibitory control, b) cognitive flexibility, and c) working 
memory. 
In the context of an exergaming intervention, individuals in 
the high task-relevant load condition will report a stronger 
sense of spatial presence compared to people in the high 
task-irrelevant condition.   

Results 
H1a: supported 
H1b: supported 
H1c: rejected 

supported 

H3a: supported 
H3b: supported 
H3c: rejected 

H4: rejected 

H2 

H3 

H4 

 

The study’s research question investigated the impacts of task-relevant and task-

irrelevant load on cognitive improvement. The findings suggest that type of task load did not 

have a main impact on participants’ cognitive improvement over the course of the four-week 

cognitive training. However, type of task load and level of immersion did have an interaction 

effect on all three cognitive measures, either at a significance or approaching significance. To be 

more specific, participants in the VR + task-relevant load showed more improved performance in 

inhibitory control (Stroop) and cognitive flexibility (trail-making) compared to the other groups 

across three posttests. Interestingly, participants in the non-VR + task-irrelevant load condition 

 

69 

improved the most in working memory (backward digit span) among the four groups across three 

posttests. Therefore, type of task load had an interaction effect with level of immersion on 

cognitive improvement despite having no main effect on all dependent measures. 

5.2. Summary of Findings on Different Time Frames 

Table 20 shows the effects of manipulations at different time frames. In the first posttest, 

the results suggest that participants in the task-irrelevant load condition had more improved 

performance in cognitive flexibility (trail-making) immediately after a single bout of exergaming 

training. For those in the VR condition, task-irrelevant load also contributed to participants’ 

cognitive improvement in working memory (backward digit span).  

In the second posttest, the task-relevant load started to have an interaction effect with 

immersion on inhibitory control (Stroop) while the impacts of task-irrelevant load on working 

memory (backward digit span) and cognitive flexibility (trail making) still persisted after two 

weeks of exergaming training.  

 

Table 20: The Effects of Manipulation at Different Time Frames 
 

After the Single 

After the Two-
Week Training 

After the Four-
Week Training 

Inhibitory Control 

(Stroop) 

Bout 
None 

Cognitive Flexibility 

TI 

(Trail-Making) 

Working Memory 

VR+TI 

(Backward Digit Span) 

VR+TR 

 

TI 

 
Non-VR+TI 

VR 

VR+TR 

VR 

Across Three 

Posttests 

VR 

VR+TR 

VR 

VR+TR 

None 

Non-VR+TI 

 

VR 

 

 

 

 

VR+TR 

VR+TI 

Non-VR+TI 

 

TI 

Virtual Reality 

Virtual Reality 

Virtual Reality 

Non-Virtual 

Task-

+ 

+ 

Reality + 

Irrelevant 

Task-Relevant 

Task-Irrelevant 

Task-Irrelevant 

 

70 

In the last posttest, no main and interaction effects of the task-irrelevant load were found 

while level of immersion positively influenced participants’ inhibitory control (Stroop) and 

cognitive flexibility (trail-making). An interaction effect for Stroop also suggested that 

participants in the VR + task-relevant load condition improved the most in inhibitory control 

among the four groups.  

5.3. The Role of Level of Immersion on Cognitive Improvement 

The main purpose of this research was to investigate the impact of immersion on 

cognitive improvement. The results listed above reveal that immersion was a predictor of 

participants’ cognitive improvement in inhibitory control and cognitive flexibility. Interestingly, 

results across different time frames indicate that the main effect of level of immersion did not 

show up immediately after a single bout of exergaming. To be more specific, the main impact of 

level of immersion on cognitive improvement did not emerge at the early stage of the 

exergaming training, despite having an overall significant effect across four weeks. A possible 

explanation is that the underlying mechanisms responsible for cognitive improvement of short-

term versus long-term exergame training may be different (Barcelos et al., 2015) and the brain 

processes associated with long-term cognitive benefits may be triggered by immersion.  

The results, based on different time frames, further confirm the speculation that different 

mechanisms are responsible for cognitive improvement of short-term versus long-term exergame 

training. The task-irrelevant load was the main contributor of cognitive improvement in 

cognitive flexibility right after a single bout of exergame training, which suggested that task-

switching is the main mechanism responsible for cognitive improvement of short-term training. 

The impacts of the task-irrelevant load on cognitive flexibility was weaker, while immersion and 

the task-relevant load had an interaction effect on inhibitory control after four sessions of 

 

71 

exergame training. After eight training sessions, immersion became the main contributor of 

cognitive improvement while the task-irrelevant load no longer had any impact on inhibitory 

control and cognitive flexibility.  

Multiple studies found an immediate improvement on trail-making results after a single 

bout of exergame training and an improvement effect on both trail-making and Stroop results 

after three months (e.g., Anderson-Hanley et al., 2012; Maillot et al., 2012). Some studies (e.g., 

Barcelos et al., 2015) also revealed that inhibitory control and cognitive flexibility are associated 

with brain processes other than task-switching in a long-term intervention (i.e. three months). 

The results of the current study found that the brain processes elicited by level of immersion are 

key factors of cognitive improvement of long-term exergame training. Thus, people in the VR 

condition had a more improved performance in cognitive flexibility and inhibitory control in 

long-term training, but not in the short-term. 

Regarding working memory, immersion was not a predictor of participants’ cognitive 

improvement in backward digit span in all three posttests. Contrary to the first research 

hypothesis (H1c), participants in the non-VR + task-irrelevant load condition improved their 

working memory the most over the course of the intervention. The reason may be that the game-

irrelevant task was memory-based training, which improved working memory as a result of “a 

transfer of training effects” (Barcelos et al., 2015, p. 774). Furthermore, in the non-VR condition, 

participants could entirely focus on the task-irrelevant memory training task. Therefore, they had 

the most improved performance in working memory (backward digit span). 

5.4. Spatial Presence as a Mediator between Immersion and Cognitive Improvement  

The other question this study aimed to answer was, does being in a more immersive 

mediated environment itself influence participants’ cognitive improvement of exergaming? If so, 

 

72 

does the feeling of spatial presence, a brain process elicited by level of immersion (Monteiro-

Junior et al., 2016), have an impact on the relationship between level of immersion and cognitive 

improvement? Research on spatial presence experiences (e.g., Hofer et al., 2012; Schubert, 2009; 

Wirth et al., 2007) suggest that a higher level of immersion will elicit a feeling of presence, 

which is a psychological and cognitive process that requires media users to mentally engage in 

the environment. Consistent with previous literature, the findings revealed that participants in the 

VR condition reported a higher level of presence.  

For the mediation relationship, the results revealed that the impacts of level of immersion 

on trail-making and Stroop were fully mediated by spatial presence. To be more specific, spatial 

presence may be the reason why participants in the VR condition improved their cognitive 

flexibility and inhibitory control after finishing the four-week exergame training. In other words, 

when players felt located within the immersive mediated environment (i.e., VR) and perceived 

the possibility of moving in that environment, they were more likely to have cognitive 

improvement after experiencing presence. The underlying mechanism for the cognitive 

improvement may be that the process of feeling presence served as a mental exercise of 

exergaming and activated the brain areas associated with executive functions (Monteiro-Junior et 

al., 2016). 

These findings also reveal that the mediating effect of spatial presence did not happen 

immediately after a single bout of exergaming, which suggests that it takes time for spatial 

presence to make an impact on participants’ cognitive improvement.  An unexpected and non-

significant finding was that spatial presence was not a mediator between immersion and 

backward digit span, which indicates that the feeling of presence as a mental exercise did not 

contribute to the improvement of working memory. It may be that spatial attention only 

 

73 

contributed to spatial-related ability, such as spatial orientation and spatial working memory  

(Fabius, Mathôt, Schut, Nijboer, & Van der Stigchel, 2017), but not digit-related working 

memory.   

5.5. Task-relevant vs. Task-irrelevant Load on Cognitive Improvement  

Previous research has examined the impacts of cognitive demands and challenges in the 

context of exergames, with an emphasis on the comparison between high versus low cognitive 

demands (e.g., Anderson-Hanley et al., 2012; Maillot et al., 2012). However, the definitions and 

manipulations of cognitive demands were different across studies, and type of cognitive load 

were not discussed. This study further investigated the impacts of task-relevant versus task-

irrelevant load on spatial presence and cognitive improvement in the context of exergaming.  

This study’s hypothesis that the task-relevant load would lead to a higher level of 

presence than the task-irrelevant load, however, was not supported by the data. It may be that 

participants did not increase, or decrease, their attention allocation to the mediated environment 

based on the type of task load. Another speculation is that merely being in the VR condition 

provided participants with enough information to build a spatial situational model. Even though 

type of task load affected the participants’ attention allocation to and cognitive improvement in 

the mediated environment, they could still create a strong spatial situational model for the feeling 

of spatial presence. 

Furthermore, cognitively engaging tasks have been found to contribute to cognitive 

improvement in the context of exergaming (Barcelos et al., 2015; Best, 2013), but the impacts 

were not consistent due to different definitions of cognitively engaging tasks; some studies 

focused on task-relevant (gaming-relevant) mental effort  (e.g., Anderson-Hanley et al., 2017; 

Best, 2012) while others focused on task-irrelevant (non-gaming relevant) tasks (e.g., 

 

74 

Radovanović et al., 2014). However, limited research compares the impacts of these two types of 

mental exercise on cognitive improvement within the context of exergaming.   

This research explored whether the type of cognitive load influence executive functions 

differently during exergaming. Two competing theories were introduced to answer the research 

question of this study: perceptual load theory and the process model of spatial presence 

experiences. First, perceptual load theory (Lavie, 1995, 2005) predicts that people playing the 

exergame in a high task-irrelevant load condition will have to allocate their attentional and 

cognitive resources to both types of tasks, which requires the use of their cognitive functions. In 

this sense, dealing with task-unrelated distractors in the low task-relevant load condition will 

improve cognitive performance due to the practice of splitting attention. However, the process 

model of the formation of spatial presence experiences (Wirth et al., 2007) predicts that people 

cannot devote their full attentional resources to the mediated environment, so they are less likely 

to invest their cognitive resources and engage in the formation of spatial presence experiences. In 

other words, dealing with task-unrelated distractors in the low task-relevant load condition will 

not improve cognitive performance due to the lack of spatial presence.  

This study’s results suggest that both types of task load did not have a main effect over 

the course of four weeks. Instead, an interaction between level of immersion and type of task 

load was found across the four-week training. This is consistent with previous studies on the 

impacts of level of cognitive demands and load (e.g., Best, 2012), which found that cognitive 

demands and load moderated the effects of physical activity of exergaming on cognitive 

improvement in older adults despite having no main effect. In the current study, type of task load 

made the effect of level of immersion on inhibitory control and cognitive flexibility more salient 

during the intervention. One speculation is that playing a more difficult game (task-relevant 

 

75 

load), which required more spatial attention and orientation, moderated the impact of immersion 

on cognitive improvement across the four-week training. 

The findings, based on different time frames, reveals that the task-irrelevant load had an 

impact on cognitive flexibility right after a single bout of exergaming, but the effect started 

fading after the second week of training. No effect of task-relevance was found on cognitive 

flexibility and inhibitory control in the last posttest. One speculation is that participants in the 

task-irrelevant load condition might have developed a strategy to deal with the game-irrelevant 

task after two weeks, so they did not require the same amount of executive functions to process 

the two tasks separately as they needed in the first posttest, which was less likely to prime the 

use of their executive functions for the cognitive measures later.   

The results, based on different time frames, also suggest that the task-relevant load had an 

interaction effect with immersion on inhibitory control across times in the second and last 

posttests. One possible explanation is that dealing with a game-relevant task in the immersive 

environment required participants’ spatial attention and orientation to target certain fruits and 

avoid hitting bombs, which also amplified the impacts of immersion on participant’s inhibitory 

control. No interaction effects between level of immersion and task-relevant load were found in 

the first posttest. A possible reason may be that the mechanisms for cognitive improvement of 

short-term versus long-term training were different, and VR + game-relevant load were more 

likely to contribute to the improvement of inhibitory control in the longer-term training (i.e., two 

weeks and four weeks). One speculation for the time effect of immersion is that it requires more 

time or training sessions for people to be familiar with the virtual environment and allocate their 

spatial attention to the environment, especially for people who did not have many gaming 

experiences. Future research could include neuroscientific evidence (e.g., functional magnetic 

 

76 

resonance imaging and electroencephalography) to investigate whether there is a difference in 

brain functional connectivity or other brain activities associated with executive functions 

between different time frames. 

 

The results, based on different time frames, also show complicated and unexpected 

interaction effects of task-irrelevant load and level of immersion. In the first posttest, an 

interaction effect was found: the task-irrelevant load had a stronger impact on participants’ 

performance in the VR condition than on those in the non-VR condition. One speculation is that 

people dealt with the task-relevant and task-irrelevant loads differently. People in the VR 

condition had to process two incompatible loads at the same time, which is different from people 

in the VR+ task-relevant load condition. In addition, the manipulation of the task-irrelevant load 

was a memory-based task, which served as a practice of the working memory task, and VR 

served as a distraction for the participants to use their executive functions to keep those numbers 

in mind. Therefore, they had a better performance in working memory right after the first 

exergame training. 

 

The interaction effects of the task-irrelevant load and level of immersion on participants’ 

working memory showed a different pattern in the second posttest. Participants in the non-VR 

condition with the task-irrelevant load improved their working memory the most among the four 

conditions. An explanation for this interaction effect is that people in the VR condition have 

already developed a strategy to process these two loads with minimum mental effort and people 

in the non-VR condition can pay more attention to perform the assigned memory tasks. 

Therefore, people in the non-VR+ task-irrelevant load condition, instead of those in the VR+ 

task-irrelevant load condition, had a better performance in working memory among the four 

groups. Moreover, no effect was found on the last working memory posttest, which may have 

 

77 

resulted from the participants’ familiarity with the memory tasks, although they still reported a 

similar level of mental effort as in the previous two posttests. 

5.6. Summary of Discussion 

Overall, level of immersion had an impact on inhibitory control (Stroop) and cognitive 

flexibility (trail-making). Furthermore, an interaction effect between level of immersion and type 

of task load was found on inhibitory control, with the results showing that VR and task-relevant 

load led to a better performance in Stroop. When immersion had a positive impact on 

participants’ cognitive performance, spatial presence mediated the relationship. These findings 

suggested that spatial presence, elicited by the immersive virtual reality, was involved in a 

cognitive process, which later led to an improved performance in visual processing speed and 

inhibitory control. Regarding working memory, using a game-unrelated memory task would 

directly contribute to participants’ working memory.  

The results, based on different time frames, further suggested that the underlying 

mechanisms responsible for cognitive improvement are different in short-term and long-term 

effects. The task-irrelevant load had a training effect. The immediate effect of the task-irrelevant 

load on working memory may suggest “a transfer of training effects” (Barcelos et al., 2015, p. 

774) after dividing their attention between two tasks when playing the exergame in the low-

difficulty mode and memorizing a seven to nine-digit number at the same time. After four 

training sessions, immersion and task-relevant load condition started to interact on participants’ 

inhibitory control. In the last posttest, the impacts of task-irrelevant load no longer existed on 

cognitive flexibility, immersion had a main effect on cognitive flexibility, and had both main and 

interaction effects on inhibitory control. Therefore, immersion and task-relevant load were 

 

78 

responsible for the long-term effect while task-irrelevant load was responsible for the short-term 

effect. 

5.7. Limitations 

Like previous studies of exergaming for cognitive improvement, this longitudinal study is 

not without limitations. First of all, the majority of the participants were white (90.2%) and most 

were females (73.2%). There was only one male participant in one condition (non-VR + task-

irrelevant condition).  Also, the results of this study may not be able to apply to all older adults 

since the average of the participants was 61 years old, which was considered as young older 

adults. Thus, the study’s findings may not be able to be generalized to all populations. Future 

research could include more participants and make the sample more similar to the demographic 

or specifically targeting at people considered as older adults (i.e. aged 65 or older). 

Second, small sample sizes and missing responses may decrease statistical power. 

However, this limitation was common in the previous studies due to limited resources. 

Furthermore, the adherent rate of the current study (96.9%) was higher than similar studies 

(Barcelos et al., 2015). Therefore, this limitation was not considered a threat to internal validity 

or statistical power. Future studies should consider implementing a large-scale longitudinal 

exergaming intervention given that the researchers have enough resources.  

Third, similar to previous research on the impacts of exergaming on cognitive 

improvement, this research found an immediate effect after a single-bout of exergaming and also 

a long-term effect after the completion of four-week training. However, the current study could 

not predict how long the impacts would last. Previous studies on cognitive and physical 

interventions showed that the improvements after interventions (e.g., ten training sessions) 

continued for two years, and the effects of some training programs continued up to seven years. 

 

79 

The results can provide insight into the potential impact of the current study. Future studies 

should further investigate the long-term impact of interventions on executive functions. 

Fourth, participants knew how well they performed when they did the backward digit 

span measure. In other words, they knew it was a memory task and could figure out how many 

digits they successfully memorized and how well they performed. If the participants did not 

perform very well at the beginning, they might have stopped trying to finish the task. Future 

studies should incorporate other working memory measures in which participants will be 

unaware of their performance. In addition, the relationship between exergaming and working 

memory might be different from exergaming and other executive functions (e.g., inhibitory 

control and selective attention). It is recommended that future research should further investigate 

what characteristics of exergaming would contribute to the improvement of working memory, 

including other types, such as spatial working memory.  

In addition, previous studies showed that the impacts of exergaming on working memory, 

inhibitory control, and cognitive flexibility were different and suggested that the relationship 

between exergaming and working memory might be different from exergaming and inhibitory 

control or exergaming and selective attention (Barcelos et al., 2015). It is recommended that 

future research should further investigate what characteristics of exergaming would contribute to 

the improvement of working memory, including other types such as spatial working memory.  

Fifth, it is hard to tell whether people allocated their attention to the memory tasks first 

and then to the game or the other way around. This study asked participants to try their best to 

perform both tasks well. Still, two participants did not focus on the game-irrelevant task and two 

participants failed to try their best to play the game in different time frames. As a result, these 

 

80 

participants were excluded from data analysis. Future research could design better game-

irrelevant tasks that participants would be more engaged in doing.  

Sixth, the study used a self-reported screening survey to recruit cognitively and 

physically healthy participants. Future study could include neuroscientific measures to better 

assess participants’ cognitive functions before the training.       

Seventh, the results suggested that the impacts of immersion did not happen until 

participants finished their fourth training (two weeks). However, it is not clear whether long-term 

effect of immersion on cognitive improvement was a result of a time effect (i.e., requiring two 

weeks for the effect to show up) or dosage effect (i.e., requires four training sessions for the 

effect to show up). Future studies could separate these factors (e.g., four sessions in one day vs. 

four sessions in two weeks) to evaluate the cognitive improvement. 

 

 

 

81 

CHAPTER 6: CONCLUSION 

 

This exploratory study of the cognitive benefits of exergame training for older adults 

particularly focused on the roles of level of immersion and type of task load. Overall, the 

findings suggest that the combination of virtual reality technology and exergames contributed to 

cognitive improvement in older adults in the four-week exergame training. Moreover, the feeling 

of presence mediated the impact of VR on cognitive improvement in the context of exergaming. 

The results, based on different time frames, further confirmed that the mechanisms responsible 

for older adults’ cognitive improvement are different between short-term and long-term training.  

This study has both theoretical and practical implications. Theoretically, this study adds 

to the literature that investigates the characteristics of exergames and improvement of executive 

functions. The results of this study support the concept of cognitive reserve (Stern, 2012), which 

argues that cognitive training interventions may be able to slow down the level of cognitive 

decline and maintain cognitive functions as much as possible. Specifically, brain reserve, which 

emphasizes hardware characteristics such as brain volume and neuronal structural integrity, can 

be strengthened by the physical exercise of exergaming, and cognitive reserve, which highlights 

the software aspects such as cognitive functioning and plasticity of neural circuits, can be 

boosted by the mental exercise (i.e., feeling of presence). The current research suggests that 

being in the VR condition required participants’ cognitive resources to locate themselves in the 

virtual environment and perceive their actionable possibilities, which increases the amount of 

mental exercise during exergaming. 

Previous studies suggested assessing potential mediators or moderators (Barcelos, et al., 

2015) on the relationship between exergaming and neurobiological/neuropsychological benefits. 

 

82 

The current study applied the theoretical framework of spatial presence experiences to exergame 

scholarship and revealed that immersion had a positive impact on executive functions after the 

four-week training and spatial presence was a mediator between immersion and cognitive 

improvement. 

Furthermore, the current research found that both perceptual load theory and the 

formation model of spatial presence experiences can be used to explain the results. Perceptual 

load theory can be used to explain the mechanism responsible for cognitive improvement in 

exergaming with a task-irrelevant load in short-term training, and the formation model of spatial 

presence can be used to explain the impacts of a task-relevant load and spatial presence on 

cognitive improvement in long-term training.  

Regarding practical implications, the preliminary results of this four-week exergame 

training suggests that immersive technologies for health would be an effective tool to increase 

the intensity of physical and mental exercises (Blondell et al., 2014), which could be applied to 

both daily life and clinical settings. For older adults, playing virtual reality-based exergames two 

times per week for one month showed greater cognitive improvement compared to a similar dose 

of non-VR exergames. 

This study shows the possibilities of applying the latest media technologies to 

exergaming, which amplifies the advantages of exergames — having the benefits of physical 

activity and attractiveness of video games. VR technologies have the potential to facilitate 

enjoyment and promote repeated use. Furthermore, the results suggested that the process of 

feeling presence served as a mental exercise within exergaming and activated brain areas 

associated with the use of executive functions. In this sense, VR-enhanced exergames, which 

combine physical and mental exercise, can serve as a good alternative for traditional exercise 

 

83 

without the limitation of time and space. As VR technologies become more popular and 

affordable, individuals who want to ameliorate cognitive decline or avoid cognitive deficits may 

benefit from the use of exergaming in VR environments. This research also provided some 

suggestions for the design and development of exergames. The findings of the current research 

have shown that VR exergaming is beneficial for older adults’ executive functions, and that the 

requirement of spatial attention and orientation may be the reason why older adults benefit from 

exergaming. In the future, the design of exergames could emphasize game features that require 

players’ spatial attention and orientation, such as avoiding or locating certain objects, to provide 

novel strategies for preventing cognitive decline in older adults. 

 

 

 

84 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

APPENDIX 

 

85 

RESEARCH PROTOCOL  

Overview 

• 

• 

• 

• 

• 

• 

• 

• 

• 

Recruit through community SONA (a screening survey asking if they can play games 

without glasses or with contact lenses) 

Room 511 

Set up equipment 

Pre-experiment online survey through SONA 

Pre-game Cognitive Tasks 

Play game for 20 minutes 

Post-game Questionnaire 

Schedule the appointment time for the next session (up to 8 sessions) 

Total time in lab: 30-35 mins 

Equipment and software 

• 

• 

• 

• 

• 

• 

• 

• 

Computer 

Monitor/ TV screen 

Oculus Rift 

Xbox Kinect Motion Sensor 

Consent form (in the pre-test survey) 

Online questionnaires 

Experimenter sheet 

Dual-task response software (Inquisit 5.0) 

Pre-Experiment 

• 

• 

• 

• 

Set up hardware 

Make sure the Oculus lenses are clean 

o  Use the air bulb to remove dust/particles 

o  Use the microfiber cloth to gently remove any grease/smudges from the lenses 

Open the software for pre-game cognitive tasks as well as a webpage with the post-

game questionnaire 

Researcher answer the first question in the questionnaire (which condition the 

participant is in). Double check the schedule if you’ve chosen the right ones! 

 

86 

• 

Load the game and double check the schedule and the conditions again  

o  Oculus/Zen Mode + memorizing a 10-digit number 

o  Oculus/ Survival Mode 

o  Screen/ Zen Mode+ memorizing a 10-digit number 

o  Screen/ Survival Mode 

Enter participant’s number and condition. 

Check settings of oculus rift. If participants are in the Oculus Condition, check if the 

light of Oculus Rift is on (blue). If participants are not in the Oculus Condition, make 

• 

• 

sure the Oculus Rift is off 

Participant Arrives 

•  Explain Study:  

“This study is about the relationship between playing video games and health 

improvement. This is a 4-week study and consists of 8 sessions. This is the first session 

and we highly encourage you to finish 8 sessions. To thank you for your continued 

participation you will be paid each consecutive week you complete.” 

““Before the study starts, we will ask you to sign the consent form. If you consent to 

participate our study, you will be playing a video game in a virtual environment. You will 

be asked to do 3 tasks before and after you play a video game for 20 minutes. This video 

game will require your body movement to move to control the games. Please let me know 

if you have any questions or concerns.” 

•  Sign the online Consent Form: 

“Please read this consent form thoroughly so that you understand it. Please let me know if 

you have any questions or concerns.” 

 

Details for conditions 

Task-relevant load | 2 conditions: (High vs. Low immersion) 

Double check they are in the survival mode 

“Today you will be exploring the games using the [Oculus Rift/ TV] and the Oculus Controller. 

You will only need to use this equipment to play the video game. 

 

87 

Then adjust the sizing of the straps of the Oculus rift or the distance between the participant and 

the TV screen. 

 “Let me know if you have any issue on controlling or seeing objects in the games.” 

“You will have a minute to practice controlling the game. After that, you will be asked to play a 

game. If you feel uncomfortable or motion sickness, you can stop at any point during your game 

play.” 

After 20 minutes, Stop the experiments and assist in removing equipment and or help the 

participant to get back to the seat. 

"Ok, that’s it, we are going to stop here [and I’m going to take off the headphones and Oculus 

now].” 

“Alright, let’s do the cognitive tasks again and also fill out another survey.” 

Load the software for cognitive tasks and the Qualtrics questionnaire on computer monitor.  

“Thank you for coming in today. This is the 15$ and we thank you for coming to this study. We 

also encourage you to keep participating this study and sign up for the next session now.” 

If they agree to stay in the intervention, then schedule the next session. 

After the participant has left, open the window to let in some fresh air. Close it before the next 

participant arrives. 

--------------------------------------------------------------------------------------------------------------------- 

Task-irrelevant load | 2 conditions: (High vs. Low immersion) 

Double check they are in the zen mode 

“Today you will be exploring the games using the [Oculus Rift/ TV] and the Oculus Controller.” 

You will only need to use this equipment to play the video game. 

Then adjust the sizing of the straps of the Oculus rift or the distance between the participant and 

the TV screen. 

 “Let me know if you have any issue on controlling or seeing objects in the games.” 

 

88 

“You will have a minute to practice controlling the game. After that, you will be asked to play a 

game. If you feel uncomfortable or motion sickness, you can stop at any point during your game 

play.” 

Before starting the game, use an online randomizer to generate a 10-digit number and give 

the participant 1 minute to memorize.  

“You should try your best to memorize this 10-digit number when you play the game. We will 

ask you to report this number after the 20-minute gameplay.”  

After 20 minutes, Stop the experiments and assist in removing equipment and or help the 

participant to get back to the seat. 

"Ok, that’s it, we are going to stop here [and I’m going to take off the headphones and Oculus 

now].” 

“Alright, let’s do the cognitive tasks again and also fill out another survey.” 

Load the software for cognitive tasks and the Qualtrics questionnaire on computer monitor.  

“Thank you for coming in today. This is the 15$ and we thank you for coming to this study. We 

also encourage you to keep participating this study and sign up for the next session now” 

If they agree to stay in the intervention, then schedule the next session. 

After the participant has left, open the window to let in some fresh air. Close it before the next 

participant arrives. 

 

 

 

 

 

 

 

 

89 

Instruction of Game Play- Zen Mode 

Fruit Ninja is very intuitive to play.  

Kinect: Your arms are swords. VR: Your holding swords. 

Use your swords to slice fruit as it flies across the screen 

All fruit, no bombs. A great way to relax or hone your ninja skills. 

 

 

 

 

 

 

 

 

 

 

90 

Instruction of Game Play- Survival Mode 

Fruit Ninja is very intuitive to play.  

Kinect: Your arms are swords. VR: You are holding swords.  

Use your body to slice fruit as it flies across the screen 

All fruit, no bombs. A great way to relax or hone your ninja skills. 

Face off against flying cannons! Slice fruit, avoid bombs 

 

 

 

 

 

 

91 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCES 

 

92 

REFERENCES 

 

Albert, M. S., & Moss, M. B. (1988). Geriatric neuropsychology. 
 
Anderson-Hanley, C., Arciero, P. J., Brickman, A. M., Nimon, J. P., Okuma, N., Westen, S. 
C., . . . Zimmerman, E. A. (2012). Exergaming and Older Adult Cognition A Cluster 
Randomized Clinical Trial. American Journal of Preventive Medicine, 42(2), 109-119. 
doi:10.1016/j.amepre.2011.10.016 

 
Anderson-Hanley, C., Maloney, M., Barcelos, N., Striegnitz, K., & Kramer, A. (2017). 

Neuropsychological Benefits of Neuro-Exergaming for Older Adults: A Pilot Study of an 
Interactive Physical and Cognitive Exercise System (iPACES). Journal of Aging and 
Physical Activity, 25(1), 73-83. doi:10.1123/japa.2015-0261 

 
Angevaren, M., Aufdemkampe, G., Verhaar, H. J. J., Aleman, A., & Vanhees, L. (2008). 

Physical activity and enhanced fitness to improve cognitive function in older people 
without known cognitive impairment. Cochrane Database of Systematic Reviews(3). 
doi:10.1002/14651858.CD005381.pub3 

 
Ball, K., Berch, D. B., Helmers, K. F., Jobe, J. B., Leveck, M. D., Marsiske, M., . . . Tennstedt, 
S. L. (2002). Effects of cognitive training interventions with older adults: a randomized 
controlled trial. Jama, 288(18), 2271-2281.  

 
Ball, K., Edwards, J. D., & Ross, L. A. (2007). The impact of speed of processing training on 

cognitive and everyday functions. The Journals of Gerontology Series B: Psychological 
Sciences and Social Sciences, 62(Special_Issue_1), 19-31.  

 
Barcelos, N., Shah, N., Cohen, K., Hogan, M. J., Mulkerrin, E., Arciero, P. J., . . . Anderson-

Hanley, C. (2015). Aerobic and Cognitive Exercise (ACE) Pilot Study for Older Adults: 
Executive Function Improves with Cognitive Challenge While Exergaming. Journal of 
the International Neuropsychological Society, 21(10), 768-779. 
doi:10.1017/s1355617715001083 

 
Best, J. R. (2012). Exergaming Immediately Enhances Children's Executive Function. 

Developmental Psychology, 48(5), 1501-1510. doi:10.1037/a0026648 

 
Best, J. R. (2013). Exergaming in Youth Effects on Physical and Cognitive Health. Zeitschrift 

Fur Psychologie-Journal of Psychology, 221(2), 72-78. doi:10.1027/2151-2604/a000137 

 
Blondell, S. J., Hammersley-Mather, R., & Veerman, J. L. (2014). Does physical activity prevent 

cognitive decline and dementia?: A systematic review and meta-analysis of longitudinal 
studies. Bmc Public Health, 14. doi:10.1186/1471-2458-14-510 

 

 

93 

Brehmer, Y., Westerberg, H., & Bäckman, L. (2012). Working-memory training in younger and 

older adults: training gains, transfer, and maintenance. Frontiers in Human Neuroscience, 
6, 63. doi:10.3389/fnhum.2012.00063 

Calamia, M., Markon, K., & Tranel, D. (2013). The Robust Reliability of Neuropsychological 

Measures: Meta-Analyses of Test–Retest Correlations. The Clinical Neuropsychologist, 
27(7), 1077-1105. doi:10.1080/13854046.2013.809795 

 
Centers for Disease Control and Prevention. (2013). The state of aging and health in America 

2013 Atlanta, GA: Centers for Disease Control and Prevention, US Department of 
Health and Human Services. Centers for Disease Control and Prevention: Centers for 
Disease Control and Prevention. 

 
Centers for Disease Control and Prevention. (2014). Behavioral risk factor surveillance system 

survey data. Atlanta, GA: US Department of Health and Human Services, Centers for 
Disease Control and Prevention. 

 
Chao, L., & Knight, R. (1997). Prefrontal deficits in attention and inhibitory control with aging. 

Cerebral cortex (New York, NY: 1991), 7(1), 63-69.  

 
Cheng, S.-T. (2016). Cognitive Reserve and the Prevention of Dementia: the Role of Physical 
and Cognitive Activities. Current Psychiatry Reports, 18(9), 85. doi:10.1007/s11920-
016-0721-2 

 
Chu, C.-H., Chen, A.-G., Hung, T.-M., Wang, C.-C., & Chang, Y.-K. (2015). Exercise and 

fitness modulate cognitive function in older adults. Psychology and Aging, 30(4), 842.  

 
Chuang, L.-Y., Hung, H.-Y., Huang, C.-J., Chang, Y.-K., & Hung, T.-M. (2015). A 3-month 

intervention of Dance Dance Revolution improves interference control in elderly females: 
a preliminary investigation. Experimental Brain Research, 233(4), 1181-1188. 
doi:10.1007/s00221-015-4196-x 

 
Crescentini, C., Urgesi, C., Fabbro, F., & Eleopra, R. (2014). Cognitive and brain reserve for 

mind-body therapeutic approaches in multiple sclerosis: a review. Restorative neurology 
and neuroscience, 32(5), 575-595.  

 
Cummings, J. J., & Bailenson, J. N. (2016). How immersive is enough? A meta-analysis of the 

effect of immersive technology on user presence. Media Psychology, 19(2), 272-309.  

 
Cunningham, J. B., & McCrum-Gardner, E. (2007). Power, effect and sample size using 

GPower: practical issues for researchers and members of research ethics committees. 
Evidence-Based Midwifery, 5(4), 132-137.  

 
Diamond, A. (2014). Want to Optimize Executive Functions and Academic Outcomes?: Simple, 
Just Nourish the Human Spirit. Minnesota Symposia on Child Psychology, 37, 205-232.  

 

 

94 

Diamond, A. (2015). Effects of Physical Exercise on Executive Functions: Going beyond Simply 

Moving to Moving with Thought. Annals of sports medicine and research, 2(1), 1011. 

  
Eggenberger, P., Wolf, M., Schumann, M., & de Bruin, E. D. (2016). Exergame and Balance 
Training Modulate Prefrontal Brain Activity during Walking and Enhance Executive 
Function in Older Adults. Frontiers in Aging Neuroscience, 8. 
doi:10.3389/fnagi.2016.00066 

 
Fabius, J., Mathôt, S., Schut, M., Nijboer, T., & Van der Stigchel, S. (2017). Focus of spatial 
attention during spatial working memory maintenance: Evidence from pupillary light 
response. Visual Cognition, 25(1-3), 10-20.  

 
Finch, C. E. (2009). The neurobiology of middle-age has arrived. Neurobiology of aging, 30(4), 

515-520.  

 
Gamito, P., Oliveira, J., Coelho, C., Morais, D., Lopes, P., Pacheco, J., . . . Barata, A. F. (2015). 

Cognitive training on stroke patients via virtual reality-based serious games. Disability 
and rehabilitation, 1-4.  

 
Gamito, P., Oliveira, J., Coelho, C., Morais, D., Lopes, P., Pacheco, J., . . . Barata, A. F. (2017). 

Cognitive training on stroke patients via virtual reality-based serious games. Disability 
and rehabilitation, 39(4), 385-388. doi:10.3109/09638288.2014.934925 

 
Gazzaley, A., Cooney, J. W., Rissman, J., & D'esposito, M. (2005). Top-down suppression 

deficit underlies working memory impairment in normal aging. Nature neuroscience, 
8(10), 1298-1300.  

 
Glisky, E. L. (2007). Changes in cognitive function in human aging. Brain aging: models, 

methods, and mechanisms, 3-20.  

 
Gupta, S. K. (2011). Intention-to-treat concept: A review. Perspectives in Clinical Research, 

2(3), 109-112. doi:10.4103/2229-3485.83221 

 
Harada, C. N., Love, M. C. N., & Triebel, K. L. (2013). Normal cognitive aging. Clinics in 

geriatric medicine, 29(4), 737-752.  

 
Havranek, M., Langer, N., Cheetham, M., & Jancke, L. (2012). Perspective and agency during 

video gaming influences spatial presence experience and brain activation patterns. 
Behavioral and Brain Functions, 8, 13. doi:10.1186/1744-9081-8-34 

 
Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A 

regression-based approach: Guilford Publications. 

 
Hillman, C. H., Erickson, K. I., & Kramer, A. F. (2008). Be smart, exercise your heart: exercise 

effects on brain and cognition. Nature reviews neuroscience, 9(1), 58-65.  

 

 

95 

Hofer, M., Sele, M., & Wirth, W. (2013). Effects of Stereoscopic View on Spatial Presence: The 

Mediating Role of Mental Models. Paper presented at the 8th Conference of the Media 
Psychology Division, Julius-Maximilians University Würzburg, Germany.  

Hofer, M., Wirth, W., Kuehne, R., Schramm, H., & Sacau, A. (2012). Structural Equation 

Modeling of Spatial Presence: The Influence of Cognitive Processes and Traits. Media 
Psychology, 15(4), 373-395. doi:10.1080/15213269.2012.723118 

 
Huang, T., Larsen, K., Ried‐Larsen, M., Møller, N., & Andersen, L. B. (2014). The effects of 
physical activity and exercise on brain‐derived neurotrophic factor in healthy humans: 
A review. Scandinavian journal of medicine & science in sports, 24(1), 1-10.  

 
Huntley, J. D., Gould, R. L., Liu, K., Smith, M., & Howard, R. J. (2015). Do cognitive 

interventions improve general cognition in dementia? A meta-analysis and meta-
regression. BMJ Open, 5(4). doi:10.1136/bmjopen-2014-005247 

 
Hurd, M. D., Martorell, P., Delavande, A., Mullen, K. J., & Langa, K. M. (2013). Monetary costs 
of dementia in the United States. New England Journal of Medicine, 368(14), 1326-1334.  

 
Jurado, M. B., & Rosselli, M. (2007). The elusive nature of executive functions: a review of our 

current understanding. Neuropsychology review, 17(3), 213-233.  

 
Kayama, H., Okamoto, K., Nishiguchi, S., Yamada, M., Kuroda, T., & Aoyama, T. (2014). 

Effect of a Kinect-Based Exercise Game on Improving Executive Cognitive Performance 
in Community-Dwelling Elderly: Case Control Study. J Med Internet Res, 16(2), e61. 
doi:10.2196/jmir.3108 

 
Khng, K. H., & Lee, K. (2014). The relationship between Stroop and stop-signal measures of 

inhibition in adolescents: influences from variations in context and measure estimation. 
PloS one, 9(7), e101356.  

 
Kirk, A., MacMillan, F., Rice, M., & Carmichael, A. (2013). An Exploratory Study Examining 

the Appropriateness and Potential Benefit of the Nintendo Wii as a Physical Activity 
Tool in Adults Aged >= 55 Years. Interacting with Computers, 25(1), 102-114. 
doi:10.1093/iwc/iws004 

 
Kober, S. E., Kurzmann, J., & Neuper, C. (2012). Cortical correlate of spatial presence in 2D and 
3D interactive virtual reality: An EEG study. International Journal of Psychophysiology, 
83(3), 365-374. doi:10.1016/j.ijpsycho.2011.12.003 

 
Kramer, A. F., Hahn, S., Cohen, N. J., Banich, M. T., McAuley, E., Harrison, C. R., . . . Boileau, 
R. A. (1999). Ageing, fitness and neurocognitive function. Nature, 400(6743), 418-419.  

 
Lam, M., Eng, G. K., Rapisarda, A., Subramaniam, M., Kraus, M., Keefe, R. S., & Collinson, S. 
L. (2013). Formulation of the age–education index: Measuring age and education effects 
in neuropsychological performance. Psychological assessment, 25(1), 61.  

 

 

96 

Lavie, N. (1995). Perceptual load as a necessary condition for selective attention. Journal of 

Experimental Psychology-Human Perception and Performance, 21(3), 451-468. 
doi:10.1037/0096-1523.21.3.451 

 
Lavie, N. (2005). Distracted and confused?: Selective attention under load. Trends in Cognitive 

Sciences, 9(2), 75-82. doi:10.1016/j.tics.2004.12.004 

 
Lavie, N., Hirst, A., de Fockert, J. W., & Viding, E. (2004). Load theory of selective attention 
and cognitive control. Journal of Experimental Psychology-General, 133(3), 339-354. 
doi:10.1037/0096-3445.133.3.339 

 
Lee, K. M. (2004). Why presence occurs: Evolutionary psychology, media equation, and 

presence. Presence: Teleoperators and virtual environments, 13(4), 494-505.  

 
Leppink, J., Paas, F., Van der Vleuten, C. P., Van Gog, T., & Van Merriënboer, J. J. (2013). 

Development of an instrument for measuring different types of cognitive load. Behavior 
research methods, 45(4), 1058-1072.  

 
Lo Priore, C., Castelnuovo, G., Liccione, D., & Liccione, D. (2003). Experience with V-STORE: 

considerations on presence in virtual environments for effective neuropsychological 
rehabilitation of executive functions. CyberPsychology & Behavior, 6(3), 281-287.  

 
Maguire, E. A., Burgess, N., Donnett, J. G., Frackowiak, R. S. J., Frith, C. D., & O'Keefe, J. 

(1998). Knowing where and getting there: A human navigation network. Science, 
280(5365), 921-924. doi:10.1126/science.280.5365.921 

 
Mansouri, F. A., Tanaka, K., & Buckley, M. J. (2009). Conflict-induced behavioural adjustment: 

a clue to the executive functions of the prefrontal cortex. Nature reviews. Neuroscience, 
10(2), 141-152.  

 
Masley, S., Roetzheim, R., & Gualtieri, T. (2009). Aerobic Exercise Enhances Cognitive 

Flexibility. Journal of Clinical Psychology in Medical Settings, 16(2), 186-193. 
doi:10.1007/s10880-009-9159-6 

 
McMahan, R. P., Bowman, D. A., Zielinski, D. J., & Brady, R. B. (2012). Evaluating Display 

Fidelity and Interaction Fidelity in a Virtual Reality Game. Ieee Transactions on 
Visualization and Computer Graphics, 18(4), 626-633.  

 
Meneghini, V., Barbosa, A. R., de Mello, A., Bonetti, A., & Guimaraes, A. V. (2016). Perception 

of the elderly regarding participation in exergaming-based exercise: a qualitative study. 
Ciencia & Saude Coletiva, 21(4), 1033-1041. doi:10.1590/1413-81232015214.11812015 

 
Monteiro-Junior, R. S., Otero Vaghetti, C. A., Nascimento, O. J. M., Laks, J., & Deslandes, A. 

C. (2016). Exergames: neuroplastic hypothesis about cognitive improvement and 
biological effects on physical function of institutionalized older persons. Neural 
Regeneration Research, 11(2), 201-204. doi:10.4103/1673-5374.177709 

 

97 

 
Nachar, N. (2008). The Mann-Whitney U: A test for assessing whether two independent samples 
come from the same distribution. Tutorials in Quantitative Methods for Psychology, 4(1), 
13-20.  

Ogawa, E. F., You, T. J., & Leveille, S. G. (2016). Potential Benefits of Exergaming for 

Cognition and Dual-Task Function in Older Adults: A Systematic Review. Journal of 
Aging and Physical Activity, 24(2), 332-336. doi:10.1123/japa.2014-0267 

 
Pachoulakis, I., Papadopoulos, N., & Ieee. (2016). Exergames for Parkinson's Disease Patients: 

the Balloon Goon game 2016 International Conference on Telecommunications and 
Multimedia (pp. 12-17). 

 
Peng, W. (2009). Design and evaluation of a computer game to promote a healthy diet for young 

adults. Health communication, 24(2), 115-127.  

 
Pett, M. A. (2015). Nonparametric statistics for health care research: Statistics for small 

samples and unusual distributions: Sage Publications. 

 
Radovanović, S., Jovičić, M., Marić, N. P., & Kostić, V. (2014). Gait characteristics in patients 
with major depression performing cognitive and motor tasks while walking. Psychiatry 
Research, 217(1–2), 39-46. doi:http://dx.doi.org/10.1016/j.psychres.2014.02.001 

 
Riva, G., Waterworth, J. A., & Waterworth, E. L. (2004). The layers of presence: a bio-cultural 

approach to understanding presence in natural and mediated environments. 
CyberPsychology & Behavior, 7(4), 402-416.  

 
Ronnlund, M., Nyberg, L., Backman, L., Nilsson, L., Rönnlund, M., & Bäckman, L. (2005). 

Stability, Growth, and Decline in Adult Life Span Development of Declarative Memory: 
Cross-Sectional and Longitudinal Data From a Population-Based Study. Psychology and 
Aging, 20(1), 3-18.  

 
Rovio, S., Spulber, G., Nieminen, L. J., Niskanen, E., Winblad, B., Tuomilehto, J., . . . Kivipelto, 

M. (2010). The effect of midlife physical activity on structural brain changes in the 
elderly. Neurobiology of aging, 31(11), 1927-1936.  

 
Salthouse, T. A. (2009). When does age-related cognitive decline begin? Neurobiology of aging, 

30(4), 507-514. doi:10.1016/j.neurobiolaging.2008.09.023 

 
Salthouse, T. A., Atkinson, T. M., & Berish, D. E. (2003). Executive functioning as a potential 

mediator of age-related cognitive decline in normal adults. Journal of Experimental 
Psychology: General, 132(4), 566.  

 
Sánchez-Cubillo, I., Periáñez, J. A., Adrover-Roig, D., Rodríguez-Sánchez, J. M., Ríos-Lago, 

M., Tirapu, J., & Barceló, F. (2009). Construct validity of the Trail Making Test: Role of 
task-switching, working memory, inhibition/interference control, and visuomotor 

 

98 

abilities. Journal of the International Neuropsychological Society, 15(3), 438-450. 
doi:10.1017/S1355617709090626 

Schubert, T. (2009). A New Conception of Spatial Presence: Once Again, with Feeling. 
Communication Theory, 19(2), 161-187. doi:10.1111/j.1468-2885.2009.01340.x 

 
Schubert, T., Friedmann, F., & Regenbrecht, H. (2001). The experience of presence: Factor 
analytic insights. Presence: Teleoperators and virtual environments, 10(3), 266-281. 

  
Singh-Manoux, A., Kivimaki, M., Glymour, M. M., Elbaz, A., Berr, C., Ebmeier, K. P., . . . 
Dugravot, A. (2012). Timing of onset of cognitive decline: results from Whitehall II 
prospective cohort study. BMJ, 344. doi:10.1136/bmj.d7622 

 
Slater, M., & Wilbur, S. (1997). A framework for immersive virtual environments (FIVE): 

Speculations on the role of presence in virtual environments. Presence: Teleoperators 
and virtual environments, 6(6), 603-616.  

 
Smith, P., Blumenthal, J., Hoffman, B., Cooper, H., Strauman, T., Welsh-Bohmer, K., . . . 
Sherwood, A. (2010). Aerobic Exercise and Neurocognitive Performance: a Meta-
Analytic Review of Randomized Controlled Trials. Psychosom Med, 72(3), 239-252.  

 
Staiano, A. E., & Calvert, S. L. (2011). Exergames for physical education courses: Physical, 

social, and cognitive benefits. Child development perspectives, 5(2), 93-98.  

 
Stern, Y. (2012). Cognitive reserve in ageing and Alzheimer's disease. Lancet neurology, 11(11), 

1006-1012. doi:10.1016/S1474-4422(12)70191-6 

 
Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal of 

communication, 42(4), 73-93.  

 
Valenzuela, M. J., Sachdev, P., Wen, W., Chen, X., & Brodaty, H. (2008). Lifespan mental 

activity predicts diminished rate of hippocampal atrophy. PloS one, 3(7), e2598.  

 
Wang, Z. Y., & Duff, B. R. L. (2016). All Loads Are Not Equal: Distinct Influences of 

Perceptual Load and Cognitive Load on Peripheral Ad Processing. Media Psychology, 
19(4), 589-613. doi:10.1080/15213269.2015.1108204 

 
Watson, K., Carlson, S., Gunn, J., Galuska, D., O'Connor, A., Greenlund, K., & Fulton, J. 

(2016). Physical Inactivity Among Adults Aged 50 Years and Older-United States, 2014. 
MMWR. Morbidity and mortality weekly report, 65(36), 954-958.  

 
Wirth, W., Hartmann, T., Böcking, S., Vorderer, P., Klimmt, C., Schramm, H., . . . Jäncke, P. 

(2007). A process model of the formation of spatial presence experiences. Media 
Psychology, 9(3), 493-525. doi:10.1080/15213260701283079 

 
Witmer, B. G., & Singer, M. J. (1998). Measuring presence in virtual environments: A presence 

questionnaire. Presence: Teleoperators and virtual environments, 7(3), 225-240.  

 

99 

 
World  Health Organization. (2015). Dementia. Retrieved from 

http://www.who.int/mediacentre/factsheets/fs362/en/ 

 
Yasamy, M., Dua, T., Harper, M., & Saxena, S. (2013). Mental health of older adults, addressing 

a growing concern. Mental health and older people, 4-9.  

 
 

 

100