Boston University
OpenBU http://open.bu.edu
Theses & Dissertations Boston University Theses & Dissertations
2017
Resting state functional
connectivity in the default mode
network and aerobic exercise in
young adults
https://hdl.handle.net/2144/23793 Boston University
BOSTON UNIVERSITY SCHOOL OF MEDICINE
Thesis
RESTING STATE FUNCTIONAL CONNECTIVITY IN THE DEFAULT MODE
NETWORK AND AEROBIC EXERCISE IN YOUNG ADULTS
by
ANDREW GOSS
B.S., University of Denver, 2013
Submitted in partial fulfillment of the requirements for the degree of
Master of Science 2017
© 2017 by
ANDREW GOSS All rights reserved
Approved by
First Reader
Karin Schon, Ph.D.
Assistant Professor of Anatomy & Neurobiology
Second Reader
Peter Cummings, M.Sc., M.D.
ACKNOWLEDGMENTS
I want to thank Dr. Karin Schon for the support, guidance, and opportunity in the field of aerobic exercise and brain plasticity that made this project possible. I also want to thank my second reader Dr. Peter Cummings for his guidance. Additionally, this project would not have been possible without Olivia Lanman, who has been instrumental in assisting me develop the procedures, commands, and persistence that this project
required. Rachel Nauer and Matthew Dunne also deserve recognition for the knowledge and support they provided. Lastly, I would like to thank all students and staff that have assisted the Brain Plasticity and Neuroimaging Laboratory in data collection, allowing me the opportunity to compete this project.
RESTING STATE FUNCTIONAL CONNECTIVITY IN THE DEFAULT MODE
NETWORK AND AEROBIC EXERCISE IN YOUNG ADULTS
ANDREW GOSS
ABSTRACT
Around the world Alzheimer’s Disease (AD) is on the rise. Previous studies have shown the default mode network (DMN) sees changes with AD progression as the disease erodes away cortical areas. Aerobic exercise with significant increases to cardiorespiratory fitness could show neuro-protective changes to delay AD. This study will explore if functional connectivity changes in the DMN can be seen in a young adult sample by using group independent component analysis through FSL MELODIC. The young adult sample of 19 were selected from a larger study at the Brain Plasticity and Neuroimaging Laboratory at Boston University. The participants engaged in a twelve-week exercise intervention in either a strength training or aerobic training group. They also completed pre-intervention and post-intervention resting-state fMRI scans to evaluate change in functional connectivity in the default mode network.
Cardiorespiratory fitness was assessed using a modified Balke protocol with pre-intervention and post-pre-intervention VO2 max percentiles being used. Through two
repeated-measure ANOVA analyses, this study found no significant increase in mean functional connectivity or cardiorespiratory fitness in the young adult sample. While improvements in mean VO2 max percentile and functional connectivity would have been
does not improve significantly, neither will functional connectivity in the default mode network.
TABLE OF CONTENTS
TITLE………...i
COPYRIGHT PAGE………...ii
READER APPROVAL PAGE………..iii
ACKNOWLEDGMENTS ... iv
ABSTRACT ... v
TABLE OF CONTENTS ... vii
LIST OF TABLES ... ix
LIST OF FIGURES ... x
LIST OF ABBREVIATIONS ... xi
1. INTRODUCTION ... 1
1.1. Aerobic Exercise in Rodents ... 2
1.2. Aerobic Exercise in Humans ... 3
1.3. Functional Magnetic Resonance Imaging ... 3
1.4. Default Mode Network ... 4
1.5. Anatomy of the Hippocampus ... 5
1.6. Cardiorespiratory Fitness ... 6
2. METHODS ... 8
2.1. Participants ... 8
2.2. VO2 Max Testing ...10
2.3. Default Mode Network Connectivity Methods ...11
2.4. MRI Data Acquisition ...11
2.5. fMRI Data Pre-Processing ...12
2.6. Statistical Analysis ...13
3. RESULTS ... 15
3.1. VO2 Max Percentile Adjusts with Exercise ...15
3.2. Group ICA Locates the Default Mode Network ...15
3.3. Default Mode Network Connectivity ...17
4. DISCUSSION ... 19
4.1. Cardiorespiratory Fitness in Young Adults ...19
4.2. Resting-State Functional Connectivity in Default Mode Network ...20
4.3. Limitations ...21
4.4. Future Directions ...22
4.5. Conclusion ...23
REFERENCES ... 24
LIST OF TABLES
Table Title Page
LIST OF FIGURES
Figure Title Page
1 Mean VO2 Max Percentile 16
2 Component 6 – The Default Mode Network 17
LIST OF ABBREVIATIONS
ACSM ... American College of Sports Medicine AD ... Alzheimer’s Disease ANOVA ... Analysis of Variance BET ... Brain Extraction Tool BOLD ... Blood Oxygen Level-Dependent CDC ... Centers for Disease Control and Prevention DMN ... Default Mode Network fMRI ... Functional Magnetic Resonance Image FMRIB ... Functional Magnetic Resonance Imaging of the Brain FSL ... FMRIB Software Library GICA ... Group Independent Component Analysis ICA ... Independent Component Analysis MELODIC ... Multivariate Exploratory Linear Optimized Decomposition into ICA MNI ... Montreal Neurological Institute NIH ... National Institutes of Health rsfMRI ... Resting State Functional Magnetic Resonance Image SPSS ... Statistical Package for the Social Sciences VO2 Max ... Maximal Oxygen Uptake in mL per kg of body weight per minute
1. INTRODUCTION
Projections show that by the year 2030 one-fifth of Americans will be 65 years or older and with the increase in age comes an increase in health care spending (CDC, 2007; He et al., 2005). One of the main components of increased health care cost is declining mental health which is a risk factor for Alzheimer’s disease (AD) (NIH, 2007; van Gelder et al., 2007). AD is the leading cause of dementia beginning with impaired memory and has neuro-pathological hallmarks including diffuse and neuritic extracellular amyloid plaques in the brain that are frequently surrounded by dystrophic neurites and intra-neuronal neurofibrillary tangles (Mayeux and Stern, 2012). These plaques and tangles starve brain areas as the disease progresses until AD ultimately causes death.
More specifically, cognitive impairments in patients with AD are associated with progressive degeneration of limbic system (Arnold et al., 1991; Klucken et al., 2003), neocortical regions (Terry et al., 1981), and basal forebrain areas (Teipel et al., 2005). These areas control functions such as emotion, behavior, motivation, senses, and long-term memory to name a few. The hippocampus and entorhinal cortex are among the first brain areas to show cognitive decline in AD, resulting in issues forming new memories (Serrano-Pozo et al., 2011).
Without a cure for AD, methods to delay the onset or diminish the severity of the disease are being explored. One of such ways is aerobic exercise through effectively remediating the function of brain networks disrupted by mild cognitive impairment (Huang et al., 2016). Previous studies have also shown the default mode network (DMN) sees changes with AD as the disease erodes away cortical areas (Chhatwal et al., 2013;
Jones et al., 2011; Thomas et al., 2014; Petrella et al., 2011). These changes continue to be explored as we work towards finding a cure for this debilitating disease.
1.1. Aerobic Exercise in Rodents
Aerobic exercise has been established to promote hippocampal neurogenesis in young adult mice that performed voluntary wheel running (Voss et al., 2013; van Praag, Kempermann, and Gage, 1999). Hippocampal neurogenesis is the process by which progenitor cells undergo mitosis and generate new cells that differentiate into
functionally integrated neurons (Bruel-Jungerman, Rampon, and Laroche, 2007). Van Praag and associates also noted the dentate gyrus can double or even triple due to
exercise in rodents. The dentate gyrus is one of the subfields within the hippocampus, and contains specific cell types that contribute to learning and memory processes (Kesner, 2007).
Similar studies of aerobic exercise in young adult mice have also shown improved resilience to stress-inducing mitochondrial reactive oxygen species in arteries (Gioscia-Ryan et al., 2016). Furthermore, these improvements in memory function and
hippocampal neurogenesis extend throughout middle-aged mice who participate in long-term running (Marlatt, Potter, Lucassen, and van Praag, 2012). This indicates long-long-term improvements can be quantified and that it is never too late to start exercising regularly.
Cognition, assessed by reference and working memory tasks in the Morris water maze, also changes with moderate treadmill running which was shown to prevent spatial learning and memory deficits in aged rats (Vanzella et al., 2017). In a similar study,
distance and latency in the Morris water maze task increased in AD-simulated rats that did not undergo running on a treadmill once a day for thirty minutes during a four-week intervention (Baek and Kim, 2016). These studies have focused on the hippocampal memory system. With our approach, we can examine how communication between the hippocampus and other regions in the brain can change in response to exercise training.
1.2. Aerobic Exercise in Humans
The findings within rodent models showing the benefits of aerobic exercise have also been demonstrated in humans. Increased performance on multiple tests of executive function have been demonstrated following an aerobic exercise intervention in older adult women (Baker et al., 2010). Furthermore, interestingly Baker and associates found that aerobic exercise increased plasma levels of insulin-like growth factor I in men.
In addition to executive function performance, aerobic exercise has been shown to increase neurogenesis in various brain areas including the hippocampus (Kim et al., 2016), which is an area attacked by AD (Serrano-Pozo et al., 2011). The hippocampus is typically larger in higher-fit adults. So the benefits are clear, yet further studies are needed to explore if these beneficial changes positively impact onset or severity of AD.
1.3. Functional Magnetic Resonance Imaging
Of the many neuroimaging modalities, functional magnetic resonance imaging (fMRI) is different because the signal intensity in fMRI is sensitive to the amount of oxygen carried by hemoglobin (Thulborn, Waterton, Matthews and Radda, 1982). The
ability to view brain activation through blood flow is usually referred to as the blood oxygen level-dependent (BOLD) signal (Ogawa, Lee, Kay and Tank, 1990). Therefore, increased brain activity can be viewed because hemoglobin represents the change in blood oxygen content. The BOLD signal correlates strongly with the underlying local field potential, a measure of peri-synaptic activity, confirming neural activity indirectly drives the BOLD signal (Logothetis, 2008). The images for fMRI analysis are taken through a typical MRI and the subject can participate in either an assigned task during the scans or simply stare at a dot, or cross, without any stimulation (i.e. task-independent scan) for resting-state fMRI.
1.4. Default Mode Network
One of the main resting-state networks, the default mode network (DMN), emerged from two distinct lineages: research on functional connectivity during the resting-state (Biswal, Yetkin, Haughton and Hyde, 1995), and observation of brain regions that were consistently higher in activity during rest, rather than during a task (Buckner, 2012; Snyder and Raichle, 2012). A resting-state network defines the activity in the brain that is not directed to the outside world. This means the brain does not
receive input from the external environment when it is in the resting-state and typically is active when the brain is in a wakeful state but experiencing no tasks (Rosazza and Minati, 2011).
The DMN is defined as including a large portion of the ventral medial prefrontal cortex extending dorsally and ventrally with medial parietal cortex also prominent,
comprising the posterior cingulate cortex, and is involved in many different functions (Uddin et al., 2009). It is known that there is a positive correlation between
cardiorespiratory fitness and functional connectivity within the default mode network in older adults, due to the increased functional connectivity following exercise (McFadden et al., 2014; Voss et al., 2010; Kramer and Erickson, 2007). Kramer and associates also suggested that functional networks most sensitive to age-related decline are also sensitive to individual differences in fitness. These functional connectivity differences improved in the DMN of older adults after a one-year aerobic exercise intervention (Voss et al., 2010).
AD has also been linked to the DMN through the destruction of associated cortical areas causing functional connectivity to be decreased (Chhatwal et al., 2013). These connectivity losses are accelerated in AD as compared to normal aging in age-matched controls (Jones et al., 2011).
1.5. Anatomy of the Hippocampus
The current basis of knowledge on hippocampal connectivity and function comes largely from studies in rodents and monkeys. From such studies, we know the
hippocampal formation is a prominent C-shaped structure along the floor of the temporal horn of the lateral ventricle with the hippocampus proper consisting of three major subfields: CA1, CA2, and CA3 (Schultz and Engelhardt, 2014). Schultz and Engelhardt also state that the other regions that together comprise the hippocampal formation consist of the dentate gyrus, the subicular complex, and the entorhinal cortex. The circuitry of
this area in the macaque monkey demonstrate that projections from the entorhinal cortex terminate in the subiculum, CA1, CA2, and CA3 of the hippocampus, and the dentate gyrus (Witter and Amaral, 1991). Lastly, extensive hippocampal integration of sensory information is established by a largely unidirectional chain of intrinsic hippocampal projections (Schultz and Engelhardt, 2014).
1.6. Cardiorespiratory Fitness
The rate of maximal volume of oxygen uptake (VO2 max) indicates an
individual’s aerobic physical fitness and has been shown to increase with aerobic exercise training (Mezzani et al., 2013). VO2 max can be determined through use of a treadmill
walking test where a participant’s breath, using a gas analyzer placed around their nose and mouth, measures VO2 max (Sousa, Vilas-Boas, and Fernades, 2014), or more
commonly, through use of a sub-maximal test where other measures such as maximal heart rate and resting heart rate are used to calculate VO2 max (Fitchett, 1985). Besides
being a measure of physical fitness, cardiorespiratory fitness has also been linked to functional connectivity within the brain (Wong et al., 2015). Wong and associates found greater anterior cingulate and supplementary motor cortex activation during performance of a dual-task in participants with higher cardiorespiratory fitness.
1.7. Independent Component Analysis
There are many ways to investigate resting-state fMRI scans, but independent component analysis (ICA) is unique as it does not require a hypothesis (i.e. “seed” for
seed-based approach) for analysis. ICA has emerged as a powerful technique for the analysis of functional neuroimaging data by identifying brain regions with temporally coherent activity and linearly decomposing that data into a set of spatial components with corresponding activations (McKeown et al., 1998; Calhoun et al., 2009). ICA is different than conventional regression approaches because it does not require a temporal model. This makes ICA a good tool for investigations where brain activation is difficult to specify, such as resting-state studies.
Furthermore, group independent component analysis (GICA) allows for
straightforward application of ICA to data sets with multiple subjects without merging components (Calhoun et al., 2001; Esposito et al., 2005). GICA uses temporal
concatenation of the individual datasets to identify group-level components that can be projected back to estimate subject-specific components. The components would include individual spatial maps and activation time courses. This approach has been used in numerous studies to describe differences between subjects within intrinsic networks (Rytty et al., 2013; Kim et al., 2009; Maneshi, Vahdat, Gotman, and Grova, 2016).
GICA will be applied for this study to analyze functional connectivity in the DMN of nineteen young adult participants who completed a twelve-week exercise intervention. We seek for a strengthened DMN following exercise as compared to baseline. Additionally, cardiorespiratory fitness will be evaluated to determine if the exercise intervention improved in this area as well.
2. METHODS
2.1. Participants
Participants for this study were selected from a database of a larger study
conducted by the Brain Plasticity and Neuroimaging Laboratory at the Boston University School of Medicine. Both this study and the larger study explore the effects of
cardiorespiratory fitness and human brain function. Participants from the larger study were native English speakers in the greater Boston area that were sedentary and between the ages of 18 and 35 or 55 and 85 years. Sedentary was defined as less than 30 minutes of moderate intensity physical activity three days each week for at least three months (American College of Sports Medicine et al., 2010). Additionally, the participants reported no neurological or psychiatric conditions at the time of the study, and provided signed, informed consent prior to participation in the study. Other MRI counter-indicators used as exclusion criteria were if the participant had any metal in the body that could not be removed, took cardio-active medication, or used any psycho-active medication. Additional exclusion criteria ascertained if each participant had any heart conditions, been diagnosed with hypertension or high cholesterol, been diagnosed with any respiratory conditions, currently have a musculoskeletal impairment, been told by a doctor they were obese, or been diagnosed with diabetes mellitus. This information was used to ensure the exercise training and testing was safe for the subject to participate.
The study itself consisted of a series of initial testing to determine baseline data, followed by a twelve-week exercise intervention, ending with a repeat of the initial tests. At the start of the exercise intervention, participants were randomly assigned and divided
into strength and aerobic groups for the duration of the intervention. The strength group was the control and trained using only the leg press, leg curl, chest press, and lat pull-down machines. The aerobic group only trained using a treadmill. All study procedures followed the guidelines set by the Code of Ethics of the World Medical Association and were approved by the Boston University Medical Campus Institutional Review Board.
From the larger study database, nineteen young adult participants (ages 18 to 35) were selected who completed: the initial resting-state fMRI scan, the entire twelve-week exercise intervention, and the final resting-state fMRI scan. Young adult participants who did not meet these requirements were excluded. See Table 1 for further participant
details.
Table 1: Study Participant Details
This table represents the participant demographics relevant to this study.
N 19 NFemale 15 Strength Group 12 Aerobic Group 7 Age Range 20 - 31 Mean Age 26 (25.6)
VO2 Max Percentile Range Pre-Intervention 2 – 96
VO2 Max Percentile Overall Median Pre-Intervention 49
VO2 Max Percentile Aerobic Group Median Pre-Intervention 49
VO2 Max Percentile Strength Group Median Pre-Intervention 50.5
VO2 Max Percentile Range Post-Intervention 4 – 96
VO2 Max Percentile Overall Median Post-Intervention 52
VO2 Max Percentile Aerobic Group Median Post-Intervention 53
2.2. VO2 Max Testing
Cardiorespiratory fitness was estimated using a modified Balke protocol to estimate VO2 max, where the walking speed was fixed while grade was increased
incrementally (American College of Sports Medicine, Thompson, Gordon, & Pescatello, 2010; Cooper & Storer, 2001). During the test, heart rate was taken every minute with blood pressure taken every three minutes. The test would be terminated if the
participant’s heart rate exceeded their projected maximum or the participant requested the test to stop. Data from the modified treadmill test was used to approximate VO2 max by applying the age and sex predicted maximum heart rate from a plot of heart rate vs. VO2. This VO2 was calculated using the American College of Sports Medicine (ACSM) metabolic equation for Gross VO2 for walking:
𝑉𝑂# = 0.1 )*/
,- ∙)/0 × 𝑆 + 1.8 )*/,- ∙)/0 𝑆𝐺 + 3.5 )*/,- ∙)/0
VO2, S, and G represent gross oxygen consumption (ml / (kg ∙ min)), speed (m / min), and
the fraction being the percent of treadmill grade, respectively (American College of Sports Medicine et al., 2010). A VO2 max percentile for each selected participant was
calculated based on VO2 max norms by age and sex (American College of Sports
2.3. Default Mode Network Connectivity Methods
The initial resting-state fMRI scan (pre-intervention scan) and final resting-state fMRI scan (post-intervention scan) were identified from a larger study conducted by the Brain Plasticity and Neuroimaging Laboratory at the Boston University School of Medicine. Older adults, participants who did not have both pre-intervention and post-intervention MR scans, and participants who did not complete the full twelve-week exercise intervention were excluded. The remaining nineteen young adult participants, with pre-intervention and post-intervention scans, determined the 38 total inputs for this study.
2.4. MRI Data Acquisition
A 3-Tesla Philips Achieva scanner, paired with an 8-channel SENSE head coil, was used to obtain the structural and functional fMRI data at the Boston University Center for Biomedical Imaging in Boston, MA, USA. For each participant, a structural T1
– weighted (T1W) magnetization prepared rapid acquisition gradient echo (MP-RAGE) structural image was obtained. Two hundred functional BOLD volumes were obtained during the ten-minute resting state fMRI scan for each participant. TR and TE were set to 3000ms and 30ms, respectively, with a slice thickness of 3.3125 mm and flip angle of 80 degrees. For the resting state scan, the participant simply stared at a dot on the screen for the duration of the entire ten minutes. The BOLD images were obtained with a T2
*
– weighted (TW2) echo-planar imaging sequence using an in plane acquisition resolution of 3.3125 mm2
2.5. fMRI Data Pre-Processing
The group resting-state fMRI data was preprocessed by normalizing anatomical T1-weighted scans and functional resting-state scans to the standard MNI template
provided by FSL using fslreorient2std. The data was then input to FSL 5.0.6 MELODIC 3.14 for Independent Component Analysis with multi-session temporal concatenation to decompose data sets into different spatial and temporal components that were common across all participants (Woolrich et al., 2009; Smith et al., 2004; Jenkinson et al., 2012).
Data sets for each participant’s scan were prepared by converting the 3-Dimensional images into a single 4-3-Dimensional (i.e. 3-Dimension image plus time) image. Structural images for use in registration were brain-extracted using a Brain Extraction Tool (BET) (Smith, 2002; Jenkinson et al., 2005). The purpose of the Brain Extraction Tool is to clean the image by removing non-brain tissue from the image. The 4D image files were then registered to each other using the FEAT functionality within MELODIC so filtered functional data is adjusted into the standard space
MNI152_T1_2mm for higher-level analysis. The MCFLIRT function in FSL was used to account for motion correction (Jenkinson et al., 2002). Recommended defaults in FSL that were applied to the data include: spatial smoothing set to 5mm, warp resolution was 10mm, with high-pass temporal filtering and resampling resolution set to 4mm (Ryu et al., 2016). Automatic dimensionality was selected to determine the number of output components, and the independent component maps threshold level for activation was set to 0.5.
Through preliminary visual inspection, component 6 was identified as the DMN. In order to confirm this identification, Yeo Liberal 7-Network cortical parcellation was transformed into FSL standard space (4mm resolution) and all component spatial maps were cross-correlated to each network (Yeo et al., 2016). The spatial maps from the group ICA analysis were then used for dual regression to generate subject specific versions of the spatial maps and associated time-series (Beckmann et al., 2009; Filippini et al., 2009). This was achieved in stages where first, through multivariate spatial
regression, the group spatial maps were regressed onto each subjects 4D dataset to give a set of time-courses. This gives a set of subject-specific time-series, with one per group-level spatial map. Then through multivariate temporal regression, those time-courses were regressed into the same 4D dataset to give a subject-specific set of spatial maps. Randomise was used for running cross-subject statistics separately for each group-ICA component (Winkler et al., 2014).
2.6. Statistical Analysis
The statistical analysis utilized a paired two-group difference matrix model with the subject-specific spatial maps together with between-subject design files to assess connectivity changes between pre-intervention and post-intervention scans for each subject. After the paired t-test with dual regression, the z scores from stage two
component 6 maps were extracted and used in SPSS as a repeated measures ANOVA to assess average connectivity changes across all voxels in the Default Mode Network.
The pre-intervention and post-intervention VO2 max percentile for each selected
participant were compared using a repeated measures ANOVA in SPSS. This was to analyze differences between pre-intervention and post-intervention percentiles as well as differences between aerobic and strength groups.
3. RESULTS
3.1. VO2 Max Percentile Adjusts with Exercise
Cardiorespiratory fitness was assessed for 19 healthy, young adult participants. 12 of the 19 were in the strength group, with the other 7 being in the aerobic group. 15 of the 19 were female, yet all fell within an age range from 20 to 31 years of age. The mean age of the group was 26 as seen in Table 1. Participant pre-intervention VO2 max
percentile ranged from 2nd
percentile to 96th
percentile per the modified Balke protocol and ACSM metabolic equation for gross VO2 for walking. The post-intervention VO2
max percentile range was 4th
percentile to 96th
percentile. Median values for VO2 max
percentile before the intervention were 49, 49, and 50.5 for the overall, aerobic, and strength groups, respectively. Post-intervention median VO2 max percentile values were
52, 53, and 50.5 for the overall, aerobic, and strength groups, respectively.
The pre-intervention and post-intervention VO2 max percentile for each selected
participant were compared using a repeated measures ANOVA in SPSS to analyze differences between pre-intervention and post-intervention percentiles as well as differences between aerobic and strength groups as shown in Figure 1. The between-subjects repeated-measures ANOVA yielded a slight increase (f(1) = 0.436, p = 0.519) as well as the between-groups measure (f(1) = 0.256, p = 0.621), see Figure 1.
3.2. Group ICA Locates the Default Mode Network
The 38 total inputs for this study, coming from 19 pre-intervention and 19 post-intervention resting-state fMRI (rsfMRI) scans, were decomposed in FSL MELODIC
GICA resulting in 34 spatially mapped independent components. After visual inspection, component 6 was identified as the DMN. To confirm, Yeo Liberal 7-Network cortical parcellation (Yeo et al., 2011) was transformed into FSL standard space (4mm resolution)
Figure 1: Mean VO2 Max Percentile
This graph shows the difference in mean VO2 max percentiles between the groups over time.
Within the bar cluster, AER (in blue) represents the aerobic exercise intervention group. STR (in green) represents the strength exercise intervention group. On the x-axis, PRE is the
pre-intervention measurements. POST is the post-pre-intervention measurements with the mean VO2 max
and all component spatial maps were cross-correlated to each network resulting in component 6, as seen in Figure 2, being identified as the strongest correlation to the DMN reference network (r = 0.605).
Figure 2: Component 6 - The Default Mode Network
This image illustrates the voxels of interest from group independent component analysis, matched to a standard brain in sagittal, coronal and transverse views from left to right. The colors indicate the intensity of overlap between our DMN and the Yeo mask, with yellow being the highest intensity overlap, followed by orange, then red areas with the least intensity overlap. For the coronal and transverse images, note that right is on the left side of the image, with left being on the right side.
3.3. Default Mode Network Connectivity
After GICA identified the default mode network as component 6, z scores from stage two component 6 maps were extracted and used in SPSS with a repeated-measures ANOVA to assess average connectivity changes across all voxels in the DMN. As seen in Figure 3, the within-subjects result showed the change in DMN connectivity from pre-intervention to post-pre-intervention (f(1) = 0.722, p = 0.407). The between-groups
assessment showed no significant change between aerobic and strength groups (f(1) = 0.913, p = 0.353).
Figure 3: Mean Connectivity in the Default Mode Network
This figure shows the component 6 z-scores for functional connectivity in the DMN. Within the bar cluster, AER (in blue) represents the aerobic exercise intervention group. STR (in green) represents the strength exercise intervention group. On the x-axis, PRE is the pre-intervention measurements. POST is the post-intervention measurements with mean connectivity on the y-axis. Error bars indicate the standard deviation.
4. DISCUSSION
The purpose of this study was to investigate changes of functional connectivity in the default mode network of 19 young adults following a 12-week exercise intervention. Due to the nature of independent component analysis, no hypothesis was used for this study. However, we expect functional connectivity within the DMN to be strengthened following aerobic exercise training. Previous studies indicate correlations between cardiorespiratory fitness and connectivity within the brain in older adults (Wong et al., 2015). This study sought to add to previous research by determining connectivity changes in the default mode network for young adults.
To achieve this, the main goals of this study were to assess cardiorespiratory fitness through VO2 max treadmill testing, use resting-state fMRI scans in group
independent component analysis with FSL MELODIC to determine functional
connectivity in the default mode network, and compare the baseline and post-intervention values along with aerobic and strength groups for both cardiorespiratory fitness and DMN connectivity. This is important because aerobic exercise has been shown to
increase neurogenesis in various brain areas including the hippocampal gyrus (Kim et al., 2016), which is an area attacked by Alzheimer’s Disease (Serrano-Pozo et al., 2011). Increasing functional connectivity in the brain may help delay the onset of AD.
4.1. Cardiorespiratory Fitness in Young Adults
In Figure 1, the mean VO2 max percentiles for the 19 participants are graphed by
we can visually see that due to the exercise intervention, VO2 max percentiles increase
slightly for both aerobic and strength groups from baseline. However, through the repeated-measures ANOVA in SPSS this study found no significant change in VO2 max
percentile from baseline to post-intervention (f(1) = 0.436, p = 0.519). The between groups change also showed no significant change (f(1) = 0.256, p = 0.621).
4.2. Resting-State Functional Connectivity in Default Mode Network
In addition to the cardiorespiratory fitness assessment, the component for the default mode network was located through group independent component analysis in FSL MELODIC and found to be component 6 from the 34 total components. Figure 2 shows this component mapped to a standard brain with the cortical parcellation for the DMN from Yeo and associates (Yeo et al., 2016). Our component was most correlated with that mask (r=0.605), indicating our component 6 is the default mode network.
Component 6 was then used to measure functional connectivity by extracting the z scores from the stage two dual regression maps and used in a repeated measures ANOVA in SPSS to view average connectivity changes. Figure 3 shows these baseline and post-intervention mean connectivity scores with aerobic and strength groups. This study found no mean connectivity change in the default mode network from baseline (f(1) = 0.722, p = 0.407), or any significant change between the aerobic and strength groups (f(1) = 0.913, p = 0.353).
We know from previous studies that cardiorespiratory fitness and DMN
that if one is insignificant the other would be as well. Just because both VO2 max and
connectivity are insignificant does not mean exercise does not change DMN connectivity. Other studies have shown a positive correlation between cardiorespiratory fitness and functional connectivity within DMN in older adults (Voss et al., 2010). By neither VO2
max percentile or connectivity within DMN changing, our findings align with that correlation.
4.3. Limitations
There were many limitations to consider that may have resulted in these findings. First, as seen in Table 1, the participant demographics show this was a smaller study with only 19 participants. As with most studies, a larger pool of participants may help yield significant findings.
Additionally, the findings for aerobic and strength group differences in
cardiorespiratory fitness and connectivity were also insignificant. This could be due to the uneven sample with the strength group having 12 and aerobic having 7 participants. Cardiorespiratory fitness typically increases more when strength and aerobic exercise is combined rather than separated (Cadore et al., 2014). More aerobic group participants may have helped achieve a significant result.
There was a substantial amount of participants that were above the 50th
percentile for cardiorespiratory fitness. This may have accounted for the VO2 max percentile scores
not increasing significantly as the baseline assessment ranged from 2nd
to 96th
We know that higher fit individuals (above 50th
percentile) will have to work harder to increase their VO2 max (Hayes et al., 2013).
This intervention was also only 12-weeks which is comparatively short to a year long intervention which could see further VO2 max percentile increases (Hill et al.,
1993). Lastly, this intervention included aerobic exercise 3 times a week. If the
intervention cannot be longer, perhaps a more intense, daily exercise, intervention may yield more significant results.
Finally, group independent component analysis in FSL is limited in doing repeated-measures ANOVA with resting-state fMRI data. If the system allowed for this statistical analysis, the baseline and post-intervention mean connectivity may have given a better result and accounted for aerobic and strength groups. Also, as with the VO2 max
percentile, a larger input sample would help make a stronger statistical model.
4.4. Future Directions
Further studies are needed to assess how cardiorespiratory fitness and default mode network connectivity may be used as a predictor of Alzheimer’s disease before it manifests. Additionally, long term studies are needed to examine if increased
cardiorespiratory fitness over years or decades impact the onset of AD. Furthermore, the subjects could be split into high and low fitness groups to determine if functional
connectivity improves more in one of those groups. Lastly, it would be of interest to determine all of the cortical areas that change in result to increased cardiorespiratory fitness because perhaps this differs by region. For example, the functional connectivity
between the DMN and posterior cingulate cortex may improve following aerobic exercise training.
4.5. Conclusion
Through group independent component analysis in FSL MELODIC, resting-state fMRI data was used to measure functional connectivity in the default mode network of 19 young adult participants who completed a 12-week exercise intervention in either an aerobic or strength training group. The baseline and post-intervention VO2 max
percentiles were also compared. While improvements in mean VO2 max percentile and
functional connectivity would have been seen with a larger sample size, neither default mode network functional connectivity or cardiorespiratory fitness changed significantly as a result of this intervention. This study adds to the literature by suggesting if fitness does not improve significantly, neither will functional connectivity in the default mode network.
REFERENCES
American College of Sports Medicine, Thompson, W. R., Gordon, N. F., and Pescatello, L.S. (2010). ACSM’s guidelines for exercise testing and prescription.
Philadelphia:Wolters Kluwer/Lippincott Williams & Wilkins.
Arnold SE, Hyman BT, Flory J, Damasio AR, Van Hoesen GW. (1991). The
topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer’s disease. Cereb Cortex 1: 103–116
Baek, S.-S., & Kim, S.-H. (2016). Treadmill exercise ameliorates symptoms of
Alzheimer disease through suppressing microglial activation-induced apoptosis in rats. Journal of Exercise Rehabilitation, 12(6), 526–534.
http://doi.org/10.12965/jer.1632858.429
Baker, L. D., Frank, L. L., Foster-Schubert, K., Green, P. S., Wilkinson, C. W., McTiernan, A., … Craft, S. (2010). Effects of Aerobic Exercise on Mild
Cognitive Impairment: A Controlled Trial. Archives of Neurology, 67(1), 71–79. http://doi.org/10.1001/archneurol.2009.307
Beckmann, C. F., Mackay, C. E., Filippini, N., and Smith, S. M. (2009) Group
comparison of resting-state FMRI data using multi-subject ICA and dual regression. OHB.
Biswal, B., Yetkin, F. Z., Haughton, V. M., & Hyde, J. S. (1995). Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magnetic
Resonance in Medicine, 34(4), 537–541.
Bruel-Jungerman, E., Rampon, C., & Laroche, S. (2007). Adult hippocampal
neurogenesis, synaptic plasticity and memory: facts and hypotheses. Reviews in
the Neurosciences, 18(2), 93–114.
Buckner, R. L. (2012). The serendipitous discovery of the brain’s default network.
NeuroImage, 62(2), 1137–1145.
https://doi.org/10.1016/j.neuroimage.2011.10.035
Cadore, E. L., Pinto, R. S., Bottaro, M., & Izquierdo, M. (2014). Strength and Endurance Training Prescription in Healthy and Frail Elderly. Aging and Disease, 5(3), 183– 195. http://doi.org/10.14336/AD.2014.0500183
Calhoun, V. D., Adali, T., McGinty, V. B., Pekar, J. J., Watson, T. D., & Pearlson, G. D. (2001). fMRI activation in a visual-perception task: network of areas detected using the general linear model and independent components analysis.
NeuroImage, 14(5), 1080–1088. https://doi.org/10.1006/nimg.2001.0921
Calhoun, V. D., Liu, J., & Adalı, T. (2009). A review of group ICA for fMRI data and ICA for joint inference of imaging, genetic, and ERP data. NeuroImage, 45(1 Suppl), S163–S172. http://doi.org/10.1016/j.neuroimage.2008.10.057
CDC (2007). The State of Aging and Health in America 2007. Washington, D.C.: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention
Chhatwal, J. P., Schultz, A. P., Johnson, K., Benzinger, T. L. S., Jack, C., Ances, B. M., … Sperling, R. A. (2013). Impaired default network functional connectivity in autosomal dominant Alzheimer disease. Neurology, 81(8), 736–744.
http://doi.org/10.1212/WNL.0b013e3182a1aafe
Cooper, C. B., and Storer, T. W. (2001). Exercise Testing and Interpretation: A Practical
Approach. Cambridge University Press.
Covey, M. K., Collins, E. G., Reynertson, S. I., & Dilling, D. F. (2014). Resistance training as a preconditioning strategy for enhancing aerobic exercise training outcomes in COPD. Respiratory Medicine, 108(8), 1141–1152.
http://doi.org/10.1016/j.rmed.2014.06.001
Erickson, K. I., Voss, M. W., Prakash, R. S., Basak, C., Szabo, A., Chaddock, L., … Kramer, A. F. (2011). Exercise training increases size of hippocampus and improves memory. Proceedings of the National Academy of Sciences of the
United States of America, 108(7), 3017–3022.
http://doi.org/10.1073/pnas.1015950108
Esposito, F., Scarabino, T., Hyvarinen, A., Himberg, J., Formisano, E., Comani, S., … Di Salle, F. (2005). Independent component analysis of fMRI group studies by self-organizing clustering. NeuroImage, 25(1), 193–205.
https://doi.org/10.1016/j.neuroimage.2004.10.042
Filippini, N., MacIntosh, B. J., Hough, M. G., Goodwin, G. M., Frisoni, G. B., Smith, S. M., Matthews, P.M., Beckmann C.F. and Mackay C.E. (2009). Distinct patterns
of brain activity in young carriers of the APOE-ε4 allele. PNAS,
Fitchett, M. A. (1985). Predictability of VO2 max from submaximal cycle ergometer and bench stepping tests. British Journal of Sports Medicine, 19(2), 85–88.
Gioscia-Ryan, R. A., Battson, M. L., Cuevas, L. M., Zigler, M. C., Sindler, A. L., & Seals, D. R. (2016). Voluntary aerobic exercise increases arterial resilience and mitochondrial health with aging in mice. Aging (Albany NY), 8(11), 2897–2914. http://doi.org/10.18632/aging.101099
Hayes, S. M., Hayes, J. P., Cadden, M., & Verfaellie, M. (2013). A review of
cardiorespiratory fitness-related neuroplasticity in the aging brain. Frontiers in
Aging Neuroscience, 5, 31. http://doi.org/10.3389/fnagi.2013.00031
He W., Sengupta M., Velkoff V. A., DeBarros K. A. (2005). 65+ In the United States: 2005. Washington, D.C.: U.S. Census Bureau
Hill, R., Storandt, M., Malley, M. (1993). The Impact of Long-term Exercise Training on Psychological Function in Older Adults. J Gerontol; 48 (1): P12-P17. doi:
10.1093/geronj/48.1.P12
Huang, P., Fang, R., Li, B.-Y., & Chen, S.-D. (2016). Exercise-Related Changes of Networks in Aging and Mild Cognitive Impairment Brain. Frontiers in Aging
Neuroscience, 8, 47. http://doi.org/10.3389/fnagi.2016.00047
Jenkinson, M., Bannister, P., Brady, J. M. and Smith, S. M. (2002). Improved
Optimisation for the Robust and Accurate Linear Registration and Motion Correction of Brain Images. NeuroImage, 17(2), 825-841.
Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., Smith, S. M. (2012).
FSL. NeuroImage, 62:782-90.
Jenkinson, M., Pechaud, M., and Smith, S. (2005). BET2: MR-based estimation of brain,
skull and scalp surfaces. In Eleventh Annual Meeting of the Organization for
Human Brain Mapping.
Jones, D. T., Machulda, M. M., Vemuri, P., McDade, E. M., Zeng, G., Senjem, M. L., … Jack, C. R. (2011). Age-related changes in the default mode network are more advanced in Alzheimer disease. Neurology, 77(16), 1524–1531.
http://doi.org/10.1212/WNL.0b013e318233b33d
Kim, D. I., Mathalon, D. H., Ford, J. M., Mannell, M., Turner, J. A., Brown, G. G., … Calhoun, V. D. (2009). Auditory Oddball Deficits in Schizophrenia: An Independent Component Analysis of the fMRI Multisite Function BIRN
Kim, D.-Y., Jung, S.-Y., Kim, K., & Kim, C.-J. (2016). Treadmill exercise ameliorates Alzheimer disease-associated memory loss through the Wnt signaling pathway in the streptozotocin-induced diabetic rats. Journal of Exercise Rehabilitation, 12(4), 276–283. http://doi.org/10.12965/jer.1632678.339
Kesner R.P. (2007). Behavioral functions of the CA3 subregion of the hippocampus. Learn Mem.;14:771–781.
Klucken J, McLean PJ, Gomez-Tortosa E, Ingelsson M, Hyman BT. (2003). Neuritic alterations and neural system dysfunction in Alzheimer’s disease and dementia with Lewy bodies. Neurochem Res 28: 1683–1691
Kramer, A. F., & Erickson, K. I. (2007). Capitalizing on cortical plasticity: influence of physical activity on cognition and brain function. Trends in Cognitive Sciences,
11(8), 342–348. https://doi.org/10.1016/j.tics.2007.06.009
Leckie, R. L., Weinstein, A. M., Hodzic, J. C., & Erickson, K. I. (2012). Potential Moderators of Physical Activity on Brain Health. Journal of Aging
Research, 2012, 948981. http://doi.org/10.1155/2012/948981
Logothetis, N. K. (2008). What we can do and what we cannot do with fMRI. Nature,
453(7197), 869–878. https://doi.org/10.1038/nature06976
Maneshi, M., Vahdat, S., Gotman, J., & Grova, C. (2016). Validation of Shared and Specific Independent Component Analysis (SSICA) for Between-Group Comparisons in fMRI. Frontiers in Neuroscience, 10, 417.
http://doi.org/10.3389/fnins.2016.00417
Marlatt, M. W., Potter, M. C., Lucassen, P. J., & van Praag, H. (2012). Running throughout middle-age improves memory function, hippocampal neurogenesis and BDNF levels in female C57Bl/6J mice. Developmental Neurobiology, 72(6), 943–952. http://doi.org/10.1002/dneu.22009
Mayeux, R., & Stern, Y. (2012). Epidemiology of Alzheimer Disease. Cold Spring
Harbor Perspectives in Medicine, 2(8), 10.1101/cshperspect.a006239 a006239.
http://doi.org/10.1101/cshperspect.a006239
McFadden, K. L., Cornier, M.-A., Melanson, E. L., Bechtell, J. L., & Tregellas, J. R. (2013). Effects of exercise on resting-state default mode and salience network activity in overweight/obese adults. Neuroreport, 24(15), 866–871.
McKeown, M. J., Jung, T.-P., Makeig, S., Brown, G., Kindermann, S. S., Lee, T.-W., & Sejnowski, T. J. (1998). Spatially independent activity patterns in functional MRI data during the Stroop color-naming task. Proceedings of the National Academy
of Sciences of the United States of America, 95(3), 803–810.
Mezzani, A., Grassi, B., Jones, A. M., Giordano, A., Corrà, U., Porcelli, S., … Giannuzzi, P. (2013). Speeding of pulmonary VO2 on-kinetics by light-to-moderate-intensity aerobic exercise training in chronic heart failure: clinical and pathophysiological correlates. International Journal of Cardiology, 167(5), 2189–2195.
https://doi.org/10.1016/j.ijcard.2012.05.124
NIH (2007). 2005–2006 Progress Report on Alzheimer's Disease. Washington, D.C.: U.S. Department of Health and Human Services, National Institute on Aging and National Institutes of Health
Ogawa, S., Lee, T. M., Kay, A. R., & Tank, D. W. (1990). Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proceedings of the
National Academy of Sciences of the United States of America, 87(24), 9868–
9872.
Petrella, J. R., Sheldon, F. C., Prince, S. E., Calhoun, V. D., & Doraiswamy, P. M. (2011). Default mode network connectivity in stable vs progressive mild cognitive impairment. Neurology, 76(6), 511–517.
http://doi.org/10.1212/WNL.0b013e31820af94e
Rosazza, C., and Minati, L. (2011). Resting-state brain networks: literature review and clinical applications. Neurological Sciences, 32(5), 773–785.
http://doi.org/10.1007/s10072-011-0636-y
Rytty, R., Nikkinen, J., Paavola, L., Abou Elseoud, A., Moilanen, V., Visuri, A., … Remes, A. M. (2013). GroupICA dual regression analysis of resting state
networks in a behavioral variant of frontotemporal dementia. Frontiers in Human
Neuroscience, 7, 461. http://doi.org/10.3389/fnhum.2013.00461
Ryu, S., Higashi, H., Tanaka, T., Nakauchi, S., & Minami, T. (2016). Spatial smoothing of canonical correlation analysis for steady state visual evoked potential based brain computer interfaces. IEEE Engineering in Medicine and Biology Society.
Annual Conference, 2016, 1516–1519.
https://doi.org/10.1109/EMBC.2016.7590998
Schultz, C., & Engelhardt, M. (2014). Anatomy of the hippocampal formation. Frontiers
Serrano-Pozo, A., Frosch, M. P., Masliah, E., & Hyman, B. T. (2011). Neuropathological Alterations in Alzheimer Disease. Cold Spring Harbor Perspectives in
Medicine: 1(1), a006189. http://doi.org/10.1101/cshperspect.a006189
Smith, S.M. (2002). Fast robust automated brain extraction. Human Brain Mapping,
17(3):143-155.
Smith, S. M., Jenkinson, M., Woolrich, M. W., Beckmann, C. F., Behrens, T. E. J., Johansen-Berg, H., Bannister, P. R., De Luca, M., Drobnjak, I., Flitney, D.E., Niazy, R., Saunders, J., Vickers, J., Zhang, Y., De Stefano, N., Brady, J. M., and Matthews, P. M. (2004). Advances in functional and structural MR image
analysis and implementation as FSL. NeuroImage, 23(S1):208-19.
Snyder, A. Z., & Raichle, M. E. (2012). A Brief History of the Resting State: the Washington University Perspective. Neuroimage, 62(2), 902–910. http://doi.org/10.1016/j.neuroimage.2012.01.044
Sousa, A. C., Vilas-Boas, J. P., & Fernandes, R. J. (2014). VO2 Kinetics and Metabolic
Contributions Whilst Swimming at 95, 100, and 105% of the Velocity at
VO2 max. BioMed Research International, 2014, 675363.
http://doi.org/10.1155/2014/675363
Teipel SJ, Flatz WH, Heinsen H, Bokde AL, Schoenberg SO, Stockel S, Dietrich O, Reiser MF, Moller HJ, Hampel H (2005). Measurement of basal forebrain atrophy in Alzheimer’s disease using MRI. Brain128: 2626–2644
Terry R, Peck A, DeTeresa R, Schechter R, Horoupian D. (1981). Some morphometric aspects of the brain in senile dementia of the Alzheimer type. Ann Neurol 10: 184–192
Thomas, J. B., Brier, M. R., Bateman, R. J., Snyder, A. Z., Benzinger, T. L., Xiong, C., … the Dominantly Inherited Alzheimer Network. (2014). Functional Connectivity in Autosomal Dominant and Late-Onset Alzheimer Disease. JAMA
Neurology, 71(9), 1111–1122. http://doi.org/10.1001/jamaneurol.2014.1654
Thulborn, K. R., Waterton, J. C., Matthews, P. M., & Radda, G. K. (1982). Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field. Biochimica Et Biophysica Acta, 714(2), 265–270.
Uddin, L. Q., Clare Kelly, A. M., Biswal, B. B., Castellanos, F. X., & Milham, M. P. (2009). Functional connectivity of default mode network components: correlation, anticorrelation, and causality. Human Brain Mapping, 30(2), 10.1002/hbm.20531. http://doi.org/10.1002/hbm.20531
van Gelder B. M., Tijhuis M. A. R., Kalmijn S., Giampaoli S., Kromhout D.
(2007). Decline in cognitive functioning is associated with a higher mortality risk. Neuroepidemiology 28, 93–10010.1159/000098552
van Praag, H., Kempermann, G., & Gage, F. H. (1999). Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nature
Neuroscience, 2(3), 266–270. https://doi.org/10.1038/6368
Vanzella, C., Sanches, E. F., Odorcyk, F. K., Nicola, F., Kolling, J., Longoni, A., … Netto, C. A. (2017). Forced Treadmill Exercise Prevents Spatial Memory Deficits in Aged Rats Probably Through the Activation of Na(+), K(+)-ATPase in the Hippocampus. Neurochemical Research. https://doi.org/10.1007/s11064-017-2196-x
Voss, M. W., Prakash, R. S., Erickson, K. I., Basak, C., Chaddock, L., Kim, J. S., … Kramer, A. F. (2010). Plasticity of Brain Networks in a Randomized Intervention Trial of Exercise Training in Older Adults. Frontiers in Aging Neuroscience, 2, 32. http://doi.org/10.3389/fnagi.2010.00032
Voss, M. W., Vivar, C., Kramer, A. F., & van Praag, H. (2013). Bridging animal and human models of exercise-induced brain plasticity. Trends in Cognitive
Sciences, 17(10), 525–544. http://doi.org/10.1016/j.tics.2013.08.001
Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE. (2014). Permutation
inference for the general linear model. NeuroImage;92:381-397.
Witter, M. P., & Amaral, D. G. (1991). Entorhinal cortex of the monkey: V. Projections to the dentate gyrus, hippocampus, and subicular complex. The Journal of
Comparative Neurology, 307(3), 437–459. https://doi.org/10.1002/cne.903070308
Wong, C. N., Chaddock-Heyman, L., Voss, M. W., Burzynska, A. Z., Basak, C., Erickson, K. I., Kramer, A. F. (2015). Brain activation during dual-task
processing is associated with cardiorespiratory fitness and performance in older adults. Frontiers in Aging Neuroscience, 7, 154.
http://doi.org/10.3389/fnagi.2015.00154
Woolrich, M. W., Jbabdi, S., Patenaude, B., Chappell, M., Makni, S., Behrens, T., Beckmann, C., Jenkinson, M., Smith, S. M. (2009). Bayesian analysis of
neuroimaging data in FSL. NeuroImage, 45:S173-86.
Yeo, B. T. T., Krienen, F. M., Sepulcre, J., Sabuncu, M. R., Lashkari, D., Hollinshead, M., … Buckner, R. L. (2011). The organization of the human cerebral cortex estimated by intrinsic functional connectivity. Journal of Neurophysiology,
