Confirmatory Replication Study - Methods and Analyses

Confirmatory Replication Study 2012

1

Interindividual Differences in Behavior and Cognition Predicted by

Local Brain Structure: A Strictly Confirmatory Replication Study.

Authors: Eric-Jan Wagenmakers, Birte Forstmann, Luam Belay, Wouter Boekel.

Abstract

We seek to conduct a purely confirmatory replication study of nine studies that have

previously reported an association between behavior and structural properties of the brain.

In order to ensure that our replication is strictly confirmatory, this document outlines the

details of our experimental design and plan for data analysis.

Background

Recently, much interest has concerned the relation between brain structure on the one

hand and behavior and/or cognition on the other hand. For instance, Kanai and colleagues

(2012) found that individuals with greater gray matter volume in specific brain regions have

larger online social networks (i.e., more Facebook friends) than individuals with less gray

matter volume in these areas. Here, we propose to conduct a strictly confirmatory

replication study for nine recently published studies, each of which associated certain

structural properties of the human brain with certain behavioral measures. Specifically, the

behavioral measures relate to behavioral activation (Xu et al., 2012), control over speed and

accuracy in perceptual decision making (Forstmann et al., 2010), percept duration in

perceptual rivalry (Kanai et al., 2010, 2011a), components of attention (i.e., executive

control and alerting; Westlye et al., 2011), aspects of social cognition (i.e., (online) social

network size; Bickart et al., 2011; Kanai et al., 2012), distractibility (Kanai et al., 2011d), and

political orientation (Kanai et al., 2011c).

In order to ensure that the data analyses and hypothesis tests are strictly

confirmatory, this document describes the nine experiments and associated analyses that

we aim to replicate. At the time of publishing, 36 participants have been tested but none of

the data are analyzed or inspected. The study-specific information is preceded by the


2

‘General Methods and Analyses’ section describing the participants, the general procedure,

the MRI data acquisition, the MRI preprocessing and analysis, and finally the Bayesian

hypothesis test for correlations.

Note that we leave open the possibility of conducting additional exploratory analyses.

However, any exploratory analyses will be clearly labeled as such.

General Methods and Analyses

Local gray matter volume and cortical thickness will be quantified by means of voxel-based

morphometry (VBM) and white matter integrity will be quantified by diffusion tensor

imaging (DTI) and probabilistic tractography.

Participants. Participants are recruited from a 43-participant MRI study that was recently

conducted by our own research group. These participants are young, healthy undergraduate

students (mean age = 20.12, SD = 1.73) from the University of Amsterdam with normal or

corrected-to-normal vision. Participants received a monetary compensation for their time and

effort.

A subset of 36 students has been tested for the present study, but data have not been

analyzed or inspected, yet. The earlier MRI experiment featured extensive measurements of

brain structure, and hence the additional effort involved in replicating the nine studies consisted

primarily in having these participants complete a battery of behavioral tests. Therefore, each of

our nine replication attempts features the same set of participants. However, participants whose

behavioral or structural measures deviate more than 2.5 standard deviations from the respective

mean will be excluded from further analysis.

Even though our initial focus is on the 36 participants who have completed the behavioral

test battery, we leave open the possibility of testing additional participants in order to obtain

clearer results. We are permitted this flexibility in data collection, because we use a Bayesian

hypothesis test to quantify the evidence for and against the null hypotheses. When using the so-

called “Bayes factor”, researchers are allowed to monitor the evidence as the data come in, and

continue testing until a point has been proven or disproven, or until the experimenter runs out of

time, money, or patience (Edwards et al., 1963).

General procedure. Prior to the test session participants receive an information brochure


3

with a brief description of all tasks and questionnaires. At the beginning of the test session

participants sign an informed consent form. Participants are tested in individual computer

booths. All instructions are shown on the computer screen or printed on top of the

questionnaires. Participants begin with filling out the following questionnaires: Behavioral

Inhibition System/Behavioral Activation System Scales, Social Network Index, Social Network

Size Questionnaire, Cognitive Failures Questionnaire, and Political Orientation

Questionnaire. The order in which participants fill out the questionnaires is randomized.

Note that all questionnaires have been translated into Dutch and can be found in Appendix

A. After completing the questionnaires, participants continue with the computerized tasks:

Random dot motion task, bistable structure-from-motion task, and attention network test

(please find a detailed description of each task below). Again, the order of these tasks is

randomized per participant. The total duration of the test session is at most 1 hour and 30

minutes.

MRI data acquisition. In the earlier MRI experiment, DTI and T1 images have been obtained

on a 3T Philips scanner using a 32-channel head coil. For each subject, a T1 anatomical scan

was acquired (T1 turbo field echo, 220 transverse slices of 1 mm, with a resolution of 1mm3,

TR = 8.2 ms, TE = 3.7 ms). In addition, four repetitions of a multi-slice spin echo (MS-SE),

single shot diffusion weighted imaging (DWI) scans were obtained using the following

parameters: TR = 7545 ms, TE = 86 ms, 60 transverse slices, 2 mm slice thickness, FOV: 224 x

224 mm2, voxel size 2 mm isotropic resolution. For each slice, one image without diffusion

weighting (b = 0) and 32 diffusion-weighted images (b=1000s/mm 2) along 32 directions

were acquired.

MRI data preprocessing and analysis. We will carry out all MRI data analyses in FMRIB’s

Software Library (FSL 4.0; www.fmrib.ox.ac.uk/fsl) and FreeSurfer

(http://surfer.nmr.mgh.harvard.edu) software.

VBM preprocessing. Voxel-Based Morphometry will be performed using FSL's

default VBM pipeline (http://www.fmrib.ox.ac.uk/fsl/fslvbm/index.html). First, non-brain

tissue will be removed from T1 images using FMRIB's Brain Extraction Tool (BET). Second,

brain-extracted images will be segmented into gray matter (GM), white matter (WM), and

cerebrospinal fluid (CSF). GM images will be non-linearly registered to GM ICBM-152, and

averaged to create a study-specific template at 2mm resolution in standard space. All GM

http://www.fmrib.ox.ac.uk/fsl

http://surfer.nmr.mgh.harvard.edu/

http://www.fmrib.ox.ac.uk/fsl/fslvbm/index.html


4

images will then be non-linearly registered to the study-specific template. During this stage,

each voxel of each registered grey matter image is divided by the Jacobian of the warp field

(Good et al., 2001). Finally, images will be smoothed using a Gaussian kernel with a sigma of

3 mm.

DTI preprocessing. All four runs of DTI will be merged and corrected for eddy

currents. Affine registration will be used to register each volume to a reference volume. A

single image without diffusion weighting (b=0) will be extracted from the merged data and

non-brain tissue will be removed using BET to create a brain-mask which will be used in

subsequent analyses. DTIFIT will be then be applied to fit a tensor model at each voxel of

the data. Finally, tract based spatial statistics (TBSS) will be run. To compute probabilistic

tractography, Bayesian estimation of diffusion parameters using sampling techniques

(BedpostX) will be applied. Bedpostx uses a dual fiber model which can account for crossing

fibers.

TBSS. Tract-Based Spatial Statistics will be performed using FSL's default TBSS

pipeline (http://www.fmrib.ox.ac.uk/fsl/tbss/index.html). First, FA images are slightly

eroded and end slices are zeroed, in order to remove likely outliers from the diffusion

tensor fitting. Second, all FA images will be aligned to 1 mm standard space using non-linear

registration to the FMRIB58_FA standard-space image. Affine registration is then used to

align images into 1 x 1 x 1 mm MNI152 space, and a skeletonization procedure is

subsequently applied to a mean FA image resulting from averaging all individual MNI-

aligned images. Subsequently, the mean skeletonized FA image is thresholded at > 0.2 in

order to accurately represent white-matter tracts. Subject's FA data are then projected onto

the mean skeletonized FA image and concatenated. Finally, the non-linear warps and

skeletonization procedures resulting from aforementioned FA-TBSS will be used in order to

construct skeletonized images of other measures (i.e., parallel eigenvalue (λ1) and mean

diffusivity (MD)).

Probabilistic tractography. First, affine registration will be used to align relevant

MNI masks to subjects’ individual high-resolution T1 images, and subsequently to subjects’

DTI images. Probabilistic tractography will then be performed using 5.000 tract-following

samples at each voxel with a curvature threshold of 0.2. Single masks will be used as seed

http://www.fmrib.ox.ac.uk/fsl/tbss/index.html


5

space, and we will use classification masks in order to estimate the number of samples

reaching the relevant target mask. In addition, contralateral exclusion masks will be used to

discard pathways crossing over to the contralateral seed mask before travelling to the

classification mask. For the estimation of tract strength, resulting images will be thresholded

at 10 samples and binarized, such that the size of a resulting image will be the amount of

voxels reliably (that is, in at least 10 samples) reaching the classification mask. This size is

then divided by the total size of the seed mask, such that the result is a proportion of the

seed mask, reliably connected to the classification mask. The same is then done for the

opposite analysis (i.e. when the seed mask and classification mask are switched), and the

average of the proportions is computed. These steps are necessary to correct for the sizes of

the masks.

Bayesian hypothesis test for (partial) correlations. The Bayes factor compares the

probability of the observed data under H1 versus H0, and hence quantifies the evidence that

the data provide for and against the models under consideration. In what follows, H0 is the

null model in which correlation is absent, and H1 is the alternative model in which

correlation is present. We distinguish between an H1 that does not commit to a direction for

the correlation (i.e., a two-sided test), and an H1 that does commit to such a direction (i.e., a

one-sided test). The latter test is recommended for replication projects, or anytime

researchers have strong prior expectations about the direction of the association.

The two-sided test. Under H1, Jeffreys assigned the correlation coefficient ρ a

uniform prior distribution from -1 to 1 (Jeffreys, 1961). Further, he assumed that the means

of the observed variables x and y are 0.

Then,

BF10 =

∫

( )

( )

(1)

where n is the number of (x, y) pairs and r is the classic Pearson correlation coefficient.

Note that when the means for x and y need to be estimated from the data, as is usually the

case, n in the above equation needs to be replaced by n - 1.


6

The above Bayes factor contains a uni-dimensional integral that can be computed

numerically in R:

# Jeffreys' earthquake data:

S <- c(-8,-5,-3,3,-3,3,2,0,0,2,6,4,-1,4,0,-1,-7,-8,-3)

SKS <- c(-10,-10,1,-6,1,0,-3,1,-4,0,8,1,0,0,0,-1,-2,-10,-4)

BF10.twosided <- function(n, r){

#Jeffreys' two-sided test for the presence of a correlation;

#Jeffreys (1961), pp. 289-292

#Note that if the means are subtracted, n needs to be replaced by n-1!

integrand <- function(rho) {((1-rho^2)^(n/2)) / ((1-rho*r)^(n-.5))}

BF10 <- integrate(integrand, lower=-1, upper=1)$value/2

return(BF10)

}

# Example:

BF10.twosided(n=length(S)-1, r=cor(S,SKS)) #27.0; note the n-1

Thus, in the above example, BF10 ≈ 27.0, which means that the data are about 27 times

more likely to have occurred under H1 than under H0.

The one-sided test. The two-sided test from the previous section contains an integral

from -1 to 1; hence the height of the uniform prior distribution on is 1/2. The extension to

the one-sided test (where we seek to establish whether or not there is a positive

correlation) simply requires that we integrate from 0 to 1 and take into account that the

height of the uniform prior distribution is now 1. Therefore:

BF

= ∫( )

( )

(2)

A test for the presence of a negative correlation is trivially obtained by multiplying one of

the observed variables with -1. Again, we can easily compute this Bayes factor numerically

in R:

BF10.onesided <- function(n, r){

#One-sided test for the presence of a **positive** correlation;

#see Jeffreys (1961), pp. 289-292

#Note that if the means are subtracted, n needs to be replaced by n-1!

integrand <- function(rho) {((1-rho^2)^(n/2)) / ((1-rho*r)^(n-.5))}

BF10 <- integrate(integrand, lower=0, upper=1)$value

return(BF10)

}

# Example:

BF10.onesided(n=length(S)-1, r=cor(S,SKS)) #53.9; note the n-1


7

Thus, in the above example, BF10 ≈ 53.9, which means that the data are about 54 times

more likely to have occurred under H1 than under H0. The evidence is more compelling for

the one-sided test than for the two-sided test—this makes sense, because the predictions of

the one-sided H1 are more daring than those of the two-sided H1; when daring predictions

are borne out by the data, this increases one’s confidence in the model. A detailed

discussion of the properties of one-sided versus two-sided Bayes factors can be found in

Wagenmakers, Lodewyckx, Kuriyal, and Grasman (2010) and Hoijtink, Klugkist, and Boelen

(2008).

We will only use the one-sided test, because we aim to replicate the specific

correlations reported in the studies-to-replicate.

Study-specific Methods and Analyses

Replication 1. Individuals with high BAS-Total scores (i.e., sensitivity to signals of reward

and non-punishment) show increased parallel eigenvalues within left corona radiata (CR)

and left superior longitudinal fasciculus (SLF). Individuals with high BAS-Fun scores (i.e.,

tendency to seek out new potentially-rewarding experiences) show increased parallel

eigenvalues and fractional anisotropy within left CR and left SLF. These individuals also

show increased mean diffusivity within left inferior longitudinal fasciculus and left inferior

fronto-occipital fasciculus.

Xu, J. Kober, H., Caroll, K. M., Rounsaville, B. J., Pearlson, G. D., & Potenza, M. N. (2012).

White matter integrity and behavioral activation in healthy subjects. Human Brain Mapping,

33, 994-1002.

BIS/BAS questionnaire and procedure. Participants fill out a Dutch version of the Behavioral

Inhibition System/Behavioral Activation System (BIS/BAS; Carver et al., 1994; see Appendix

A) scale. The BIS/BAS is a 20-item questionnaire. Our interest will be focused on the BAS

scale which comprises 13 items (BAS-Total) and has three subscales, Drive (BAS-Drive), Fun-

Seeking (BAS-Fun), and Reward-Responsiveness (BAS-Reward).

Behavioral analysis. The behavioral measures of interest are BAS-Total scores and BAS-Fun

scores. BAS-Total scores assess the sensitivity to signals of reward and non-punishment.


8

BAS-Fun scores assess the tendency to seek out new potentially-rewarding experiences. For

each participant these scores will be imported into R (R Foundation for Statistical

Computing, http://www.R-project.org) for the Bayesian correlation test.

TOI generation. Xu and colleagues (2012) reported significant positive correlations between

the BAS-Total scores and parallel eigenvalue (λ1) within left corona radiata (CR) and left

superior longitudinal fasciculus (SLF). Furthermore, they reported positive correlations

between the BAS-Fun scores and λ1 as well as fractional anisotropy (FA) within left CR and

left SLF. The authors also reported significant positive correlations between the BAS-Fun

scores and mean diffusivity (MD) within left inferior longitudinal fasciculus (ILF) and left

inferior fronto-occipital fasciculus (IFOF). We defined all these white matter (WM) tracts as

our tracts of interest (TOIs). Dr. Xu kindly provided us with the masks/templates that were

used in the original study. We will use these for the segmentation of our confirmatory TOIs.

Correlational analysis. Before performing the Bayesian hypothesis test for correlations (as

described above), we will extract FA, MD, and λ1 values from all voxels contained in the

respective TOIs and average them. This will be done for every subject. These WM tract

measures are then corrected for age and gender using partial correlations. Unlike Xu and

colleagues we will not correct for education. Our participants are all Psychology freshmen,

therefore we can rule out substantial differences in education. The corrected mean WM

tract measures per TOI will be imported into R (R Foundation for Statistical Computing,

http://www.R-project.org) software for the Bayesian correlation test. Specifically, we will

test for positive correlations between BAS-Total scores and mean λ1 within left CR and left

SLF. Furthermore, we will test for positive correlations between BAS-Fun scores and mean λ1

as well as mean FA within left CR and left SLF. Finally, we will test for positive correlations

between BAS-Fun scores and mean MD within left ILF and left IFOF.

Replication 2. Individuals with high LBA flexibility (i.e., good control over speed and

accuracy in perceptual decision making) show increased tract strength of white matter

fibers connecting right pre-SMA and right striatum.

Forstmann, B. U., Anwander, A., Schäfer, A., Neumann, J., Brown, S., Wagenmakers, E.-J.,

Bogacz, R., et al. (2010). Cortico-striatal connections predict control over speed and accuracy

about:blank

about:blank


9

in perceptual decision making. Proceedings of the National Academy of Sciences of the

United States of America, 107(36), 15916-15920.

Random dot motion task and procedure. We use the same random dot motion (RDM) task

(Gold & Shadlen, 2001) as Forstmann and colleagues (2010; see Figure 1).

The experiment features 360 trials in total, with 180 speed and 180 accuracy trials. A cloud

consisting of 120 white dots with 50% coherently moving dots and 50% randomly moving

dots is presented against a black background. A single dot consists of 3 pixels and the whole

cloud has a diameter of 250 (uniformly distributed) pixels. At the beginning of each trial

either a speed- or an accuracy-cue is presented for 1000 ms. The speed-accuracy tradeoff

(SAT) in this task is manipulated by pseudo-randomized presentation of the two different

cue-types. The speed cue instructs participants to adopt a liberal level of cautiousness,

responding as quickly as possible. The accuracy cue, however, instructs participants to adopt

a more conservative level of cautiousness, responding as accurate as possible. After the cue,

a fixation cross is presented at the center of the screen for 500 ms. Subsequently, the RDM

stimulus is shown until a response is made, but at most for 1500 ms. Responses have to be

made within this time window. Participants respond by pressing ‘a’ with their left index

finger when they perceive a leftward motion and ‘l’ with their right index finger when they

perceive a rightward motion. Immediately after the response participants receive feedback

Figure 1. Random Dot Motion paradigm with cues emphasizing speed (SN Dutch abbreviation for fast) and accuracy (AC Dutch abbreviation for accurate).

Note. Figure taken from Forstmann et al. 2010 and edited.


10

(400 ms) that is dependent on the trial-type. They see either “te traag” (Dutch for ‘too

slow’)/”fout” (Dutch for incorrect) or “op tijd” (Dutch for ‘in time’)/“goed” (Dutch for

correct). After 120 and 240 trials participants can choose to take a break of up to 45

seconds. The entire experiment takes approximately 20 minutes.

LBA model. The linear ballistic accumulator (LBA; Brown et al., 2008) model serves to

decompose the response time and accuracy measures into latent psychological processes.

The decision process of interest, here, is response caution, which can be quantified by

means of the LBA. We will apply the same parameter constraints as Forstmann and

colleagues (2010). In this design only one parameter - response threshold (b) - will be free to

vary with the speed vs. accuracy cue, while all other parameters (start point distribution (A),

drift rate for the response (v), variability of the drift rate (s) and nondecision time (t0)) will

be fixed.

Behavioral data analysis. The behavioral measure of interest is the LBA flexibility

parameter, assessing efficacy of changing response caution. It is assumed that “changes in

response caution originate from adjustments of response thresholds (Forstmann et al.,

2010; page 1516)”. Therefore, LBA flexibility is computed as the difference between the LBA

threshold estimates for the accuracy (baccuracy) and the speed (bspeed) conditions. We will fit

the LBA model to each participant's RT and accuracy measures on speed and accuracy trials

separately. The only parameter allowed to vary will be the response threshold b. The

resulting individual LBA flexibility estimates will be imported into R for the Bayesian

correlation test.

TOI analysis. Forstmann and colleagues (2010) reported a significant positive correlation

between LBA flexibility and tract strength of white matter (WM) fibers connecting right pre-

SMA and right striatum. We defined this WM fiber tract as our tract-of-interest (TOI). We

will use the mask/template that was used in the original study for the segmentation of our

confirmatory TOI.

Probabilistic tractography. We will limit our tractography to delineate tracts that the

authors found to significantly correlate with LBA flexibility. Hence, probabilistic tractography

will be performed only on fibers connecting right pre-SMA and right striatum. We will


11

perform the probabilistic tractography conform the protocol stated in the general methods

and analyses section (see above).


described above), we will extract tract strength values from all voxels contained in the

relevant TOI and average them. This will be done for every subject. These tract strength

measures are then corrected for age and gender using partial correlations. The corrected

mean tract strength measures will be imported into R (R Foundation for Statistical

Computing, http://www.R-project.org) software for the Bayesian correlation test.

Specifically, we will test for a positive correlation between LBA flexibility and WM tract

strength of fibers connecting right pre-SMA and right striatum.

Replication 3. Individuals with short percept durations in perceptual rivalry show

increased cortical thickness within superior parietal lobe (SPL) and postcentral gyrus,

increased gray matter volume within SPL, and increased fractional anisotropy of white

matter fibers underneath SPL.

Kanai, R., Bahrami, B., & Rees, G. (2010). Human parietal cortex structure predicts individual

differences in perceptual rivalry. Current Biology, 20(18), 1626-1630.

Bistable SFM task. We use the same ambiguous rotating structure-from-motion (SFM)

stimulus task as Kanai and colleagues (2010; see Figure 2).

Figure 2. Bistable Structure-From-Motion Stimulus.

Note. Figure taken from Kanai et al. 2010.

The task includes 8 trials and 1 practice trial. Participants are presented with an ambiguous

rotating sphere, consisting of 200 full white dots. The dots move sinusoidally at a constant

speed (151 deg/s). However, the rotation of the sphere can be perceived as either rightward

about:blank


12

or leftward. A red fixation dot is presented at the center of the screen and participants are

instructed to steadily fixate. The trial duration is 48 seconds. Participants report the

duration of their percept of the rotation direction (right or left) of the SFM stimulus by

holding the spatially compatible key (‘left arrow’ or ‘right arrow’ on a regular keyboard) with

their left or right index finger until the percept switches to the other direction. The entire

experiment takes approximately 10 minutes.

Behavioral data analysis. The behavioral measure of interest is percept duration, assessing

bistable perception. Besides computing the mean percept duration, we will also compute its

reciprocal, the switch rate, for each participant and import these values into R (R

Foundation for Statistical Computing, http://www.R-project.org) software for the Bayesian

correlation test.

ROI generation. Kanai and colleagues (2010) reported significant negative correlations

between percept duration and cortical thickness (CT) within bilateral superior parietal lobe

(SPL) and bilateral postcentral gyrus (PCG). For gray matter (GM) volume they reported

significant negative correlations with percept duration within right and left SPL. White

matter fractional anisotropy (FA) was negatively correlated with percept duration

underneath right and left SPL. We defined all these regions as our regions of interest (ROIs).

Dr. Kanai kindly provided us with the masks/templates that were used in the original study.

We will use these for the segmentation of our confirmatory ROIs.


described above), we will extract CT, GM, and FA values from all voxels contained in the

respective ROIs and average them. This will be done for every subject. These structural brain

measures are then corrected for age and gender using partial correlations. The corrected

mean CT, GM, and FA measures per ROI will be imported into R (R Foundation for Statistical


Specifically, we will test for negative correlations between percept duration and mean CT

within bilateral SPL and bilateral PCG. We will also test for negative correlations between

percept duration and mean GM volume within right and left SPL. Finally, we will test for

negative correlations between percept duration and mean FA underneath right and left SPL.

about:blank

about:blank


13

Replication 4. Individuals with long percept durations in perceptual rivalry show increased

gray matter volume within right anterior superior parietal lobe.

Kanai, R., Carmel, D., Bahrami, B., & Rees, G. (2011a). Structural and functional fractionation

of right superior parietal cortex in bistable perception. Current Biology, 21(3), R106-R107.

Bistable SFM task. We use the same ambiguous rotating structure-from-motion (SFM)

stimulus task as Kanai and colleagues (2011a; see Figure 2). For a detailed description of the

task please refer back to Replication 3.

Behavioral data analysis. The behavioral measure of interest is percept duration (and

switch rate), assessing bistable perception. We will compute the mean percept duration for

each participant and import these values into R (R Foundation for Statistical Computing,

http://www.R-project.org) software for the Bayesian correlation test.

ROI generation. Kanai and colleagues (2011a) reported a significant positive correlation

between percept duration and gray matter (GM) volume within right anterior superior

parietal lobe (aSPL). We defined this region as our region of interest (ROI). Dr. Kanai kindly

provided us with the mask/template that was used in the original study. We will use this

mask/template for the segmentation of our confirmatory ROI.


described above), we will extract GM values from all voxels contained in the ROI and

average them. This will be done for every subject. These GM measures are then corrected

for age and gender using partial correlations. The corrected mean GM measures will be

imported into R (R Foundation for Statistical Computing, http://www.R-project.org)

software for the Bayesian correlation test. Specifically, we will test for a positive correlation

between percept duration and mean GM volume within right aSPL.

Replication 5. Individuals with a small difference in reaction times between trials with

incongruent and congruent stimuli (i.e., good executive control) show increased cortical

thickness within left caudal anterior cingulate cortex, left superior temporal lobe, and

right middle temporal lobe. Individuals with a small difference in reaction times between

about:blank

about:blank


14

trials with no cue and a central cue (i.e., good alerting ability) show increased cortical

thickness within medial and lateral aspects of the left superior temporal lobe.

Westlye, L. T., Grydeland, H., Walhovd, K. B., & Fjell, A. M. (2011). Associations between

regional cortical thickness and attentional networks as measured by the attention network

test. Cerebral Cortex, 21(2), 345-356.

Attention Network Test. We use the same Attention Network Test (ANT) as Westlye and

colleagues (2011; downloaded from Dr. Jin Fan’s website

www.sacklerinstitute.org/users/jin.fan; see Figure 3). The task includes 2 runs of 96 trials

and 20 practice trials. Each trial begins with the presentation of a fixation cross in the center

of the screen for variable durations (400, 800, 1200, or 1600 ms). Subsequently, one of

three cues is presented for 100 ms: (1) no cue, (2) center cue (*, replacing fixation cross), or

(3) spatial cue (*, above or below fixation cross). Then the target is presented for a

maximum duration of 1700 ms (until a response is made). The target is an arrow in the

centre of a row of 5 arrows, presented below or above the fixation cross. The flanking

arrows can be (1) two congruent arrows (pointing in the same direction as the target), (2)

two incongruent arrows (pointing in the opposite direction of the target), or (3) two lines on

each side of the target (neutral). Participants are instructed to report the direction (left or

right) of the target arrow by pressing the spatially compatible key (‘left mouse button’ and

‘right mouse button’) with their left or right thumb. The entire experiment takes

approximately 15 minutes.

Figure 3. Attentional Network Test Paradigm.

Behavioral data analysis. The behavioral measures of interest are executive control (EC)

and alerting (A) network scores, assessing the executive control and the alerting

http://www.sacklerinstitute.org/users/jin.fan


15

components of attention, respectively. We will apply the same processing steps as

described by Westlye and colleagues (2011) before we compute the two network scores:

“To remove outliers, all RTs > 1500 ms and < 200 ms were removed (…). Next, since error responses

are assumed to originate from a different RT distribution than correct responses, we only analyzed

correct responses. Also, because responses following erroneous responses typically are slower than

responses following correct responses (posterror slowing), we also removed responses following

erroneous responses. Since RTs are not normally distributed, we used median RT per condition as

raw scores for each subject. (…). (page 348).”

However, we will not adjust the component scores with the baseline RT in order to control

for an effect of age on RT, because our participants form a homogenous age group

(Psychology freshmen).

Based on median RT the EC score will be computed as follows:

EC = [RTincongruent - RTcongruent] / RTcongruent

Based on median RT the A score will be computed as follows:

A = [RTno cue – RTcenter cue] / RTcenter cue

For each participant the resulting scores will be imported into R (R Foundation for Statistical


ROI generation. For their subsample of young participants, Westlye and colleagues (2011)

reported significant negative correlations between EC scores and cortical thickness (CT)

within left caudal anterior cingulate cortex (ACC), left superior temporal lobe (STL), and right

middle temporal lobe (MTL). The A scores showed a significant negative correlation with CT

within left superior parietal lobe (SPL). We defined all these regions as our regions of

interest (ROIs). Dr. Westlye kindly provided us with the masks/templates (i.e., labels used in

FreeSurfer (http://surfer.nmr.mgh.harvard.edu) software) that were used in the original

study. We will use these for the segmentation of our confirmatory ROIs.


described above), we will extract CT values from all voxels contained in the ROIs and

average them. This will be done for every subject. These CT measures are then corrected for

age and gender using partial correlations. The corrected mean CT measures will be imported

about:blank


16

into R (R Foundation for Statistical Computing, http://www.R-project.org) software for the

Bayesian correlation test. Specifically, we will test for negative correlations between EC

scores and mean CT within left caudal ACC, left STL and right MTL. Furthermore, we will also

test for a negative correlation between A scores and mean CT within left SPL.

Replication 6. Young individuals with high scores on the Social Network Index (i.e., large

social networks) show increased gray matter volume within bilateral amygdala and

increased cortical thickness within right subgenual anterior cingulate cortex, left caudal

superior frontal gyrus and left caudal inferior temporal sulcus.

Bickart, K. C., Wright, C. I., Dautoff, R. J., Dickerson, B. C., & Barrett, L. F. (2011). Amygdala

volume and social network size in humans. Nature Neuroscience, 14(2), 163-164.

Social Network Index questionnaire and procedure. We use a Dutch version of the Social

Network Index (SNI) questionnaire. The questionnaire has 12 items, measuring aspects of

social cognition: Social network diversity (SND), social network size (SNS) and social network

complexity (SNC). Administration time is 10 minutes.

Behavioral data analysis. The behavioral measures of interest are SNS (i.e., the total

number of people with whom the respondent has regular contact) and SNC (i.e., the

number of different groups that these contacts belong to). Thus, for each participant the

resulting SNS and SNC scores will be imported into R (R Foundation for Statistical


ROI generation. For their subsample of young participants, Bickart and colleagues (2011)

reported significant positive correlations between SNS/SNC (similar results) and gray matter

(GM) volume within left and right amygdala. Furthermore they reported significant positive

correlations between SNS/SNC (similar results) and cortical thickness (CT) within right

subgenual anterior cingulate cortex (sgACC), left caudal superior frontal gyrus (cSFG), and

left caudal inferior temporal sulcus (cITS). We defined all these regions as our regions of

interest (ROIs). Prof. Feldman Barrett and Kevin Bickart kindly provided us with the

masks/templates that were used in the original study. We will use these for the

segmentation of our confirmatory ROIs.


about:blank

about:blank


17

described above), we will extract GM and CT values from all voxels contained in the

respective ROIs and average them. This will be done for every subject. These GM and CT

measures will then be corrected for age and gender using partial correlations. The GM

measures will additionally be corrected for total intracranial volume. The corrected mean

GM and CT measures will be imported into R (R Foundation for Statistical Computing,

http://www.R-project.org) software for the Bayesian correlation test. Specifically, we will

test for positive correlations between SNS/SNC and mean GM volume within left and right

amygdala as well as mean CT within right sgACC, left cSFG, and left cITS.

Replication 7. Individuals with a large number of friends on Facebook (i.e., large online

social network) show increased gray matter volume within the following regions: left

middle temporal gyrus, right posterior superior temporal sulcus, right entorhinal cortex,

and bilateral amygdala. Individuals with high scores on the Social Network Size

Questionnaire (i.e., large real-world social network) show increased gray matter volume

within right amygdala.

Kanai, R., Bahrami, B., Roylance, R., & Rees, G. (2012). Online social network size is reflected

in human brain structure. Proceedings of the Royal Society Biological sciences, 279(1732),

1327-1334.

Social Network Size Questionnaire and procedure. Participants fill out a Dutch version of

the Social Network Size questionnaire (Stileman & Bates, 2007; see Appendix A). This

questionnaire consists of 9 items. One of its items is: “How many friends do you have on

‘Facebook’?”. We ask participants to make a note of the number of friends they have on

‘Facebook’ or an alternative comparable social network site such as ‘myspace’ or the Dutch

‘Hyves’ and bring it to the test session. The administration time is approximately 10

minutes.

Behavioral data analysis. The behavioral measures of interest are online social network size

(i.e., the number of Facebook friends (FBN)) and real-world social network size. Subjects’

answers to the 9 subquestions contained in this questionnaire will be square-root

transformed to correct for skewness. We will compute the FBN as the square root of

subject’s answer to the question: “How many friends do you have on ‘Facebook’?”. A

normalized real-world social network size score will be computed per participant by

about:blank


18

averaging the z-scores for the questionnaire items 1, 2, 4, 5, 6, 8 and 9 after skewness

correction. Thus, for each participant an online social network size score and a real-world

social network size score will be imported into R (R Foundation for Statistical Computing,

http://www.R-project.org) software for the Bayesian correlation test.

ROI generation. Kanai and colleagues (2012) reported significant positive correlations

between online social network size and gray matter (GM) volume within left middle

temporal gyrus (MTG), right superior temporal sulcus (STS), right entorhinal cortex (EC), and

bilateral amygdala. Real-world social network size was positively correlated with GM only

within right amygdala. We defined all these regions as our regions of interest (ROIs). Dr.

Kanai kindly provided us with the masks/templates that were used in the original study. We

will use these for the segmentation of our confirmatory ROIs.


described above), we will extract gray matter (GM) values from all voxels contained in the

ROIs and average them. This will be done for every subject. These GM measures will then be

corrected for age, gender and total gray matter volume. The corrected mean GM measures

will be imported into R (R Foundation for Statistical Computing, http://www.R-project.org)

software for the Bayesian correlation test. Specifically, we will test for positive correlations

between FBN and mean GM volume within left MTG, right STS, right EC, and bilateral

amygdala. Furthermore, we will test for a positive correlation between real-world network

size scores and mean GM volume within right amygdala. Since our Bayesian correlation test

allows us to quantify evidence in favor of a null-effect, we will also test for the absence of

positive correlations between real-world social network size and mean GM volume within

left MTG, right STS, right EC, and left amygdala.

Replication 8. Individuals with high scores on the Cognitive Failure Questionnaire (i.e.,

high distractibility) show increased gray matter volume within left superior parietal lobule

and decreased gray matter volume within left mid prefrontal cortex.

Kanai, R., Dong, M. Y., Bahrami, B., & Rees, G. (2011d). Distractibility in daily life is reflected

in the structure and function of human parietal cortex. The Journal of Neuroscience, 31(18),

6620-6626.

about:blank

about:blank


19

Cognitive Failures Questionnaire and procedure. Participants fill out a Dutch version of the

Cognitive Failures Questionnaire (CFQ, Broadbent et al., 1982; see Appendix A). The

administration time is approximately 5 minutes.

Behavioral data analysis. The behavioral measure of interest is distractibility as assessed by

the CFQ. As in Kanai et al. (2011d), we will quantify distractibility by computing the standard

loadings derived from a previous factor analysis (Wallace et al., 2002). Specifically, we will

use the following 9 items: 1, 2, 3, 4, 15, 19, 21, 22, and 25. Scores on these items will be

imported into R (R Foundation for Statistical Computing, http://www.R-project.org)

software for the Bayesian correlation test.

ROI generation. Kanai and colleagues (2011d) reported a significant positive correlation

between CFQ scores and gray matter (GM) volume within left superior parietal lobe (SPL).

Furthermore, the authors reported a weak negative correlation between CFQ scores and

GM volume within left mid prefrontal cortex (mPFC). We defined these regions as our

regions of interest (ROIs). Dr. Kanai kindly provided us with the masks/templates that were

used in the original study. We will use these for the segmentation of our confirmatory ROIs.



respective ROIs and average them. This will be done for every subject. These GM measures

are then corrected for age, gender and total gray matter volume using partial correlations.

The corrected mean GM measures will be imported into R (R Foundation for Statistical


Specifically, we will test for a positive correlation between CFQ scores and mean GM

volumes within left SPL and for a negative correlation within left mPFC.

Replication 9. Individuals with high scores on the Political Orientation Questionnaire (i.e.,

conservative) show increased gray matter volume within right amygdala and left insula,

and decreased gray matter volume within right entorhinal cortex. Individuals with low

scores on the Political Orientation Questionnaire (i.e., liberal) show increased gray matter

volume within anterior cingulate cortex.

about:blank

about:blank


20

Kanai, R., Feilden, T., Firth, C., & Rees, G. (2011c). Political orientations are correlated with

brain structure in young adults. Current Biology, 21, 677-680.

Political Orientation Questionnaire and procedure. Participants fill out a Dutch version of

the Political Orientation Questionnaire (POQ) used by Kanai and colleagues (2011c; see

Appendix A). The POQ consist of a five-point scale: (1) very liberal, (2) liberal, (3) middel-of-

the-road, (4) conservative, (5) very conservative. The administration time is approximately 1

minute.

Behavioral data analysis. The behavioral measure of interest is political orientation. For

each participant a political orientation score will be imported into R (R Foundation for

Statistical Computing, http://www.R-project.org) software for the Bayesian correlation test.

ROI generation. Kanai and colleagues (2011c) reported significant positive correlations

between high POQ scores (i.e., conservatism) and gray matter (GM) volume within right

amygdala and left insula. GM volume was negatively correlated with conservatism within

right entorhinal cortex (EC). Furthermore, the authors reported a significant negative

correlation between low POQ scores (i.e., liberalism) and GM volume within anterior

cingulate cortex (ACC). We defined all these regions as our regions of interest (ROIs). Dr.

Kanai kindly provided us with the masks/templates that were used in the original study. We

will use these for the segmentation of our confirmatory ROIs.



respective ROIs and average them. This will be done for every subject. These GM measures

are then corrected for age, gender, and total gray matter volume using partial correlations.

The corrected mean GM measures will be imported into R (R Foundation for Statistical


Specifically, we will test for positive correlations between conservatism and mean GM

volume within right amygdala and left insula. We will also test for a negative correlation

between conservatism and mean GM volume within right entorhinal cortex. Furthermore,

we will test for a negative correlation between liberalism and mean GM volume within ACC.

Finally, we will attempt to replicate the absence of a significant positive correlation between

conservatism and mean GM volume within left amygdala.

about:blank

about:blank


21


22

Appendix A:

Translated questionnaires used in the present replication study.

1. Dutch Translation of Behavioral Inhibition System/Behavioral Activation System

(Carver et al., 1994)

Op de volgende bladzijden vindt je een aantal beweringen. De bedoeling is dat je deze beweringen doorleest en dat je nagaat of zij van toepassing zijn.

Naast elke bewering staan vier antwoordmogelijkheden die variëren van "helemaal mee eens" tot "helemaal mee oneens". Het is de bedoeling dat je telkens met een kruisje in één van de hokjes aangeeft in hoeverre een bewering op jou van toepassing is.

Laat geen vraag onbeantwoord en beperk je tot de gegeven antwoordmogelijkheden.

Neem je tijd, maar denk niet al te lang na over een vraag.

1. Als ik denk dat er iets onprettigs gaat gebeuren, raak ik meestal behoorlijk "opgefokt".

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens

0 helemaal mee oneens

2. Ik ben bezorgd om het maken van fouten.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


3. Als ik iets wil, ga ik er meestal helemaal voor.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens



23

4. Vaak doe ik dingen om geen andere reden dan dat het wel eens leuk zou kunnen zijn.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


5. Kritiek of een standje raken mij behoorlijk.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


6. Als ik iets krijg wat ik wil, voel ik me opgewonden en opgeladen.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


7. Ik doe een hoop moeite om dingen die ik wil te krijgen.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


8. Ik verlang sterk naar spanning en nieuwe sensaties.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


24


9. Ik voel me behoorlijk overstuur als ik denk of weet dat iemand boos op me is.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


10. Ik ben altijd bereid iets nieuws te proberen als ik denk dat het leuk zal zijn.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


11. Als ik iets goed doe, wil ik er graag mee doorgaan.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


12. Zelfs als mij iets ergs staat te gebeuren, ervaar ik zelden angst of nervositeit.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


13. Ik handel vaak zoals het moment me ingeeft.

0 helemaal mee eens


25

0 beetje mee eens

0 beetje mee oneens


14. Als ik een kans zie iets te krijgen wat ik wil, ga ik er meteen op af.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


15. Als mij goede dingen overkomen, raakt dat me sterk.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


16. Ik voel me bezorgd als ik denk dat ik slecht heb gepresteerd op iets.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


17. Ik zou het spannend vinden een wedstrijd te winnen.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens



26

18. Vergeleken met mijn vrienden heb ik erg weinig angsten.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


19. Als ik een mogelijkheid zie iets te krijgen wat ik leuk vind, word ik direct opgewonden.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens


20. Als ik ergens werk van maak, gooi ik ook mijn volle gewicht er tegenaan.

0 helemaal mee eens

0 beetje mee eens

0 beetje mee oneens



27

2. Dutch Translation of Social Network Index (Cohen et al., 1997)

Instructies: Deze vragenlijst gaat over met hoeveel mensen u regelmatig afspreekt of praat, inclusief familie, vrienden, collega's, buren, etc. Lees en beantwoord elke vraag nauwkeurig. Beantwoord subvragen waar nodig.

1. Welk van onderstaande alternatieven beschrijft je huwelijkse staat het best?

.... (1) Op dit moment getrouwd en samenwonend, of samenwonend met een vaste relatie

.... (2) Nooit getrouwd geweest, en nooit samengewoond met een vaste relatie

.... (3) Uit elkaar

.... (4) Gescheiden, of voorheen samengewoond met iemand in een vaste relatie

.... (5) Weduwe/weduwnaar

2. Hoeveel kinderen heb je? (Als je geen kinderen hebt, schrijf dan '0' op en ga verder met vraag 3.)

....

2a. Met hoeveel van je kinderen heb je minstens eens per twee weken contact?

....

3. Zijn je vader en moeder nog in leven? (Als je beide ouders overleden zijn, zet dan een kruisje bij '0' en ga verder met vraag 4.)

.... (0) Geen van beide

.... (1) Alleen moeder nog in leven

.... (2) Alleen vader nog in leven

.... (3) Beide nog in leven

3a. Zie je of spreek je je vader en moeder minstens eens per twee weken?


.... (1) Alleen moeder

.... (2) Alleen vader

.... (3) Beide


28

4. Leven je schoonvader en je schoonmoeder nog (of de ouders van je partner)? (Als je geen schoonouders hebt, zet dan een kruisje bij 'Niet van toepassing' en ga verder met vraag 5.)


.... (1) Alleen schoonmoeder leeft nog

.... (2) Alleen schoonvader leeft nog

.... (3) Beide leven nog

.... (4) Niet van toepassing

4a. Zie je of spreek je je schoonouders minstens eens per twee weken?


.... (1) Alleen schoonmoeder

.... (2) Alleen schoonvader

.... (3) Beide

5. Aan hoeveel familieleden (anders dan uw echtgeno(o)t(e), ouders, en kinderen) voel je je gehecht? (Als dit er geen zijn, schrijf dan op '0', en ga verder met vraag 6.)

....

5a. Hoeveel van deze familieleden zie je of spreek je minstens eens per twee weken?

....

6. Hoeveel goede vrienden heb je? (met goede vrienden bedoelen we hier mensen bij wie u zich gemakkelijk voelt, met wie u over persoonlijke zaken kunt praten, en die u om hulp kunt vragen)

....

6a. Hoeveel van deze vrienden zie je of spreek je minstens eens per twee weken?

....

7. Ben je lid van een religieuze groep? (bv. een kerkgenootschap) Als dit niet het geval is, zet dan een kruisje bij 'nee' en ga verder met vraag 8.

... (1) Ja

... (2) Nee


29

7a. Hoeveel leden van je religieuze groep spreek je minstens eens per twee weken? (Inclusief gesprekken rondom bijeenkomsten en diensten).

....

8. Volgt u regelmatig onderwijs (via een school, universiteit, technische training, of volwassenen onderwijs)? Zo niet, zet dan een kruisje bij 'nee' en ga door met vraag 9.

... (1) Ja

... (2) Nee

8a. Hoeveel medestudenten of docenten spreek je minstens eens per twee weken? (Inclusief gesprekken rondom de lessen)

....

9. Werkt u op dit moment voltijds of in deeltijd? (Als dit niet het geval is, zet dan een kruisje bij 'nee' en ga door met vraag 10.)

.... (0) nee

.... (1) ja, in mijn eigen bedrijf/als zelfstandig ondernemer

.... (2) ja, in loondienst

9a. Over hoeveel mensen heeft u de supervisie?

....

9b. Hoeveel collega's (uitgezonderd hen die u superviseert) spreekt u minstens eens per twee weken?

....

10. Hoeveel van je buren bezoek je of spreek je minstens eens per twee weken?

....

11. Verricht je op dit moment regelmatig vrijwilligerswerk? (Zo nee, zet dan een kruisje bij 'nee' en ga verder met vraag 12.)

... (1) Ja

... (2) Nee

11a. Hoeveel mensen die betrokken zijn bij dit vrijwilligerswerk spreek je minstens eens per twee weken over zaken die te maken hebben met het vrijwilligerswerk?

....


30

12. Hoor je bij groepen waarin je met een of meerdere leden minstens eens per twee weken over groepsgerelateerde zaken praat? Voorbeelden van zulke groepen zijn onder andere gezelligheidsverenigingen; hobbyverenigingen; vakbonden; commerciële groepen; beroepsorganisaties; groepen die te maken hebben met kinderen, zoals ouderverenigingen op scholen of scouting; groepen die te maken hebben met de gemeenschap waarin je woont; etc. (Als je niet bij zulke groepen hoort, zet dan een kruisje bij 'nee' en sla het volgende deel van de vragenlijst over.)

... (1) Ja

... (2) Nee

Bedenk in welke van deze groepen je minstens eens per twee weken met een groepslid praat, en geef de volgende informatie voor elke groep: De naam of het soort groep, en het totale aantal groepsleden in die groep met wie je minstens eens per twee weken praat.

Naam/soort groep Aantal groepsleden


31

3. Dutch Translation of Social Network Size Questionnaire (Stileman & Bates, 2007)

Hoeveel mensen waren er in totaal aanwezig op het feest voor je 18de of 21ste verjaardag? .... Als je nu een feestje zou geven, hoeveel mensen zou je dan uitnodigen? .... Wat is het totaal aantal vrienden in de contactlijst van je telefoon? .... Schrijf de namen op van alle mensen die je een sms-bericht zou sturen bij een feestelijke gebeurtenis (bijv. een verjaardag, Kerstmis, nieuwe baan, goed examenresultaat, etc.). Hoeveel mensen zijn dit in totaal? .... Schrijf de namen op van alle mensen in de contactlijst van je telefoon die je zou kunnen ontmoeten voor een praatje in een besloten groep van twee tot vier personen. Hoeveel mensen zijn dit in totaal? .... Met hoeveel vrienden uit je schooltijd zou je nu nog vriendschappelijk gesprek kunnen voeren? .... Hoeveel vrienden heb je op Facebook? .... Hoeveel vrienden heb je buiten de universiteit? .... Schrijf de namen op van alle mensen waarvan je vindt dat je ze om een gunst kan vragen in de verwachting dat die ook wordt verleend. Hoeveel mensen zijn dit in totaal? ....


32

4. Dutch Translation of Cognitive Failures Questionnaire (adapted from Broadbent et al.,

1982)


33

5. Dutch Translation of Political Orientation Questionnaire (adapted from Kanai et al., 2011c)

Geef alstublieft je politieke voorkeur aan door een van de onderstaande alternatieven te omcirkelen: (1) erg progressief (2) progressief (3) niet progressief maar ook niet conservatief (4) conservatief (5) erg conservatief


34

References

Bickart, K. C., Wright, C. I., Dautoff, R. J., Dickerson, B. C., & Barrett, L. F. (2011). Amygdala

volume and social network size in humans. Nature Neuroscience, 14(2), 163-164.

Brown, S. D., & Heathcote, A. (2008). The simplest complete model of choice response time:

linear ballistic accumulation. Cognitive Psychology, 57(3), 153-78.

Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and

affective responses to impending reward and punishment: The BIS/BAS scales. Journal

of Personality and Social Psychology, 67, 319-333.

Cohen, S., Doyle, W. J., Skoner, D. P., Rabin, B. S., & Gwaltney, J. M., Jr. (1997). Social ties and

susceptibility to the common cold. Journal of the American Medical Association, 277,

1940-1944.

Davis, M. H. (1980). A multi-dimensional approach to individual differences in empathy.

JCAS Catalog of Selected Documents in Psychology, 75, 989-1015.

De Corte, K., Buysse, A., Verhofstadt, L. L., Roeyers, H., Ponnet, K., & Davis, M. H. (2007).

Measuring empathic tendencies: Reliability and validity of the Dutch version of the

interpersonal reactivity index. Psychologica Belgica, 47(4), 235-260.

Edwards, W., Lindman, H., & Savage, L. J. (1963). Bayesian statistical inference for

psychological research. Psychological Review, 70, 193–242.

Fan, J., McCandliss, B. D., Sommer, T., Raz, A., & Posner, M. I. (2002). Testing the efficiency

and independence of attentional networks. Journal of Cognitive Neuroscience, 14(3),

340-347.

Forstmann, B. U., Anwander, A., Schäfer, A., Neumann, J., Brown, S., Wagenmakers, E.-J.,

Bogacz, R., et al. (2010). Cortico-striatal connections predict control over speed and

accuracy in perceptual decision making. Proceedings of the National Academy of

Sciences of the United States of America, 107(36), 15916-15920.

Gold, J. I., & Shadlen, M. N. (2007). The neural basis of decision making. Annual Review of

Neuroscience, 30, 535-74.

Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak, R. S.


35

(2001). A voxel-based morphometric study of aging in 465 normal adult human brains.

NeuroImage, 14, 21-36.

Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping the

moral domain. Journal of Personality and Social Psychology, 101, 366-85.

Hammers, A., Allom, R., Koepp, M. J., Free, S. L., Myers, R., Lemieux, L., Mitchell, T. N., et al.

(2003). Three-dimensional maximum probability atlas of the human brain, with

particular reference to the temporal lobe. Human Brain Mapping, 19(4), 224-47.

Hoekstra, H. A., Ormel, H., & De Fruyt, F. (1996). NEO-PI-R personality inventory Dutch

manual. Lisse, The Netherlands: Swets & Zeitlinger.

Hoijtink, H.J.A., Klugkist, I.G. & Boelen, P.A. (2008). An introduction to Bayesian evaluation of

informative hypotheses. In H.J.A. Hoijtink, I.G Klugkist & P.A. Boelen (Eds.), Bayesian

Evaluation of Informative Hypotheses (pp. 1-6). New York: Springer.

Jeffreys, H. (1961). Theory of Probability, 3rd ed. Oxford Classic Texts in the Physical Sciences.

Oxford Univ. Press, Oxford.

Kanai, R., Bahrami, B., & Rees, G. (2010). Human parietal cortex structure predicts individual

differences in perceptual rivalry. Current Biology, 20(18), 1626-1630.

Kanai, R., Carmel, D., Bahrami, B., & Rees, G. (2011a). Structural and functional fractionation

of right superior parietal cortex in bistable perception. Current Biology, 21(3), R106-

R107.

Kanai, R., & Rees, G. (2011b). The structural basis of inter-individual differences in human

behaviour and cognition. Nature Reviews Neuroscience, 12(April), 231-242.

Kanai, R., Feilden, T., Firth, C., & Rees, G. (2011c). Political orientations are correlated with

brain structure in young adults. Current Biology, 21, 677-680.

Kanai, R., Dong, M. Y., Bahrami, B., & Rees, G. (2011d). Distractibility in daily life is reflected

in the structure and function of human parietal cortex. The Journal of Neuroscience,

31(18), 6620-6626.

Kanai, R., Bahrami, B., Roylance, R., & Rees, G. (2012). Online social network size is reflected

in human brain structure. Proceedings of the Royal Society Biological sciences,


36

279(1732), 1327-1334.

Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity and Inhibitory Control.

Psychological Science, 8(1), 60-64.

Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E.,

Watkins, K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., & Behrens, T.E.J. (2006).

Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data.

NeuroImage, (31), 1487-1505.

Stileman, E. & Bates, T. 2007 Construction of the Social Network Score (SNS) Questionnaire

for undergraduate students, and an examination of the pre-requisites for large social

networks in humans? Unpublished undergraduate thesis.

See http://hdl.handle.net/1842/2553.

Wagenmakers, E.-J., Lodewyckx, T., Kuriyal, H., and Grasman, R. (2010). Bayesian hypothesis

testing for psychologists: A tutorial on the Savage-Dickey method. Cognitive

Psychology, 60, 158-189.

Wagenmakers, E.-J., Wetzels, R., Borsboom, D., & van der Maas, H. L. J. (2011). Why

psychologists must change the way they analyze their data: the case of psi: comment

on Bem (2011). Journal of Personality and Social Psychology, 100(3), 426-432.

Wallace J. C., Kass, S. J., & Stanny, C. J. (2002). Cognitive failures questionnaire revisited:

Correlates and dimensions. Journal of General Psychology, 129, 238 –256.

Westlye, L. T., Grydeland, H., Walhovd, K. B., & Fjell, A. M. (2011). Associations between

regional cortical thickness and attentional networks as measured by the attention

network test. Cerebral Cortex, 21(2), 345-356.

Xu, J., Kober, H., Caroll, K. M., Rounsaville, B. J., Pearlson, G. D., & Potenza, M. N. (2012).

White matter integrity and behavioral activation in healthy subjects. Human Brain

Mapping, 33, 994-1002.

http://www.ejwagenmakers.com/2010/WagenmakersEtAlCogPsy2010.pdf

http://www.ejwagenmakers.com/2010/WagenmakersEtAlCogPsy2010.pdf

Confirmatory Replication Study - Methods and Analyses

Documents

bayesian hypothesis

mri data acquisition

skeletonized

t1 images

fa images

probabilistic

sided test

seed mask