Internet addiction and functional brain networks: task-related fMRI study

Abstract

A common brain-related feature of addictions is the altered function of higher-order brain networks. Growing evidence suggests that Internet-related addictions are also associated with breakdown of functional brain networks. Taking into consideration the limited number of studies used in previous studies in Internet addiction (IA), our aim was to investigate the functional correlates of IA in the default mode network (DMN) and in the inhibitory control network (ICN). To observe these relationships, task-related fMRI responses to verbal Stroop and non-verbal Stroop-like tasks were measured in 60 healthy university students. The Problematic Internet Use Questionnaire (PIUQ) was used to assess IA. We found significant deactivations in areas related to the DMN (precuneus, posterior cingulate gyrus) and these areas were negatively correlated with PIUQ during incongruent stimuli. In Stroop task the incongruent_minus_congruent contrast showed positive correlation with PIUQ in areas related to the ICN (left inferior frontal gyrus, left frontal pole, left central opercular, left frontal opercular, left frontal orbital and left insular cortex). Altered DMN might explain some comorbid symptoms and might predict treatment outcomes, while altered ICN may be the reason for having difficulties in stopping and controlling overuse.

Full text

1
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
Internet addiction and functional
brain networks: task-related fMRI
study
Gergely Darnai1,2,3*, Gábor Perlaki3,4,5, András N. Zsidó1, Orsolya Inhóf1, Gergely Orsi3,4,5,
Réka Horváth2, Szilvia Anett Nagy3,4,5,6, Beatrix Lábadi1, Dalma Tényi2, Norbert Kovács2,3,
Tamás Dóczi3,5, Zsolt Demetrovics
7 & József Janszky2,3
A common brain-related feature of addictions is the altered function of higher-order brain networks.
Growing evidence suggests that Internet-related addictions are also associated with breakdown of
functional brain networks. Taking into consideration the limited number of studies used in previous
studies in Internet addiction (IA), our aim was to investigate the functional correlates of IA in the default
mode network (DMN) and in the inhibitory control network (ICN). To observe these relationships,
task-related fMRI responses to verbal Stroop and non-verbal Stroop-like tasks were measured in 60
healthy university students. The Problematic Internet Use Questionnaire (PIUQ) was used to assess IA.
We found signicant deactivations in areas related to the DMN (precuneus, posterior cingulate gyrus)
and these areas were negatively correlated with PIUQ during incongruent stimuli. In Stroop task the
incongruent_minus_congruent contrast showed positive correlation with PIUQ in areas related to
the ICN (left inferior frontal gyrus, left frontal pole, left central opercular, left frontal opercular, left
frontal orbital and left insular cortex). Altered DMN might explain some comorbid symptoms and might
predict treatment outcomes, while altered ICN may be the reason for having diculties in stopping and
controlling overuse.
Internet – along with new technologies – has improved many aspects of our lives and it is now essential part of
the everyday routine, including professional and social functioning. e benets that Internet brought into our
life are multiple, however, excessive use can contribute to various psychological and medical problems, such as
depression1, anxiety2, body image disturbance3, sleeplessness4, and poor dietary behavior3. Problematic Internet
use is considered as a relatively new, fast growing behavioral addiction5 that has a potential threat to public
health6.
ough Internet addiction (IA) is not considered as a distinct mental disorder a more specic form, problem-
atic video gaming, operationalized as ‘Internet gaming disorder’ (IGD), was included in Section 3 (‘Emerging
Measures and Models’) of DSM-5, as a condition warranting further study7. It is important to note that the
11th revision of the International Classication of Diseases (ICD-11) also includes gaming disorder in section
“Disorders Due to Substance Use or Addictive Behaviours”8. Several authors claim that it is important to dis-
tinguish between IA and IGD. E.g. Montag and colleagues investigated general Internet addiction and specic
forms of Internet addiction (incl. video gaming, shopping, social network and pornography) in three European
and one Asian country. ey found that only online social network showed constant correlation patterns with
generalized IA, other specic forms must be distinguished from IA during scientic investigations9. Over the
last few years, researchers have increased eorts to investigate brain-related alterations to understand the phe-
nomena deeper. ese neuroimaging studies – utilizing mainly structural and functional magnetic resonance
imaging (MRI) – identied abnormalities in frontal brain regions (orbitofrontal and prefrontal cortex) and the
brain’s reward system (putamen, nucleus accumbens) that play crucial role in associative learning10 and cognitive
control11,12. Moreover, IA shares several aspects of substance addiction, such as obsessive thinking about the
1Institute of Psychology, University of Pécs, Pécs, Hungary. 2Department of Neurology, University of Pécs, Medical
School, Pécs, Hungary. 3MTA-PTE Clinical Neuroscience MR Research Group, Pécs, Hungary. 4Pécs Diagnostic
Centre, Pécs, Hungary. 5Department of Neurosurgery, University of Pécs, Medical School, Pécs, Hungary. 6MTA-PTE
Stress Neurobiology Research Group, Pécs, Hungary. 7Institute of Psychology, Eötvös Loránd University, Budapest,
Hungary. *email: darnai.gergely@pte.hu
OPEN
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
substance/Internet (daydreaming, rumination, and fantasizing), neglecting everyday activities, social life and
essential needs and diculties in controlling the use5,13.
Another common brain-related feature of substance and behavioral addictions is the altered function of
higher-order brain networks. Growing evidence suggests that addictions are not only associated with structural
and functional breakdown of isolated regions but rather with system-level alterations between brain regions14.
Functional brain imaging data have revealed that the human brain is topologically organized into a set of coherent
spatio-temporal networks, such as default-mode network (DMN)15. e DMN was rst mentioned by Shulman
in 199716 who noted several brain areas in the cerebral cortex that constantly decreased their activity while per-
forming highly demanding tasks. It can be divided into two main subdivisions: the medial prefrontal cortex, and
the posterior cingulate cortex with the nearby precuneus and lateral parietal cortex17. e DMN-related func-
tional MRI (fMRI) studies can be divided into two types of design. In the resting-state experiments, the subjects
lie passively (with closed eyes or with eyes focusing on a xation cross) without any tasks during scanning. In
task-related experiments, fMRI data are acquired during a certain cognitive task and researchers usually focus on
deactivations in the brain18. However, there are some studies that revealed task-induced activations in the DMN,
e.g. when internally directed/self-related cognition is required19.
Several studies revealed altered DMN in addictions. ese researches focused primarily on gambling dis-
order20 and Internet gaming disorder15,20 or substance addictions, such as heroin21, alcohol22, nicotine23, and
cannabis24. It was also demonstrated that functional connectivity within DMN may predict successful quitting25,
the intensity of withdrawal-induced craving26 and the degree of cognitive decline27 in addictions. ese observa-
tions suggest that dysfunctions in the DMN play important role in the pathogenesis and persistence of addictive
disorders. To our knowledge the only study focusing on DMN in adult Internet addicts was conducted by Li and
colleagues28. ey assessed grey matter volumetry and functional connectivity to investigate brain alterations in
healthy young adults with an IA tendency. ey found altered relationship between the dorsolateral prefrontal
cortex (as key node of the cognitive control network) and the anterior cingulate/prefrontal cortices (as key nodes
of the DMN).
Impaired inhibitory control is another important feature of addictions (including IA)29,30. Pre-existing inhib-
itory control problems may increase vulnerability to develop addictive disorders, and may serve as a risk factor
for their maintenance31. Dong et al. tried to identify neural correlates of response inhibition in IA. ey used
Stroop-related fMRI30 and event-related brain potential29 techniques and showed that cingulate and medial fron-
tal cortices are impaired in IA. ere are some practical and theoretical dierences in our study, compared with
that one conducted by Dong and colleagues. Firstly, they investigated only males with relatively low sample size
(24 participants in total). Secondly, we claim, that in the lack of well-established diagnostic criteria and clear
cuto points, continuous measure of IA is recommended. irdly, since Stroop task is considered as a highly
demanding task, in our study we focused on DMN activation as well (using dierent contrast in the higher level
analysis). Congdon et al.32 suggested that brain areas related to response inhibition ability should be considered as
elements of the “inhibitory control network” (ICN). According to their study, the inferior frontal and medial fron-
tal gyri, the opercular, insular, orbital posterior cingulate and posterior parietal cortices are involved in the ICN.
Taking into consideration the limited number of studies and methods used in previous functional studies in
IA, our aim was to investigate the functional correlates of IA in the DMN and in the ICN. To observe these rela-
tionships, task-related fMRI responses to verbal Stroop and non-verbal Stroop-like tasks were investigated. We
hypothesized that blood-oxygen-level dependent (BOLD) signal changes in the regions of the DMN and ICN are
correlated with IA scores. In the lack of well-established diagnostic criteria, we decided to use a multidimensional
continuous measure of IA.
Materials and Methods
Participants. Our study was conducted through online recruiting. A total of 602 adults participated in the
online survey on problematic internet use. According to gender, age and MRI safety parameters sixty healthy
Caucasian university students (30 males) aged between 18 and 30 (mean ± SD: 22.0 ± 2.08 years) were included.
All participants underwent a brief interview by a clinical psychologist and neurological expert to screen out
participants with a current psychiatric and neurologic diagnosis. Subjects with chronic illnesses, neurological or
psychiatric disorders were not included (for more details about the selection procedure, see the Fig.1). All sub-
jects had right-hand dominance according to the Edinburgh Handedness Inventory (Oldeld, 1971). Participants
spend on average 2.75 (SD = 2.74) hours online per day and consider themselves 45.27% (SD = 29.05) addicted
to the Internet. ey were either paid or received course credits for their participation and were naive with regard
to the purpose of the experiment.
e study was approved by the Regional Research Ethics Committee of the Clinical Center, University of Pécs.
All procedures performed in studies involving human participants were in accordance with the ethical stand-
ards of the institutional and national research committee and with the 1964 Helsinki declaration and its later
amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.
Assessment. Without clear diagnostic criteria, it is highly recommended to measure excessive Internet use
with a continuous questionnaire without using unclear cut-o scores11. erefore, we used the Problematic Internet
Use Questionnaire (PIUQ), a validated self-report scale with good reliability and validity characteristics13,33.
e questionnaire contains 18 items, each scored on a 5-point Likert-type scale ranging from 1 (never) to 5
(always). A conrmatory factor analysis veried the three factor model of questionnaire, each subscale contains
six items. Obsession subscale refers to obsessive thinking about the Internet (daydreaming, rumination and fan-
tasizing) and with- drawal symptoms caused by the lack of Internet use (anxiety, depression). (“How oen do
you feel tense, irritated, or stressed if you cannot use the Internet for as long as you want to?”). Neglect subscale
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
contains items about neglecting everyday activities, social life and essential needs (“How oen do you spend time
online when you’d rather sleep?”). Control disorder subscale reects diculties in controlling time spent on the
Internet (“How oen do you realize saying when you are online, ‘just a couple of more minutes and I will stop’?”).
Since in this study we focused on global psychological consequences of Internet addiction, we used PIUQ total
score in statistical analyses, that was computed by summing the scores on all the items of the scale.
Stimuli. Using Presentation soware (Neurobehavioral Systems, Inc., Berkeley, CA, USA), the visual stimuli were
presented via MRI-compatible goggles (VisualSystem, NordicNeuroLab AS, Bergen, Norway) and subjects’ responses
were collected via MRI-compatible response buttons (ResponseGrip, NordicNeuroLab AS, Bergen, Norway).
Since indirect behavioral evidences suggest that language network and verbal processes might be impaired
in IA34, we decided to use two dierent tasks in this study for controlling this possible interfering eect: verbal
Stroop task and non-verbal Stroop-like task.
Verbal stroop task. A series of colored words were displayed against a black background. Half of the words
were written in the same ink color as the meaning of the word (congruent stimuli, e.g., the word “green” displayed
in green color), while the other half were written in colors other than the word’s meaning (incongruent stimuli,
e.g., the word “green” displayed in blue color). Four colored words (blue, green, red, and yellow) and their corre-
sponding colors were used. Subjects had to choose the ink color (and neglect the meaning) of the words by using
the four response buttons (le thumb = red; le index = blue; right thumb = green; right index = yellow). Colored
circle thumbnails were presented on the bottom of screen to aid the subjects in which button to use for dierent
colors (Fig.1).
Non-verbal stroop-like task. A series of white arrows pointing either le or right was displayed against
a black background either on the le or right side of a centered xation cross. Half of the stimuli were pointing
in the same direction as their position on the screen (congruent stimuli, e.g., a leward pointing arrow on the
Figure 1. e selection of the nal sample. 602 subjects completed the online survey. 139 of them reported
smoking habits or excessive alcohol drinking and 28 individuals did not fall within the expected age range
(18–30). Since there was overlap between the groups (5 subjects), totally 162 subjects were excluded at this stage
of the process. Finally, 30 male and 30 female participants with no neurological and psychiatric symptoms and
no risk factors related to MRI measurements were chosen randomly from the remaining 440 subjects.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
le side of the xation cross), while the other half were pointing in an opposite direction as their position on the
screen (incongruent stimuli, e.g., leward pointing arrow on the right side of the xation cross). Subjects had to
choose the direction (and neglect the position) of the arrows by pressing one of the two buttons with their le or
right index ngers.
Experimental design. Stimulus presentations were analogous in both tasks. Stimuli were presented for
1200 ms each, with an interstimulus duration of 800 ms. Sixteen stimuli were presented in each session (32 s),
with 8 congruent and 8 incongruent stimuli, presented pseudo randomly. Half of the stimuli required le-handed,
while the other half required right-handed responses within each session.
e experiment started with a 20 s baseline session with a xation cross in the center of the screen. e base-
line section was followed by the rst nonverbal Stroop-like task session that started with a 6 s instruction period
followed by the congruent and incongruent stimuli. en, a 20 s baseline session with xation cross was con-
ducted following by the verbal Stroop task (again, starting with 6 s instruction period). e verbal Stroop and
non-verbal Stroop-like task sessions were presented successively, always interleaved with 20 s baseline. e whole
experiment consisted of 5–5 repetitions with 10 baseline sessions, resulting in a total measurement time of 580 s
(Fig.2). Reaction times (RTs) and error rates (ERs) were recorded for each condition (verbal_congruent, ver-
bal_incongruent, non-verbal_congruent, non-verbal_incongruent).
Imaging data acquisition and visual analysis. All measurements were performed on a 3 T
Magnetom TIM Trio human whole-body MRI scanner (Siemens AG, Erlangen, Germany) with a 12-channel
head coil. Functional images were acquired using a 2D single-shot echo-planar imaging (EPI) sequence (TR/
TE = 2000/30 ms; Flip angle = 76°; 36 axial slices with a thickness of 3 mm; FOV = 192 × 192 mm2; matrix
size = 64 × 64; receiver bandwidth = 2170 Hz/pixel; no gap; interleaved slice order to avoid crosstalk between
contiguous slices). For distortion correction purposes, eld mapping sequence (TR/TE1/TE2 = 400/4.92/7.38 ms;
Flip angle = 60°; 36 axial slices; FOV = 228 × 228 mm2; matrix size = 76 × 76; receiver bandwidth = 259 Hz/pixel)
with the same voxel size, orientation and adjustment parameters as the fMRI scan was acquired right aer the
fMRI measurement. Anatomical images were obtained using an isotropic T1-weighted 3D-MPRAGE sequence
(TR/TI/TE: 2530/1100/3.37 ms; Flip angle = 7°; 176 sagittal slices with a thickness of 1 mm; FOV = 256 × 256
mm2; matrix size = 256 × 256; receiver bandwidth = 200 Hz/pixel). e MPRAGE anatomical images were vis-
ually checked by MRI experts. ere were no brain abnormalities according to the visual analysis of the MR
images.
Functional MRI data processing and analysis. Pre-processing and statistical analyses were performed
using FEAT (FMRI Expert Analysis Tool) Version 6.00, part of FSL (FMRIB’s Soware Library, http://www.fmrib.
ox.ac.uk/fsl). Pre-processing included MCFLIRT motion correction, slice timing correction, brain extraction,
spatial smoothing with 5 mm full width at half maximum, and a high-pass temporal lter of 104 s (estimated
using FSL’s cutocalc tool). e temporal ltering applied to the data was used for the model as well. Whole brain
general linear model (GLM) time-series statistical analyses of individual data sets were carried out using FILM
(FMRIB’s Improved Linear Model) with local autocorrelation correction.
Figure 2. Verbal Stroop and Non-verbal Stroop-like task: design and stimuli examples. e experiment started
with a baseline session. It was followed by task sessions that were presented successively. Subjects had to choose
the direction of the arrows (Stroop-like task) or ink color of the words (Stroop task).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
First level analysis included ve regressors: verbal (Stroop task) congruent, verbal incongruent, non-verbal
(Stroop-like) congruent and non-verbal incongruent conditions, and an additional regressor to model BOLD
changes induced by the task instructions.
The single-session data sets were registered into the MNI152 standard space using a two-step process.
First, functional (EPI) image of each subject was registered to that subject’s T1 structural scan using BBR (6
degrees-of-freedom) incorporating simultaneous B0 eld unwarping (i.e. distortion correction) with a combina-
tion of FUGUE and BBR tools35,36. en, each subject’s T1 image was registered to the 2 mm MNI152 standard
space T1 image using a 12 degrees-of-freedom linear t followed by nonlinear registration (FNIRT, warp resolu-
tion = 8 mm). Next, for each subject, these two registrations were combined and applied to the rst-level statisti-
cal maps to take them into standard space.
Second-level mixed-eects analyses were carried out using FLAME (FMRIB’s Local Analysis of Mixed Eects,
stage 1 and 2) with outlier de-weighting to investigate the following questions:
i. Deactivation pattern during high-demand incongruent condition (verbal and non-verbal separately).
ii. Correlation between PIUQ and BOLD signal change during high-demand incongruent condition (verbal
and non-verbal tasks separately).
iii. Activation pattern during the incongruent condition (verbal and non-verbal separately).
iv. Correlation of PIUQ with BOLD signal change between incongruent and congruent conditions (incongru-
ent_minus_congruent; verbal and non-verbal separately).
Statistical maps were considered to be signicant at Z > 2.3 and a family-wise error corrected cluster signi-
cance threshold of p = 0.05.
Brain areas were considered to be part of the DMN if the following assumptions were satised for the local
maximas (LMs): a) deactivation measured during incongruent condition and b) according to the Harvard-Oxford
Cortical Structural Atlas (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Atlases) the LMs are in the areas related to the
DMN (incl. medial prefrontal, posterior cingulate cortex, precuneus or lateral parietal cortex)17.
Brain areas were considered to be part of the ICN if the following assumptions were satised for the LMs: a)
activation measured during incongruent_minus_congruent contrast indicating congruency eect or interference
eect; b) according to the Harvard-Oxford Cortical Structural Atlas the LMs are in the areas related to the ICN
(incl. inferior frontal and medial frontal gyri, the opercular, insular, orbitofrontal, posterior cingulate and poste-
rior parietal cortices)32.
Statistical analyses for behavioral data. Statistical analyses were performed using IBM SPSS Statistics
for Windows, Version 22.0 (IBM Corp. Released 2013. Armonk, NY, USA). Mean reaction time (RT) and error
rate (ER) dierences between congruent and incongruent stimuli were assessed by paired samples t-test or
Wilcoxon signed-rank test depending on the distribution of the data. Since PIUQ scores did not show normal
distribution, Spearman’s rank correlation was used to study the associations between RTs/ERs and PIUQ.
Results
Behavioral results. e mean of the total score on the PIUQ for our sample is 32.85 (SD = 12.4, 95% CI:
[29.62–36.08].
Signicant dierences were found between congruent and incongruent stimuli in RTs and ERs regarding both
the verbal Stroop and the non-verbal Stroop-like task (Table1). No signicant correlations were found between
task performance scores (incl. reaction times – RT, error rates – ER and congruency eect [incongruent_minus_
congruent for RTs and ERs) and PIUQ total.
fMRI results. e correlations between PIUQ and BOLD fMRI results for incongruent and incongruent_
minus_congruent contrasts are presented in Table2 (see also Fig.3). During the incongruent stimuli in the verbal
Stroop task, signicant negative correlations were found in the bilateral posterior cingulate gyri and bilateral
Tas k Condition Mean (SD) or
Median (min-max) t/Z rho
Correlation with PIUQ
Verbal Stroop RTaCong. 0.87 (0.09) 4.96*** 0.077
Incong. 0.96 (0.17) 0.077
Verbal Stroop ERbCong. 1 (0–18) 5.80*** −0.01
Incong. 5 (0–18) −0.02
Non-verbal Stroop RTaCong. 0.65 (0.10) 8.43*** −0.11
Incong. 0.68 (0.10) −0.12
Non-verbal Stroop ERbCong. 0 (0–4) 4.16*** 0.15
Incong. 1 (0–5) 0.04
Table 1. Dierences in task performances between conditions and correlations with PIUQ. aPaired samples
t-test, mean (SD) and t values are presented; bWilcoxon signed-rank test, median (min-max) and Z scores are
presented; ***p < 0.001, RT – reaction time, ER – error rate, PIUQ – Problematic Internet Use Questionnaire,
rho – Spearman’s correlation coecient.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
precuneus. During the incongruent stimuli in non-verbal Stroop-like task, signicant negative correlations were
detected in the bilateral precuneus, the right middle frontal gyrus and right precentral gyrus. BOLD signal was
decreased with the severity of Internet addiction in both tasks.
Incongruent_minus_congruent contrast for Stroop task positively correlated with PIUQ in several le sided
cortical areas including the inferior frontal gyrus (pars opercularis), frontal pole, central and frontal opercular
cortex, frontal orbital cortex and insular cortex. e same contrast for non-verbal Stroop-like task showed no
signicant relationship with PIUQ.
Compared to the baseline, signicant deactivations were found in the areas related to the DMN during incon-
gruent stimuli in Stroop and Stroop-like tasks (see Fig.S1). Areas related to the ICN showed signicant activation
during incongruent stimuli in Stroop task (Fig.S2).
Discussion
In this study, functional correlates of Internet addiction were demonstrated during verbal Stroop and non-verbal
Stroop-like task in young adult Internet users. Since the dierent task-related fMRI contrasts represent dierent
psychological and neural domains they will be discussed separately.
Incongruent stimuli induced BOLD signal changes were found to be negatively correlated with PIUQ in bilat-
eral precuneus and posterior cingulate gyri (PCG) during Stroop task and bilateral precuneus, during Stroop-like
task. Moreover, we found signicant deactivations in these areas during incongruent stimuli for both verbal
Stroop and non-verbal Stroop-like tasks. Brain areas in the cerebral cortex with constantly decreasing activity
during demanding tasks are considered being part of the DMN16. Previous studies investigating Internet addic-
tion and Internet gaming disorder (IGD) also found alterations in the DMN. However, these studies revealed
the involvement of the anterior part of the network and used merely resting-state fMRI. Li et al.28 demonstrated
decreased anticorrelation between the right dorsolateral prefrontal cortex and the anterior part of the DMN
(medial prefrontal cortex and anterior cingulate gyrus) that might lead to dysfunctions of the cognitive control
network and DMN, including diminished cognitive eciency. Wang et al. also revealed reduced functional con-
nectivity in the anterior part of the DMN, as well as decreased interactions between the salient network and DMN
in adolescents37. Authors claim that this may serve as neural background of the uncontrolled Internet use and
they also suggest that IA may share similar neurobiological abnormalities than other addictions. Similar impaired
connectivity patterns were found in IGD15. In addition, functional connectivity among DMN regions were in
negative correlation with anger and depression38, as well as with executive control dysfunction39. e results
of our study have signicant contribution to advance our knowledge on the eld. We are the rst to show that
task-related changes in activation of DMN-related structures are also signicantly related to IA. e highlighted
areas (precuneus and PCG) are parts of the posterior part of the DMN and previous studies revealed that the
middle frontal gyrus observed in the current study is also strongly related to the DMN40.
Impaired DMN may explain some comorbid symptoms that oen occur in IA, thereby it may have important
clinical implications. Firstly, many evidences indicate that DMN plays an important role in cognition, particu-
larly in attention and memory processes. Wide range of studies showed that vigilance41, semantic processing42,
Cluster Area #voxels Max.
Z-score
MNI coordinates
x y z
PIUQ INCONGRUENT
Verbal Stroop task
1.
Le posterior cingulate gyrusa506 4.09 0 −42 34
Right precuneusa14 −50 46
Right posterior cingulate gyrusa434 3.78 8 −40 30
2. Le precuneusa−4−58 10
Non-verbal Stroop-like task
1. Le precuneusa386 3.95 −2−60 52
Right precuneusa8−68 50
2. Right middle frontal gyrusa378 4 40 8 38
Right precentral gyrusa42 2 42
PIUQ INCONGRUENT_MINUS_CONGRUENT
Verbal Stroop task
1.
Le inferior frontal gyrus pars opercularisb275 3.45 −50 16 0
Le frontal poleb−36 40 8
Le central opercular cortexb−38 8 10
Le frontal opercular cortexb−42 24 2
Le frontal orbital cortexb−44 20 −6
Le insular cortexb−42 10 −2
Non-verbal Stroop-like task n.s.
Table 2. Correlations between PIUQ and fMRI BOLD signal changes. aNegative correlation bpositive
correlation; n.s. not signicant; MNI coordinates are listed for the local maximas.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
episodic, autobiographical memory43 and creative problem solving44 are strongly associated with DMN activa-
tion, moreover, diminished vascular functionality for the DMN in mid-life may serve as a preclinical marker for
brain dysfunction in later life45. ese observations may implicate long term negative eect of IA on cognition
- similarly to those presented in nicotine addiction27, and suggest that DMN alterations serve as the common
neuro-functional background behind IA and attention-related decits46. Although, it must be note that IA is still
a relatively young phenomenon and more evidences are needed to prove this assumption. Secondly, according
to previous studies connectivity within DMN can be changed with pharmacotherapies via bottom-up and with
cognitive behavioral therapy (CBT) via top-down mechanisms in some psychiatric disorders (for review see47),
however, the exact mechanisms are unclear. Furthermore, individual variations in pre-treatment functional con-
nectivity may be a reliable predictor of treatment ecacy in schizophrenia48, major depression49 and smoking27.
It is important to note that these studies did not investigate the DMN as a holistic functional brain system, they
focused rather on its connection with other networks (executive control network)27, on a highlighted component
of the DMN (dorsolateral prefrontal cortex)48 and subnetworks of the DMN (anterior and posterior subnetworks;
abnormal functional connectivity within the posterior part was normalized aer antidepressant treatment, while
anterior subnetwork remained persistent against treatment)49. Since CBT seems to be successful methods for
improving IA status50, individual dierences in DMN might also predict treatment outcome and be reliable and
objective tool for assessing therapeutic ecacy. Further studies are needed to explore this.
In Stroop task the incongruent minus congruent contrast showed signicant positive correlation with PIUQ
in the le inferior frontal gyrus, le frontal pole, le central opercular, le frontal opercular, le frontal orbital
and le insular cortex. Activation in these areas were reported to be correlated with individual dierences in task
performances requiring good inhibition ability32,51. Congdon et al. used Stop-signal task-related probabilistic
ICA (independent component analysis) and referred two components that included similar regions found in
our study32. ey demonstrated that engagement of networks that include these regions is positively related to
response inhibition that is the ability to suppress a habitual response or routine behavior, including motor actions
and higher-order responses (thoughts, emotions etc.). erefore, these networks are essential to stop problem-
atic behaviors – such as excessive Internet use. Ergo, it is not surprising that Internet addiction was found to be
associated with impaired inhibition in previous studies30,52. Here we suggest that a special “inhibitory control
network (ICN)” that was found by Congdon and colleagues32 should be considered as a possible pathogenic
factor in the development of IA. However, signicant correlations were only found during the Stroop task as an
important unexpected result. e most plausible explanation is that the Stroop-like non-verbal task was not sen-
sitive enough to detect individual dierences due to the relatively simple nature of the task even in incongruent
condition. Despite there were signicant dierences between incongruent and congruent conditions in RT and
ER, incongruency eect did not occur in an expected way. Another possible explanation might be that the verbal
domain is altered in IA. is is supported by the presence of le hemisphere specialization in our study that is
known to be typical for language processing53. At this stage of knowledge, this explanation seems highly specula-
tive, and further studies are needed.
Some limitations must be considered. First, the cross-sectional nature of the study limits our ability to dis-
criminate between cause and eect. Does excessive Internet use lead to brain-related changes or vica versa?
Secondly, IA is known to have various comorbid symptoms such as depression, attention decit hyperactivity
disorder, anxiety disorders etc. Although, we excluded patients with diagnosed comorbid diseases, subjects with
subclinical symptoms may have inuenced our results. Another limitation is related to recruitment through
online platforms. e major problems with this method are that the representativeness of the sample cannot be
fully determined and problematic Internet users could be overrepresented in the sample.
Figure 3. Group level negative associations between PIUQ and BOLD signal changes during incongruent
condition in verbal Stroop task (A) and non-verbal Stroop-like task (B). Group level positive associations
between PIUQ and BOLD signal changes during incongruent_minus_congruent contrast in Stroop task (C).
Images were thresholded using clusters determined by Z > 2.3 and a corrected cluster signicance threshold of
p = 0.05. Axial slices are shown in radiological convention for MNI slice coordinates from Z = 8 to 48 mm in
(A), Z = 22 to 62 mm in (B) and Z = −16 to 24 mm in (C).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

8
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
Taken together, our results suggest that problematic Internet use in healthy young adults has functional brain
correlates in regions that are related to the DMN. Since similar associations were found in IGD54, smoking18 and
other substance related addictions21 our results suggest long-term negative eects of IA on brain functions. As
described before, DMN is strongly related to human cognitive performance and some features of the DMN were
able to predict therapy outcomes in psychiatric disorders. erefore, we suggest, that altered DMN might explain
some comorbid symptoms (including attention decit) and might predict treatment outcomes (for further infor-
mation about Internet-related changes in human cognition see the review article of Loh & Kanai55). Although
we must highlight, that in the lack of well-designed, reliable researches, these assumptions are speculative at the
moment. Activation in the ICN was also found to be correlated with problematic Internet use suggesting that
diculties in stopping and controlling overuse are not results of “weak character of the person” but the impaired
brain mechanisms responsible for behavioral inhibition. However, it must be highlighted that, since we investi-
gated only healthy young people, our conclusions should be applied only to this group. Longitudinal studies and
studies with other age groups (e.g. adolescents) are highly needed to test if these ndings can be generalized to
other groups and whether these functional correlates are the result of overall additive tendencies or specic to
maladaptive Internet use. It would be also important to know that controlled Internet use or complete abstinence
could reduce these brain-related changes.
Received: 16 May 2019; Accepted: 11 October 2019;
Published: xx xx xxxx
References
1. im, . et al. Internet addiction in orean adolescents and its relation to depression and suicidal ideation: A questionnaire survey.
International Journal of Nursing Studies 43, 185–192 (2006).
2. Norett, L. Quantitative research. Nurs. Stand. 27, 59–59 (2013).
3. odgers, . F., Melioli, T., Laconi, S., Bui, E. & Chabrol, H. Internet Addiction Symptoms, Disordered Eating, and Body Image
Avoidance. Cyberpsychology, Behav. Soc. Netw. 16, 56–60 (2013).
4. Cheung, L. M. & Wong, W. S. e eects of insomnia and internet addiction on depression in Hong ong Chinese adolescents: An
exploratory cross-sectional analysis. J. Sleep Res. 20, 311–317 (2011).
5. Young, . S. Internet Addiction: e Emergence of a New Clinical Disorder. CyberPsychology Behav. 1, 237–244 (1998).
6. Mihajlov, M. & Vejmela, L. Internet addiction: A review of the rst twenty years. Psychiatr. Danub. 29, 260–272 (2017).
7. irály, O., Griffiths, M. D. & Demetrovics, Z. Internet Gaming Disorder and the DSM-5: Conceptualization, Debates, and
Controversies. Curr. Addict. Reports 2, 254–262 (2015).
8. eed, G. M. et al. Innovations and changes in the ICD-11 classication of mental, behavioural and neurodevelopmental disorders.
World Psychiatry 18, 3–19 (2019).
9. Montag, C. et al. Is it meaningful to distinguish between generalized and specic Internet addiction? Evidence from a cross-cultural
study from Germany, Sweden, Taiwan and China. Asia-Pacic. Psychiatry 7, 20–26 (2015).
10. Day, J. J., oitman, M. F., Wightman, . M. & Carelli, . M. Associative learning mediates dynamic shis in dopamine signaling in
the nucleus accumbens. Nat. Neurosci. 10, 1020–1028 (2007).
11. Altbäcer, A. et al. Problematic internet use is associated with structural alterations in the brain reward system in females. Brain
Imaging Behav. 10, 953–959 (2016).
12. ühn, S. & Gallinat, J. Brains online: Structural and functional correlates of habitual Internet use. Addict. Biol. 20, 415–422 (2015).
13. Demetrovics, Z., Szeredi, B. & ózsa, S. e three-factor model of Internet addiction: e development of the Problematic Internet
Use Questionnaire. Behav. Res. Methods 40, 563–574 (2008).
14. Steele, V. ., Ding, X. & oss, T. J. Addiction: Informing drug abuse interventions with brain networs. Connectomics 101–122,
https://doi.org/10.1016/b978-0-12-813838-0.00006-6 (2019).
15. Zhang, J. T. et al. Altered coupling of default-mode, executive-control and salience networs in Internet gaming disorder. Eur.
Psychiatry 45, 114–120 (2017).
16. Shulman, G. L. et al. Common Blood Flow Changes across Visual Tass: II. Decreases in Cerebral Cortex. J. Cogn. Neurosci. 9,
648–663 (1997).
17. aichle, M. E. e Brain’s Default Mode Networ. Annu. Rev. Neurosci. 38, 433–447 (2015).
18. Ding, X. & Lee, S.-W. Changes of Functional and Eective Connectivity in Smoing eplenishment on Deprived Heavy Smoers: A
esting-State fMI Study. PLoS One 8, e59331 (2013).
19. Zhang, . & Volow, N. D. Brain default-mode networ dysfunction in addiction. Neuroimage 200, 313–331 (2019).
20. Yuan, . et al. Core brain networs interactions and cognitive control in internet gaming disorder individuals in late adolescence/
early adulthood. Brain Struct. Funct. 221, 1427–1442 (2016).
21. Ma, N. et al. Abnormal brain default-mode networ functional connectivity in drug addicts. PLoS One 6, e16560 (2011).
22. Arcurio, L. ., Finn, P. . & James, T. W. Neural mechanisms of high-ris decisions-to-drin in alcohol-dependent women. Addict.
Biol. 20, 390–406 (2015).
23. Lin, F., Wu, G., Zhu, L. & Lei, H. Altered brain functional networs in heavy smoers. Addict. Biol. 20, 809–819 (2015).
24. Wetherill, . . et al. Cannabis, cigarettes, and their co-occurring use: Disentangling dierences in default mode networ functional
connectivity. Drug Alcohol Depend. 153, 116–123 (2015).
25. Weiland, B. J., Sabbineni, A., Calhoun, V. D., Welsh, . C. & Hutchison, . E. educed executive and default networ functional
connectivity in cigarette smoers. Hum. Brain Mapp. 36, 872–882 (2015).
26. Huan g, W. et al. The development and expression of physical nicotine dependence corresponds to structural and functional
alterations in the anterior cingulate-precuneus pathway. Brain Behav. 4, 408–417 (2014).
27. Cole, D. M. et al. Nicotine replacement in abstinent smoers improves cognitive withdrawal symptoms with modulation of resting
brain networ dynamics. Neuroimage 52, 590–599 (2010).
28. Li, W. et al. Brain structures and functional connectivity associated with individual dierences in Internet tendency in healthy young
adults. Neuropsychologia 70, 134–144 (2015).
29. Dong, G., Zhou, H. & Zhao, X. Male Internet addicts show impaired executive control ability: Evidence from a color-word Stroop
tas. Neurosci. Lett. 499, 114–118 (2011).
30. Dong, G., DeVito, E. E., Du, X. & Cui, Z. Impaired inhibitory control in ‘internet addiction disorder’: A functional magnetic
resonance imaging study. Psychiatry Res. Neuroimaging 203, 153–158 (2012).
31. de Wit, H. Impulsivity as a determinant and consequence of drug use: a review of underlying processes. Addict. Biol. 14, 22–31
(2009).
32. Congdon, E. et al. Engagement of large-scale networs is related to individual dierences in inhibitory control. Neuroimage 53,
653–663 (2010).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

9
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
33. oronczai, B. et al. Conrmation of the ree-Factor Model of Problematic Internet Use on O-Line Adolescent and Adult Samples.
Cyberpsychology, Behav. Soc. Netw. 14, 657–664 (2011).
34. Nie, J., Zhang, W. & Liu, Y. Exploring depression, self-esteem and verbal uency with dierent degrees of internet addiction among
Chinese college students. Compr. Psychiatry 72, 114–120 (2017).
35. Greve, D. N. & Fischl, B. Accurate and robust brain image alignment using boundary-based registration. Neuroimage 48, 63–72
(2009).
36. Jeninson, M., Bannister, P., Brady, M. & Smith, S. Improved Optimization for the obust and Accurate Linear egistration and
Motion Correction of Brain Images Improved Optimization for the obust and Accurate Linear egistration and Motion Correction
of Brain Images. Neuroimage 17, 825–841 (2002).
37. Wang, L. et al. Altered default mode, fronto-parietal and salience networs in adolescents with Internet addiction. Addict. Behav. 70,
1–6 (2017).
38. Dong, G., Li, H., Wang, L. & Potenza, M. N. e correlation between mood states and functional connectivity within the default
mode networ can dierentiate Internet gaming disorder from healthy controls. Prog. Neuro-Psychopharmacology Biol. Psychiatry
77, 185–193 (2017).
39. Wang, L. et al. Dysfunctional default mode networ and executive control networ in people with Internet gaming disorder:
Independent component analysis under a probability discounting tas. Eur. Psychiatry 34, 36–42 (2016).
40. Uddin, L. Q., elly, A. M., Biswal, B. B., Castellanos, F. X. & Milham, M. P. Functional connectivity of default mode networ
components: correlation, anticorrelation, and causality. Hum. Brain Mapp. 30, 625–37 (2009).
41. Hinds, O. et al. Computing moment-to-moment B OLD activation for real-time neurofeedbac. Neuroimage 54, 361–368 (2011).
42. Binder, J. . & Desai, . H. e neurobiology of semantic memory. Trends in Cognitive Sciences 15, 527–536 (2011).
43. Philippi, C. L., Tranel, D., Du, M. & udrauf, D. Damage to the default mode networ disrupts autobiographical memory retrieval.
Soc. Cogn. Aect. Neurosci. 10, 318–326 (2015).
44. ühn, S. et al. e Importance of the Default Mode Networ in Creativity-A Structural MI Study. J. Creat. Behav. 48, 152–163
(2014).
45. Haight, T. J. et al. Vascular ris factors, cerebrovascular reactivity, and the default-mode brain networ. Neuroimage 115, 7–16
(2015).
46. Yen, J.-Y., o, C.-H., Yen, C.-F., Wu, H.-Y. & Yang, M.-J. e Comorbid Psychiatric Symptoms of Internet Addiction: Attention
Decit and Hyperactivity Disorder (ADHD), Depression, So cial Phobia, and Hostility. J. Adolesc. Heal. 41, 93–98 (2007).
47. Weingarten, C. P. & Strauman, T. J. Neuroimaging for psychotherapy research: current trends. Psychother. Res. 25, 185–213
(2015).
48. umari, V. et al. Dorsolateral Prefrontal Cortex Activity Predicts esponsiveness to Cognitive-Behavioral erapy in Schizophrenia.
Biol. Psychiatry 66, 594–602 (2009).
49. Li, B. et al. A treatment-resistant default mode subnetwor in major depression. Biol. Psychiatry 74, 48–54 (2013).
50. Winler, A., Dörsing, B., ief, W., Shen, Y. & Glombiewsi, J. A. Treatment of internet addiction: A meta-analysis. Clin. Psychol. Rev.
33, 317–329 (2013).
51. Goghari, V. M. & MacDonald, A. W. e neural basis of cognitive control: esponse selection and inhibition. Brain Cogn. 71, 72–83
(2009).
52. Dong, G., Lu, Q., Zhou, H. & Zhao, X. Impulse inhibition in people with Internet addiction disorder: Electrophysiological evidence
from a Go/NoGo study. Neurosci. Lett. 485, 138–142 (2010).
53. Vigneau, M. et al. Meta-analyzing le hemisphere language areas: Phonology, semantics, and sentence processing. NeuroImage 30,
1414–1432 (2006).
54. Ding, W. N. et al. Altered Default Networ esting-State Functional Connectivity in Adolescents with Internet Gaming Addiction.
PLoS One 8, e59902 (2013).
55. Loh, . . & anai, . How Has the Internet eshaped Human Cognition? Neuroscientist 22, 506–520 (2016).
Acknowledgements
is paper was supported by the PTE ÁOK-KA-2017-05 and PTE ÁOK-KA-2017-06, Hungarian Brain Research
Program 20017-1.2.1-NKP −2017-00002 and KTIA_NAP_13-2-2014-0019 government-based funds, EFOP-
3.6.2-16-2017-00008. e role of neuro-inammation in neurodegeneration: from molecules to clinics”, ÚNKP-
17-3 -I.-PTE-173, ÚNKP-17- 4 -I.- PTE-311, ÚNKP-17-4-I-PTE-76, ÚNKP-17-4-III-PTE-93, and ÚNKP-17-3-
III-PTE-315 New National Excellence Program of the Ministry of Human Capacities, N.K. and J.J. was supported
by the SNN125143 research grant, G.P. and G.O. was supported by the János Bolyai Research Scholarship of the
Hungarian Academy of Sciences. Z.D. was supported by the Hungarian National Research, Development and
Innovation Oce (Grant numbers: K111938, KKP126835).
Author contributions
G.D. was responsible for writing the manuscript and supervising the whole process. G.P., A.N.Z., I.O. and
G.O. were responsible for M.R.I. measurements and the statistical analysis. R.H., S.A.N., B.L. and D.T. were
responsible for the manuscript review and critique. N.K., T.D., Z.D. and J.J. contributed equally to the conception,
organization, and execution of the research project.
Competing interests
e authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41598-019-52296-1.
Correspondence and requests for materials should be addressed to G.D.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

10
SCIENTIFIC REPORTS | (2019) 9:15777 | https://doi.org/10.1038/s41598-019-52296-1
www.nature.com/scientificreports
www.nature.com/scientificreports/
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2019
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com