Conceived and designed the experiments: FL YZ YD JX HL. Performed the experiments: YZ LQ ZZ. Analyzed the data: FL HL. Contributed reagents/materials/analysis tools: YZ YD FL. Wrote the paper: FL HL.
The authors have declared that no competing interests exist.
Internet addiction disorder (IAD) is currently becoming a serious mental health issue around the globe. Previous studies regarding IAD were mainly focused on associated psychological examinations. However, there are few studies on brain structure and function about IAD. In this study, we used diffusion tensor imaging (DTI) to investigate white matter integrity in adolescents with IAD.
Seventeen IAD subjects and sixteen healthy controls without IAD participated in this study. Whole brain voxel-wise analysis of fractional anisotropy (FA) was performed by tract-based spatial statistics (TBSS) to localize abnormal white matter regions between groups. TBSS demonstrated that IAD had significantly lower FA than controls throughout the brain, including the orbito-frontal white matter, corpus callosum, cingulum, inferior fronto-occipital fasciculus, and corona radiation, internal and external capsules, while exhibiting no areas of higher FA. Volume-of-interest (VOI) analysis was used to detect changes of diffusivity indices in the regions showing FA abnormalities. In most VOIs, FA reductions were caused by an increase in radial diffusivity while no changes in axial diffusivity. Correlation analysis was performed to assess the relationship between FA and behavioral measures within the IAD group. Significantly negative correlations were found between FA values in the left genu of the corpus callosum and the Screen for Child Anxiety Related Emotional Disorders, and between FA values in the left external capsule and the Young's Internet addiction scale.
Our findings suggest that IAD demonstrated widespread reductions of FA in major white matter pathways and such abnormal white matter structure may be linked to some behavioral impairments. In addition, white matter integrity may serve as a potential new treatment target and FA may be as a qualified biomarker to understand the underlying neural mechanisms of injury or to assess the effectiveness of specific early interventions in IAD.
Internet addiction disorder (IAD), also called problematic or pathological Internet use, is characterized by an individual's inability to control his or her use of the Internet, which may eventually result in marked distress and functional impairments of general life such as academic performance, social interaction, occupational interest and behavioral problems
Current studies about IAD have focused on case summaries, behavioral components, negative consequences in daily life, along with clinical diagnosis, epidemiology, associated psychosocial factors, symptom management, psychiatric comorbidity and treatment outcome
To date, only few neuroimaging studies had been performed to investigate brain structural and functional changes associated with IAD. A previous voxel-based morphometry (VBM) study reported decreased gray matter density in the left anterior cingulate cortex, posterior cingulate cortex, insula and lingual gyrus of IAD adolescents
We hypothesize that IAD subjects are also associated with impairments of white matter fibers connecting these regions and such changes can be detected by diffusion tensor imaging (DTI), a non-invasive MRI technique with capable of providing a quantitative measure of white matter damage
In this study, we used DTI to investigate the white matter integrity in adolescents with IAD. An observer-independent tract-based spatial statistics (TBSS) analysis method was used to analyze the DTI data. This method retains the strengths of voxel-based analysis while addressing some of its drawbacks, such as aligning images from multiple subjects and the arbitrariness of the choice of spatial smoothing
Eighteen adolescents with IAD were recruited from the Department of Child and Adolescent Psychiatry, Shanghai Mental Health Center, all of whom met the modified Young's diagnostic questionnaire for internet addiction criteria by Beard and Wolf
CON (n = 16) | IAD (n = 17) | ||
(Mean±SD) | (Mean±SD) | ||
Age | 17.78±2.46 | 17.01±2.50 | 0.38 |
Gender (M/F) | 14/2 | 15/2 | 0.95 |
Education (years) | 11.50±2.99 | 10.47±2.40 | 0.28 |
Young's Internet Addiction Scale (YIAS) | 37.00±10.64 | 64.71±12.58 |
|
Time Management Disposition Scale (TMDS) | 123.60±20.17 | 124.00±22.80 | 0.96 |
Strength and Difficulties Questionnaire (SDQ) | 16.40±3.87 | 21.76±3.46 |
|
Barratt Impulsiveness Scale-11 (BIS) | 67.20±7.83 | 69.82±12.34 | 0.49 |
The Screen for Child Anxiety Related Emotional Disorders (SCARED) | 24.71±6.16 | 38.59±9.90 |
|
Family Assessment Device (FAD) | 117.73±10.89 | 129.12±13.93 |
|
Abbreviation. CON: controls; IAD: Internet addiction disorder; SD: standard deviation.
Two-sample
The study was approved by the Ethics Committee of RenJi Hospital of Shanghai Jiao Tong University Medical School. The participants and their parents/legal guardians were informed of the aims of our study before MRI examinations. Full written informed consent was obtained from the parents/guardians of each participant.
All subjects underwent a simple physical examination including blood pressure and heart rate measurements, and were interviewed by a psychiatrist regarding their medical history on nervous, motion, digestive, respiratory, circulation, endocrine, urinary and reproductive systems. They were then screened for psychiatric disorders with the Mini International Neuropsychiatric Interview for Children and Adolescents (MINI-KID)
The diagnostic standard for IAD was adapted from the modified Young's Diagnostic Questionnaire for Internet Addiction criteria by Beard and Wolf
Six questionnaires were used to assess the participants' behavioral features, namely the Young's Internet Addiction Scale (YIAS)
Diffusion tensor imaging was performed on a 3.0-Tesla Phillips Achieva medical scanner. A single-shot echo planar diffusion weighted imaging with alignment of the anterior-posterior commissures plane was done according to the following parameters: repetition time = 8,044 ms; echo time = 68 ms; SENSE factor = 2; acquisition matrix = 128×128 zero-filled to 256×256; field of view = 256×256 mm2; slice thickness = 4 mm without gap. A total of 34 sections covered the whole brain including the cerebellum. The diffusion sensitizing gradients were applied along 15 non-collinear gradient encoding directions with b = 800 s/mm2. One additional image without diffusion gradients (b = 0 s/mm2) was also acquired. To enhance signal to noise ratio, imaging was repeated three times.
All DTI data were preprocessed by the FMRIB's Diffusion Toolbox (FDT) within FMRIB's Software Library (FSL;
Whole brain analysis of FA images was performed by using TBSS
To identify FA differences between IAD subjects and normal controls, the skeletonized FA data were fed into the voxel-wise statistics analysis which is based on non-parametric approach utilizing permutation test theory. The testing was performed by the FSL randomise program, which uses 5000 random permutations. Two contrasts were estimated: IAD subjects greater than controls and controls greater than IAD subjects. Age was entered into the analysis as a covariate to ensure that any observed difference of FA between groups was independent of age-related changes. Threshold-free cluster enhancement (TFCE)
In order to explore the microstructural mechanisms of the observed FA changes, volume-of-interest (VOI) analysis was performed to investigate changes of diffusivity indices (Da, Dr and MD) in the regions showing FA abnormalities. To do so, the VOI masks were first extracted based on the clusters showing significant inter-group FA differences. These VOIs masks were then back- projected to the original images of each subject, and the mean values of the diffusion indices within the VOIs were calculated. After confirming normal distribution of the data by a one-sample Kolmogorov-Smirnov test, one-way analysis of covariance (ANCOVA) with group as the independent variable and diffusion indices as the dependent variables was performed, controlling for age of subjects. A statistical significance level of
Pearson correlation analyses were used to test the correlations between FA changes within the VOIs and behavioral measures. A
A value of 0.2 was used to threshold the mean FA skeleton volume such that a total of 131962 voxels were entered into voxel-wise TBSS analysis. The spatial distribution of the brain regions showing reduced FA in the IAD group is presented in
Areas in red are regions where FA was significantly lower (
Anatomic region | Hemisphere | MNI coordinates (mm) | Cluster size(mm3) | ||||
X | Y | Z | |||||
Frontal Lobe | Orbital frontal WM | R | 8 | 40 | −20 | 0.008 | 86 |
Orbital frontal WM | L | −13 | 41 | −16 | 0.007 | 119 | |
Commissural fiber | Genu of corpus callosum | R | 14 | 28 | 15 | 0.002 | 288 |
Genu of corpus callosum | L | −15 | 31 | 15 | 0.004 | 241 | |
Body of corpus callosum | R | 14 | 15 | 26 | 0.003 | 368 | |
Body of corpus callosum | L | −14 | 16 | 27 | 0.004 | 322 | |
Splenium | R | 19 | −34 | 32 | 0.008 | 81 | |
Associate fiber | Inferior fronto-occipital fasciculus | R | 28 | 34 | 0 | 0.003 | 150 |
Inferior fronto-occipital fasciculus | L | −27 | 32 | 9 | 0.004 | 174 | |
Cingulum | R | 9 | 2 | 33 | 0.007 | 37 | |
Cingulum | L | −8 | 3 | 31 | 0.009 | 34 | |
Projection fiber | Anterior corona radiation | R | 17 | 28 | 20 | 0.002 | 877 |
Anterior corona radiation | L | −24 | 29 | 7 | 0.003 | 1037 | |
Superior corona radiation | R | 17 | 15 | 30 | 0.003 | 276 | |
Superior corona radiation | L | −17 | −6 | 36 | 0.004 | 275 | |
Posterior corona radiation | R | 22 | −30 | 34 | 0.007 | 106 | |
Posterior corona radiation | L | −19 | −30 | 35 | 0.009 | 25 | |
Anterior limb of internal capsule | R | 22 | 21 | 7 | 0.007 | 45 | |
Anterior limb of internal capsule | L | −21 | 20 | 8 | 0.005 | 78 | |
Precentral gyrus | L | −19 | −12 | 43 | 0.007 | 149 | |
External capsule | R | 28 | 12 | 17 | 0.007 | 25 | |
External capsule | L | −26 | 17 | 8 | 0.005 | 24 |
Abbreviation. MNI: Montreal Neurological Institute; WM: white matter; R: right; L: left.
Note. Coordinates for the peak voxels are displayed.
The 22 brain regions showing significantly reduced FA in the IAD group were extracted for VOI-based analysis of other diffusion indices. The results are listed in
Anatomic region | Da (×10−3 mm2/s) (Mean±SD) | Dr (×10−3 mm2/s) (Mean±SD) | MD (×10−3 mm2/s) (Mean±SD) | ||||||
CON | IAD | F value | CON | IAD | F value | CON | IAD | F value | |
Orbital frontal WM R | 1.12±0.08 | 1.17±0.08 | 3.24 | 0.66±0.08 | 0.74±0.06 | 11.93 |
0.81±0.08 | 0.89±0.07 | 8.89 |
Orbital frontal WM L | 1.36±0.09 | 1.33±0.07 | 0.66 | 0.58±0.04 | 0.62±0.04 | 7.62 | 0.84±0.05 | 0.86±0.04 | 1.23 |
Genu of corpus callosum R | 1.51±0.08 | 1.56±0.11 | 1.79 | 0.45±0.04 | 0.53±0.05 | 29.33 |
0.80±0.05 | 0.88±0.06 | 14.19 |
Genu of corpus callosum L | 1.58±0.08 | 1.64±0.10 | 2.66 | 0.42±0.04 | 0.49±0.04 | 24.04 |
0.81±0.04 | 0.87±0.05 | 15.08 |
Body of corpus callosum R | 1.37±0.07 | 1.43±0.09 | 4.13 | 0.48±0.05 | 0.57±0.05 | 28.47 |
0.78±0.04 | 0.86±0.05 | 22.83 |
Body of corpus callosum L | 1.37±0.09 | 1.38±0.06 | 0.29 | 0.48±0.04 | 0.55±0.05 | 21.26 |
0.78±0.04 | 0.83±0.04 | 15.94 |
Splenium R | 1.51±0.07 | 1.51±0.06 | 0.003 | 0.40±0.06 | 0.44±0.03 | 6.99 | 0.77±0.06 | 0.80±0.03 | 3.19 |
Inferior fronto-occipital fasciculus R | 1.17±0.04 | 1.16±0.06 | 0.46 | 0.61±0.04 | 0.66±0.03 | 15.83 |
0.79±0.04 | 0.83±0.04 | 5.06 |
Inferior fronto-occipital fasciculus L | 1.15±0.05 | 1.16±0.05 | 0.21 | 0.58±0.03 | 0.63±0.03 | 16.82 |
0.77±0.03 | 0.80±0.04 | 7.63 |
Cingulum R | 1.27±0.12 | 1.32±0.10 | 1.48 | 0.46±0.07 | 0.55±0.06 | 18.30 |
0.73±0.07 | 0.80±0.05 | 14.91 |
Cingulum L | 1.27±0.11 | 1.30±0.14 | 0.40 | 0.44±0.06 | 0.53±0.05 | 20.24 |
0.72±0.06 | 0.79±0.06 | 10.68 |
Anterior corona radiation R | 1.31±0.04 | 1.31±0.08 | 0.04 | 0.55±0.03 | 0.61±0.04 | 31.42 |
0.80±0.03 | 0.84±0.04 | 11.02 |
Anterior corona radiation L | 1.27±0.05 | 1.26±0.05 | 0.19 | 0.55±0.04 | 0.60±0.03 | 18.53 |
0.79±0.03 | 0.82±0.03 | 7.57 |
Superior corona radiation R | 1.23±0.04 | 1.22±0.05 | 0.01 | 0.51±0.02 | 0.56±0.03 | 37.68 |
0.75±0.02 | 0.78±0.02 | 17.12 |
Superior corona radiation L | 1.25±0.05 | 1.23±0.05 | 1.02 | 0.50±0.03 | 0.55±0.03 | 20.65 |
0.75±0.03 | 0.77±0.02 | 9.89 |
Posterior corona radiation R | 1.19±0.04 | 1.16±0.05 | 2.53 | 0.58±0.04 | 0.61±0.03 | 7.94 | 0.78±0.03 | 0.79±0.02 | 1.66 |
Posterior corona radiation L | 1.22±0.08 | 1.14±0.07 | 9.25 | 0.58±0.04 | 0.61±0.05 | 2.41 | 0.80±0.04 | 0.78±0.04 | 0.52 |
Anterior limb of internal capsule R | 1.18±0.08 | 1.22±0.08 | 2.78 | 0.50±0.04 | 0.58±0.05 | 24.28 |
0.73±0.04 | 0.79±0.05 | 17.92 |
Anterior limb of internal capsule L | 1.23±0.07 | 1.23±0.08 | 0.04 | 0.46±0.05 | 0.52±0.04 | 10.64 | 0.72±0.05 | 0.76±0.05 | 5.32 |
Precentral gyrus L | 1.27±0.07 | 1.24±0.06 | 1.72 | 0.46±0.02 | 0.50±0.03 | 20.89 |
0.73±0.03 | 0.75±0.03 | 3.88 |
External capsule R | 1.13±0.06 | 1.14±0.07 | 0.24 | 0.60±0.04 | 0.67±0.05 | 16.90 |
0.78±0.03 | 0.83±0.05 | 10.13 |
External capsule L | 1.23±0.05 | 1.28±0.09 | 4.05 | 0.44±0.04 | 0.50±0.04 | 22.17 |
0.70±0.03 | 0.76±0.04 | 26.65 |
*
Abbreviation. WM: white matter; CON: controls; IAD: Internet addiction disorder; Da: axial diffusivity; Dr: radial diffusivity; MD: mean diffusivity; R: right; L: left. SD: standard deviation.
For the 22 VOIs, Pearson correlation analysis demonstrated significantly negative correlations between FA values in the left genu of the corpus callosum and SCARED (r = −0.621,
To aid visualization, regions showing significant correlations (red) are thickened using the tbss_fill script implemented in FSL.
In this study, we used DTI to investigate the integrity of white matter in IAD adolescents by the observer-independent whole brain voxel-wise TBSS analysis. Compared with the age, gender and education matched controls, IAD subjects had significantly reduced FA in the orbito-frontal white matter, together with cingulum, commissural fibers of the corpus callosum, association fibers including the inferior front-occipital fasciculus, and projection fibers comprising the corona radiation, internal capsule and external capsule (
The orbito-frontal cortex has extensive connections with prefrontal, visceromotor, and limbic regions, as well as the association areas of each sensory modality
Anterior cingulate cortex (ACC) connects to the frontal lobes and the limbic system, playing an essential role in cognitive control, emotional processing and craving
Another major structure showing reduced FA in IAD subject is the corpus callosum, which is the largest white matter fiber tract connecting neocortex of the two hemispheres
Compared to controls, IAD subjects also showed significantly decreased FA in the anterior limb of the internal capsule, external capsule, corona radiation, inferior fronto-occipital fasciculus and precentral gyrus. Again, similar white matter abnormalities had also been observed in other forms of addiction. For example, white matter alterations in the anterior limb of the internal capsule and external capsule have been reported in alcohol abuse
Overall, our findings indicate that IAD has abnormal white matter integrity in brain regions involving in emotional generation and processing, executive attention, decision making and cognitive control. The results also suggest that IAD may share psychological and neural mechanisms with other types of substance addiction and impulse control disorders.
Although decreased FA is a well-established biomarker for impaired white matter integrity, its exact neurobiological meaning remains to be understood fully. FA of white matter fibers/bundles may be affected by many factors including myelination, axon size and density, path geometry, and extracellular water space between fibers
Behavioral acessment demonstrated that the IAD subjects had significantly higher scores on YIAS, SDQ, SCARED and FAD, compared to control. These findings are consistent with the results of previous neuropsychological studies on IAD subjects
In this study, we investigate the behavioral correlates of FA reduction in the affected brain regions in the IAD subjects. Reduction of FA in the left genu of the corpus callosum of the IAD subjects correlated significantly with increase of SCARED score; while higher YIAS scores appeared to be associated with more severely impaired white matter integrity in the left external capsule.
The SCARED is a reliable and valid self-report questionnaire that measures symptoms of anxiety disorders in children
In addition, the associations between white matter integrity and behavioral features indicate a novel potential target for treatment of IAD subjects, which is consistent with recent calls to focus on cognitive enhancement among addicted populations including IAD subjects
Our previous study showed that there was no white matter atrophy in the same cohort IAD subjects
There are several limitations that should be mentioned in this study. Firstly, the diagnosis of IAD was mainly based on results of self-reported questionnaires, which might cause some error classification. Therefore, the diagnosis of IAD needs to be refined with standardized diagnostic tools to improve the reliability and validity. Secondly, although we tried our best to exclude comorbid substance and psychiatric disorders, it is acknowledged that this may not have been done sufficiently (i.e., no urine test was given, sleep habits and schedules and daily sleepiness were not controlled in the experiment design), such that the white matter changes observed may not be attributed to IAD per se. It is also admitted that this is not a controlled study of effects of internet use on brain structure. Thirdly, the sample size in this study was relatively small, which might reduce the power of the statistical significance and generalization of the findings. Owing to this limitation, these results should to be considered preliminary, which need to be replicated in future studies with a larger sample size. Lastly, as a cross-sectional study, our results do not clearly demonstrate whether the psychological features preceded the development of IAD or were a consequence of the overuse of the Internet. Therefore, future studies should attempt to identify the causal relations between IAD and the psychological measures.
In conclusion, we used DTI with TBSS analysis to investigate the microstructure of white matter among IAD adolescents. The results demonstrate that IAD is characterized by impairment of white matter fibers connecting brain regions involved emotional generation and processing, executive attention, decision making, and cognitive control. The findings also suggest that IAD may share psychological and neural mechanisms with other types of impulse control disorders and substance addiction. In addition, the associations between FA values in white matter regions and behavioral measures indicate that white matter integrity may serve as a potential new treatment target for IAD, and DTI may be valuable in providing information on prognosis for IAD, and FA may be a qualified biomarker to assess the effectiveness of specific early interventions in IAD.
We thank the two anonymous reviewers for their constructive remarks and suggestions. We also thank the adolescent students and families who so willingly participated in this study.