As related industries, such as smart devices, mature, the Internet has become an essential tool for learning, working, and playing. According to Internet World Stats (IWS) (1), as of March 2021, the number of Internet users has reached 5.17 billion, with Asia accounting for 53.4% of the world's Internet users. Such a high Internet penetration rate has brought a severe social problem, namely Pathological Internet Use (PIU). The PIU refers to an inability to control one's Internet use that adversely affects daily life, also known as “Internet Addiction (IA)” (2–4). The global average prevalence is about 7.02% and is still on the rise (5). Previous studies have suggested that PIU has a similar neuropathological basis to substance addiction (6, 7). Although more attention has been paid to PIU, the pathological mechanism of PIU is still unclear.

Functional MRI (fMRI) is a common neuroimaging technique to explore the neuropathological mechanism of diseases (8). Some fMRI indicators based on whole-brain analysis, such as functional connectivity (FC), the amplitude of low-frequency fluctuation (ALFF), regional homogeneity (ReHo), functional connection density (FCD), voxel-mirrored homotopic connectivity (VMHC), and cerebral blood flow (CBF), provided technical support for the comprehensive exploration of spontaneous neural activity in the brain of subjects with PIU. Previous studies have suggested that the limbic system, which is involved in the reward-processing circuit, and the prefrontal lobe, which is involved in the cognitive control circuit, are the physiological basis for the formation of PIU (9–12). However, these findings have been controversial in resting-state fMRI studies. For example, some researchers found increased spontaneous brain activity in the superior frontal gyrus (SFG) of subjects with PIU (13), while others found a decreased neural activity in this region (14, 15). In addition, the neural activity in superior temporal gyrus (STG) of subjects with PIU in resting-state was found to be increased (13), while another research found it decreased (16). The inconsistent results might be related to the differences in sample size, imaging analysis indexes, and demographic characteristics. These differences make it difficult to understand the neural mechanism of PIU, so further quantitative exploration is needed.

A systematic and quantitative meta-analysis, such as signed differential mapping (SDM) analysis, can find consistent local resting-state abnormalities regardless of all method differences (17). The SDM meta-analysis method could address heterogeneity between studies by reconstructing positive and negative graphs in the same image, thereby counterbalancing the effect of reporting findings in opposite directions (18). This voxel-based neuroimaging meta-analysis method has been validated in psychiatric disorders, such as depression, autism, and behavioral addiction (19–21). Moreover, researchers have used this method to find consistent activation results of brain regions in subjects with PIU in task-state fMRI (22–24). However, the changes of spontaneous brain activity of PIU in the resting state remain to be explored.

Hence, in this study, we aimed to unearth the consistency of changes in spontaneous brain functional activity in subjects with PIU during resting state. The findings of this meta-analysis will help to understand the pathological basis of PIU and give evidence for PIU prevention and intervention in the future.

This study was reported according to preferred reporting items for systematic review and meta-analysis (PRISMA) guidelines (25) and registered on PROSPERO (registration No: CRD42021258119).

A total of 781 articles were retrieved. After the screening procedure, 10 studies (13–16, 32–37) met eligibility criteria and were included in the systematic review (see Figure 1).

This study is the first to use quantitative SDM meta-analysis to integrate abnormal neural activity in resting-state fMRI studies of subjects with PIU. We found that local neural activity of STGtp, AMY, MCC, and IN increased in the subjects with PIU, while the local neural activity of SFGdl and MFG decreased. These abnormal brain regions were closely related to the cognitive executive control and emotional regulation functions of the subjects with PIU. Our findings provided a reference for exploring the pathological mechanism of PIU.

Executive control dysfunction is a common feature in the subjects with PIU (39). In our study, multiple brain regions are involved in the executive control function of the brain, mainly including STGtp, SFGdl, and MFG. The STGtp is an important component of the temporal lobe, located in front of STG, and participates in a variety of important cognitive functions (40). Studies have found that the patients with cognitive impairment had reduced STG cortical thickness and increased neural activity compared with normal subjects (41–43). Besides, a recent neuroimaging meta-analysis of PIU has shown that STG was hyperactivated in the executive control tasks, suggesting that STG was involved in the executive control function of the subjects with PIU (24). These studies suggested that increased neural activity of STGtp in the subjects with PIU in the resting state may be a compensatory mechanism for executive control function deficits in the subjects with PIU.

The SFGdl and MFG were also important to brain regions for the executive control function. These regions were located in the frontal cortex, associated with the executive control network, and regulated other cognitive functions (44–46). A previous meta-analysis found that the subjects with PIU showed abnormal frontal cortex activation in multiple cognitive tasks (22). A recent study has revealed that the reduced SFG gray matter volume played a mediating role in the influence of emotional reflection in PIU (47). Other studies found that abnormal activation of SFG and MFG might be involved in the control of impulsivity in the subjects with PIU (48, 49). In our study, the decreased neural activity in SFGdl and MFG in the subjects with PIU may indicate the decreased control ability of the subjects with PIU to impulsivity.

Mood changes are another common complication in the subjects with PIU (50). Our meta-analysis showed that there were brain regions associated with emotion regulation, such as the AMY. This finding is consistent with previous reports of AMY dysfunction in PIU (51, 52), as well as with other addictions (53, 54). The AMY is an important component of the basal ganglia and is involved in integrating and processing information about emotions and rewards (55–57). Previous studies have suggested that AMY could regulate both positive and negative emotions (58). Subsequent studies have found the role of the central nucleus of the AMY in negative emotions accompanying reward loss (59) and the effect of the glutamate pathway from basolateral AMY to nucleus accumbens (NAc) on controlling reward-seeking behaviors (60). Therefore, the increase in spontaneous neural activity in AMY reflects the higher susceptibility to mood changes in the subjects with PIU and abnormalities in the reward system in the brain of the subjects with PIU.

Additionally, abnormalities in the salience network (e.g., IN, MCC) were found in the subjects with PIU. The salience network is a neural system for perceiving and responding to homeostatic demands and is closely related to diseases such as addiction and depression (61–63). The IN and cingulate are the key nodes of the Salience network (64). Previous studies have found that subjects with PIU have increased cortical thickness in IN (65), while decreased gray matter density in IN and MCC than HCs (66–68). These findings suggest structural abnormalities in the salience network of the subjects with PIU, which may lead to functional problems of the salience network. A recent meta-analysis of behavioral addiction has also found that neural activity in MCC increased when behavioral addicts were exposed to addiction-related cues (69). Furthermore, neural activity in the IN significantly increased when the subjects with PIU were exposed to game cues (70, 71). Therefore, the salience network abnormalities we found in the subjects with PIU were related to their excessive craving for addictive cues.

Furthermore, our study confirmed that PIU and addiction had similar neuronal activation patterns. The results of a systematic review of addiction showed that the changes of spontaneous neural activity in frontal and temporal regions were found in both substance addiction and behavior addiction (72, 73). These abnormal functional activity changes in brain areas were mainly related to emotion and cognitive control. This suggested that both PIU and addiction showed functional impairments in brain regions associated with cognitive and emotional processing. The difference was that the patients with substance addiction showed spontaneous changes in neural activity in the striatum in their resting state (72). Although the striatum was also found to be involved in the processing of rewards in the subjects with PIU during task fMRI studies (74, 75). However, in our study, no consistent neural activity was found in the striatum of the subjects with PIU in the resting state. Moreover, these findings were indirect rather than direct comparisons of neuronal activity patterns between PIU and addiction. More research is needed in the future to explore differences in neuronal activity patterns between PIU and other addictive disorders.

There are several limitations to this study. The first is that our study only included fMRI data from the whole-brain analysis, not data based on the region of interest (ROI) analysis. As different researchers choose different ROIs, this also increases the difficulty of data combination. Thus, whole-brain analysis studies avoid inconsistencies in the choices of ROI to researchers. Second, the studies we included used different analysis methods and imaging modalities. Hence, the heterogeneity due to method differences cannot be completely excluded. For example, the results of FC, FCD, and ICA analysis can reflect the connections between different brain regions or brain networks. The results of the ReHo and ALFF methods can reflect the characteristics of local neural activity in the brain (76, 77). The CBF is the main indicator of arterial spin labeling (ASL). The ASL is also an fMRI technique, which can reflect the brain metabolism and neural activity of subjects in the resting state (78). However, the research results of different modes and analytical methods could reflect the situation of a certain field more comprehensively. In addition, previous fMRI meta-analysis studies also combined fMRI studies with different analysis methods (79, 80). Therefore, when the number of studies using a single analysis method is not enough, it is feasible to combine the results of fMRI studies using multiple analysis methods. Second, due to the small number of included studies, the results might be subject to random error. It is worth noting that the heterogeneity test showed low heterogeneity in the results of this study. Sensitivity analysis also suggested that the results of this study were highly reproducible. Therefore, our study could be used as a preliminary study to reflect the characteristics of spontaneous neural activity changes in the brain of the subjects with PIU in resting state. Third, PIU may have some comorbidities (such as anxiety and depression disorders) that were not considered in our study. Due to the small number of studies that we included, a subgroup analysis was not possible. Finally, the cases in our study were all Asian. Therefore, our results can only represent the neuroimaging characteristics of the subjects with PIU in Asia. We also look forward to further neuroimaging studies of PIU in more countries.

Our study identified consistent changes in brain regions in subjects with PIU from different fMRI studies. The subjects with PIU showed abnormal functional activity in brain regions and functional brain networks involved in cognitive executive control and emotional regulation, which constitute the core symptoms of PIU. These consistent changes in brain regions may provide important targets for the future diagnosis and intervention of PIU.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

WP, SY, HL, and TZ designed this study. WP, QH, HG, JW, and YT were involved in the process of literature selection, data collection, and quality assessment. WP performed the data analysis and wrote the manuscript. SY, YW, HL, and TZ critically revised the work. All authors have approved the final manuscript.

This work was supported by the National Natural Science Foundation of China (81072852 and 81574047), the Key R&D Project of Sichuan Province (2019YFS0175), the Xinglin Scholars Scientific Research Promotion Program of Chengdu University of Traditional Chinese Medicine (XSGG2019007), and the Training Funds of Academic and Technical Leader in Sichuan Province.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.