Brain anatomy alterations associated with Social Networking Site (SNS) addiction | Scientific Reports

image

Abstract

This study relies on knowledge regarding the neuroplasticity of dual-system components that govern addiction and excessive behavior and suggests that alterations in the grey matter volumes, i.e., brain morphology, of specific regions of interest are associated with technology-related addictions. Using voxel based morphometry (VBM) applied to structural Magnetic Resonance Imaging (MRI) scans of twenty social network site (SNS) users with varying degrees of SNS addiction, we show that SNS addiction is associated with a presumably more efficient impulsive brain system, manifested through reduced grey matter volumes in the amygdala bilaterally (but not with structural differences in the Nucleus Accumbens). In this regard, SNS addiction is similar in terms of brain anatomy alterations to other (substance, gambling etc.) addictions. We also show that in contrast to other addictions in which the anterior-/ mid- cingulate cortex is impaired and fails to support the needed inhibition, which manifests through reduced grey matter volumes, this region is presumed to be healthy in our sample and its grey matter volume is positively correlated with one’s level of SNS addiction. These findings portray an anatomical morphology model of SNS addiction and point to brain morphology similarities and differences between technology addictions and substance and gambling addictions.

Introduction

Notwithstanding the positive impacts of technologies on humans, technology-related addictions seem to be fairly prevalent12; A recent meta-analysis suggests that globally the prevalence rate is about 6% and that it varies by country, ranging from 2.6% to 10.9%3. While the negative outcomes of such addictions may not always be as devastating as those generated by severe substance addictions, they attack the vulnerable population of adolescents and young-adults45 and can have a myriad of negative effects on individuals’ work, school and social functioning, wellbeing and psychological states2, as well as on their sleep hygiene and long-term cardio-metabolic health5. Therefore, these addictions have been recognized as an important topic that merits further research6 and the fifth edition of diagnostic and statistical manual for mental disorders has included the concept of “Internet Gaming Disorder” in the appendix (section 3, potential disorders requiring further research)7. Conceptual psychological-neurobiological models8 as well as functional brain imaging studies9 suggest that such addictions involve an interaction of sensitized reward processing and cue-reactivity with diminished prefrontal inhibitory control. Yet, more research is needed for understanding the structural neural underpinnings of this phenomenon10. Specifically, even though addictions are recognized as “brain diseases” by the American Medical Association, little is known regarding potential brain structural alterations associated with such addictions; this knowledge can help researchers and medical practitioners develop interventions for preventing or treating such addictions.

As such, this study seeks to examine potential brain structural alterations associated with an important instance of technology addictions, namely addiction to a social networking site. Social Networking Site (SNS) addiction is a subcategory of the technology/Internet spectrum of addictions11 and is defined as a user’s maladaptive psychological state of dependency on the use of an SNS, which is manifested through an obsessive pattern of seeking and using this SNS such that these acts infringe normal functioning and produce a range of typical behavioral addiction symptoms, including salience, withdrawal, relapse, growing tolerance, conflict and mood modification12. While there is stronger consensus regarding the prevalence of maladaptive technology use patterns which result in addiction-like symptoms12, it is not clear yet if the term “addiction” is best, and whether other terms such as “use disorder” may be more appropriate. This study, however, uses the term “addiction” in line with prior research in this field, even though the medical community still debates if this term is appropriate6. Furthermore, in line with this line of work13 we treat addiction as a continuous concept, i.e., we capture the level of addiction-like symptoms all people have, rather than trying to medically classify people as addicts or non-addicts using non-established criteria.

This study specifically focuses on brain anatomy modulations in terms of the grey matter volumes (GMV; see glossary of neuroscience terms in Appendix A) of brain regions, which are arguably associated with SNS addiction and are flexible or prone to anatomical modulations. These alterations are presumed to take place in central and necessary regions of the dual-system which governs behavior14, the deficiency of which is associated with addictions15. These regions are: (1) the Nucleus Accumbens (NAc), which has been implicated in playing a primary role in addictive behaviors through the processing of rewards that motivate behavior, including problematic behaviors; (2) the amygdala, which has been implicated in playing a key role in triggering impulsive behaviors from conditioned cues; presumably by linking environmental cues to neural systems involved in negative reinforcement (e.g., the relief from an aversive condition such as withdrawal), as well as positive reward and reward expectancy, such as those mediated by the NAc16; and (3) the midcingulate cortex (MCC), i.e., the dorsal region of the anterior cingulate cortex (ACC), which is involved with self-control or inhibition processes in response to impulsions generated through the impulsive system. The glossary in Appendix A provides details regarding these neural substrates.

Addiction is often initiated by hyperactivity of the system that assesses rewards17 and drives impulsive behaviors15. This includes the NAc, the key substrate where mesolimbic dopamine is released, and reward seeking behavior is elicited, and it also includes the amygdala, which is thought to link environmental cues to reward systems in the striatum, including the NAc. This system can become over-sensitized through repetitive enactment of a rewarding behavior and recurring strong intrinsic rewards, which can lead to a constant state of “wanting” to enact the addictive behavior18. The NAc is a central and necessary component of this reward system19, but the amygdala has also been argued as a necessary component of a broader neural system underlying automatic, habit, and impulsive behaviors152021. Hence, addictions are typically advanced by hyperactivity of the extended amygdala circuit which includes the NAc and amygdala16. Many subcortical reward-system regions10, as opposed to cortical regions, are morphologically flexible and can easily adjust to new environmental demands22. Hence, it is reasonable to assume that addiction-associated morphology changes (see glossary of neuroscience terms in Appendix A), if exist; can apply to the NAc and amygdala.

Oftentimes, the increased efficiency of the extended amygdala (reward) system is manifested through pruning wasteful and redundant neurons, and specifically reducing the GMV of the amygdala such that lean, fast and competent, bundles of neurons are retained. Achieving higher performance through pruning is very common23 and is especially relevant in subcortical areas24. It should be noted that while grey matter volume reduction changes to such regions are similar across addictions1920, the processes that lead to such changes may differ between addictions. In many cases, substances such as cocaine, which bind to dopamine receptors, create direct neurobiological changes in the operation and GMV of such brain regions25. In behavioral addictions, in contrast, such as addiction to SNS or videogame use, the implicated systems are typically affected indirectly, by environmental behaviors2627, through changing the work demands imposed on these brain regions, e.g., through increasing the need for reward or task-conflict processing and the resultant natural brain adaptations28.

Regardless of the process, negative associations between the GMV of the (typically bilateral) amygdala and other addictions have been observed in both substance and behavioral addictions, including for example in cases of abuse of cannabis29, alcohol30, cocaine31, prescription opioids32, as well as in problematic behaviors such as gambling33. Given possible neural and behavioral similarities between other addictions and technology-related addictions34, and the shared neural basis of different addictions21 including behavioral ones27, it is reasonable to expect that such negative associations also exist in the cases of SNS addiction. We hence hypothesize that (H1) the grey matter volume of the amygdala will be negatively associated with one’s SNS addiction score; after controlling for age, gender, number of contacts on the SNS, SNS use frequency, years of experience with the SNS, and the whole brain GMV. We suggest controlling for demographic and SNS use variables to ensure that the observed variation in GMV is associated with addiction per-se. We also suggest cleaning any variance in GMV which may be attributed to general brain volume of grey matter, across regions, which may differ from one individual to another and influence the GMV of the examined regions of interest (ROIs) regardless of addiction.

While the NAc is a central and active region in all addiction phases16, the existence and direction of possible structural differences in the NAc in relation to addictions are not clear. Some studies, for example, show GMV reduction in right NAc in alcoholism cases30 or left NAc in heroin-dependent patients35; whereas others show increased GMV of left NAc in cannabis users29 and frequent video-gamers36. Some studies, albeit focusing on connectivity, did not find correlations of NAc connectivity with sharing of self-related information on social media37. Given these mixed findings, and also the fact that the NAc is anatomically difficult to define with precision on scan images, we refrain from hypothesizing about the existence and direction of structural differences in the NAc. Nevertheless, given the centrality of the NAc in reward processing, including in the case of social media use38 we explore post-hoc whether structural differences in the NAc are associated with SNS addiction.

In addition to the abovementioned hyperactivity of the impulsive/reward assessment brain system, addictions typically also involve hypo-activity of the reflective or inhibition brain system15. This hypo-activity is often reflected in these areas of the brain through reduced grey matter394041. The ACC/MCC is of particular interest since it is relevant for weak inhibition abilities and consequent addictions; and the grey matter morphology of the ACC/MCC has been linked to addictive and excessive behaviors42. In such cases the ACC/MCC changes typically manifest through reduced GMV of this region, e.g., among methamphetamine users39, excessive-eaters40, cocaine users41, and Internet addicts43. Hence, if SNS addiction is similar to other addictions, and given that addictions are largely similar in their neural roots21 it is reasonable to assume that it is negatively associated with the GMV of the ACC/MCC, possibly reflecting lower efficiency of this inhibition system region. We hence hypothesize that (H2) the grey matter volume of the ACC/MCC will be negatively associated with one’s SNS addiction score; after controlling for age, gender, number of contacts, SNS use frequency, years of experience with the SNS, and the whole brain GMV.

Results

The addiction scale was statistically valid and reliable (α = 0.91, composite reliability = 0.93, and Average Variance Extracted = 0.63). Hence, its average represented participants’ presumed addiction levels or at least their levels of addiction-like symptoms. Addiction scores (average = 2.200, SD = 0.718, Range = 1.071 to 3.653 on a 1–5 Likert type scale) as well as all extracted GMVs did not deviate from normality (Kolmogorov-Smirnov and Shapiro-Wilk test statistics all with p-values > 0.20). Hence, correlational analyses were deemed to be appropriate. Addiction scores were not significantly correlated with age, years of Facebook experience, and number of contacts; but correlated with sex (r = 0.44, p < 0.05) and frequency of use (r = 0.74, p < 0.000). Hence, women in our sample had higher addiction scores than men had, and higher addiction scores were, as expected, associated with more frequent use of Facebook.

Voxel-wise based morphometry (VBM, see glossary in Appendix A) analyses revealed that SNS addiction scores, as hypothesized, negatively correlated with GMV in the bilateral amygdala (left amygdala, local maxima in MNI coordinates x, y, z = −30, −8, −18, geometric center x, y, z = −24.3, −5.6, −17.4; right amygdala, local maxima x, y, z = 30, 0, −14, geometric center x, y, z = 23.7, −3.5, −18.0), after accounting for control variable effects (see Table 1 and Fig. 1). Hence, H1 was supported. The VBM analyses, however, revealed, in contrast to our expectation, that the SNS addiction score was positively (and not negatively) correlated with GMV in the ACC/MCC (local maxima x, y, z = 4, −8, 34, geometric center x, y, z = 3.3, 0.2, 36.1), after accounting for control variables (see Table and Fig. 1). Hence, H2 was not supported. In addition, voxel-wise analysis did not show significant correlation between addiction scores and GMV in bilateral NAc.

Figure 1: Visualization of the voxel-wise VBM results*.

  • Voxel-wise VBM results illustrated in three different views: rendered brain (A), coronal view (B), and sagittal view (C). The SNS addiction score was negatively correlated with GMV in bilateral amygdala (blue areas) and positively correlated with GMV in the anterior/mid cingulate cortex (ACC/MCC, yellow area). Slices are displayed in radiological view (right is on the viewer’s left). Scatter plots (D–F) show the pattern of correlation between GMV (D): ACC/MCC; (E): left amygdala; (F): right amygdala) and SNS addiction score.

To supplement the voxel-wise analysis, additional theory driven ROI analyses were performed. The average GMVs in five anatomically defined ROIs (bilateral amygdala, ACC/MCC, and bilateral NAc) were extracted and partially correlated with the addiction score. Results suggested that addiction scores were negatively correlated with the left and right amygdala volumes (r = −0.67, p < 0.01 and r = − 0.65, p < 0.01, respectively) and positively correlated with MCC volumes (r = 0.57, p < 0.01). Left and right NAc volumes did not significantly correlate with addiction scores (p < 0.52 and p < 0.76, respectively). In order to alleviate possible power concerns, the analyses were repeated without the control variables and produced similar results. In order to further examine the non-significant association of NAc volumes with addiction scores, cross-validation with 100 re-samples was performed. It produced non-significant confidence intervals for this association (Left: −0.587 to 0.340; right: −0.520 to 0.344). Thus, while it is possible that the non-significant NAc associations are due to low power, our findings indicate that there may also be other reasons for this.