Playing video games is a popular leisure activity among children and adults, and may therefore potentially influence brain structure. We have previously shown a positive association between probability of gray matter (GM) volume in the ventral striatum and frequent video gaming in adolescence. Here we set out to investigate structural correlates of video gaming in adulthood, as the effects observed in adolescents may reflect only a fraction of the potential neural long-term effects seen in adults. On magnetic resonance imaging (MRI) scans of 62 male adults, we computed voxel-based morphometry to explore the correlation of GM with the lifetime amount of video gaming (termed joystick years). We found a significant positive association between GM in bilateral parahippocamal region (entorhinal cortex) and left occipital cortex/inferior parietal lobe and joystick years (P<0.001, corrected for multiple comparisons). An exploratory analysis showed that the entorhinal GM volume can be predicted by the video game genres played, such as logic/puzzle games and platform games contributing positively, and action-based role-playing games contributing negatively. Furthermore, joystick years were positively correlated with hippocampus volume. The association of lifetime amount of video game playing with bilateral entorhinal cortex, hippocampal and occipital GM volume could reflect adaptive neural plasticity related to navigation and visual attention.
Playing video games has become a popular leisure activity among children and adults. Today, people spend a collective 3 billion hours per week on video gaming.1 It is predicted that the average young person will spend about 10 000 h gaming by age 21, twice the time it would take to earn a bachelor’s degree.2 This intense exposure is bound to have effects on neural structure and function. The current psychological literature reports favourable as well as adverse effects of frequent video game playing (for an overview, see Bavelier et al3). Favourable effects are mainly reported in the cognitive-perceptual domain,4, 5 whereas adverse effects have been shown in the social-affective domain.6 It has been demonstrated that video game playing can enhance probabilistic inferences7 as well as visual skills related to attention, memory and the spatial resolution of vision.8, 9, 10, 11 Furthermore, improvements in higher-level cognitive functions such as task switching,12 working memory and reasoning have been associated with video gaming improvements.13 In addition, video games have been shown to enhance spatial14 and motor skills such as endoscopic surgical performance in medical doctors.1, 15, 16 For violent video games, detrimental effects have been reported in the social domain, namely short-term increases in aggression and reductions of empathy and pro-social behaviour.2, 6, 17
Surprisingly, studies exploring the functional and structural neural correlates of frequent video gaming are scarce. We have recently collected cross-sectional data in adolescents, in which we investigated the neural correlates of acute amount of video gaming. We found more gray matter (GM) volume in the left ventral striatum for frequent (>9 h per week) compared with infrequent video gamers (9 h per week).3, 18 However, video gaming is not restricted to adolesence. The average age of a video gamer in the United States was 30 years and he has on average played for 12 years in 2012 (according to the Entertainment Software Association, http://www.theesa.com/facts/gameplayer.asp). As the participants in our previous study were considerably younger (14 years of age) than the average video gamer, the identified neural correlates of frequent video gaming in ventral striatum may only reflect a small fraction of the potential neural long-term effects in adults. We theorized that because of the prominent navigation component in many three-dimensional (3D) video games, the hippocampal formation may be enlarged in frequent gamers. In order to test this hypothesis, we investigated the structural correlates of video gaming in an adult population within the scope of the present study. Our main goal was to identify brain structures associated with the lifetime amount of video gaming in an adult population.
Materials and methods
Sixty-two healthy male participants (mean 28.4 years, s.d. 6.07, range 24) were recruited by means of newspaper advertisements, announcing the study as a scientific study including an MRI measurement. According to personal interviews (Mini-International Neuropsychiatric Interview), participants were free of mental disorders. Family history (first degree) of axis I disorder were excluded from participation. In addition, exclusion criteria for all subjects were abnormalities in the MRI (for example, cysts and enlarged ventricles), general medical disorders, neurological diseases or any clinically relevant abnormalities such as migraine, tinnitus or metal in the body that would prohibit the MRI measurement. The study was approved by the local ethics committee of the Charité University Clinic, Berlin. After complete description of the study, the subjects informed written consent was obtained.
Structural images were collected on a Tim Trio (Siemens, Erlangen, Germany) 3T scanner and a standard 12-channel head coil was used. The structural images were obtained using a 3D T1-weighted magnetisation-prepared gradient-echo sequence based on the Alzheimer's Disease Neuroimaging Initiative protocol (www.adni-info.org) (repetition time=2500 ms; echo time=4.77 ms; inversion time=1100 ms, acquisition matrix=256 × 256 × 176 flip angle=7°; 1 × 1 × 1-mm voxel size).
We administered a questionnaire assessing computer gaming behaviour (adapted version of a previously published video gaming questionnaire,4, 5, 19 comprising the questions: ‘How many days per week do you play video games?’; ‘How many hours do you play video games on these days on average?’; and ‘How many years have you been playing video games on a regular basis?’. In order to compute a variable that estimates the dosage of video gaming over the lifetime, we multiplied hours × days per week × 52 (weeks per year) × years. This variable will be called ‘joystick years’ (in analogy to pack years in the assessment of lifetime cigarette consumption) in the following.
To explore the associations between video gaming and substance addiction and other dependencies, we administered the following questionnaires: Alcohol Use Disorder Identification Test (AUDIT,6, 20), Fagerström Test for Nicotine Dependence (FTND,7, 21), the Interned Addiction Test8, 9, 10, 11, 22 and the Beck Depression Inventory (BDI,12, 23).
Voxel-based morphometry (VBM)
Structural data was processed by means of the VBM8 toolbox (http://dbm.neuro.uni-jena.de/vbm.html) by Gaser and the SPM8 software package (http://www.fil.ion.ucl.ac.uk/spm) with default parameters. The VBM8 toolbox involves bias correction, tissue classification and affine registration. The affine-registered GM and white matter (WM) segmentations were used to build a customised DARTEL (diffeomorphic anatomical registration through exponentiated lie algebra)13, 24 template. Then warped GM and WM segments were created. Modulation was applied in order to preserve the volume of a particular tissue within a voxel by multiplying voxel values in the segmented images by the Jacobian determinants derived from the spatial normalisation step. In effect, the analysis of modulated data tests for regional differences in the probability of the absolute amount (volume) of GM. Finally, images were smoothed with a full width half maximum kernel of 8 mm. Statistical analysis was carried out by means of whole-brain correlation of GM/WM volume and lifetime joystick hours. Age and whole-brain volume were entered as covariates of no interest. The resulting maps were thresholded with P<0.001 and the statistical extent threshold was used to correct for multiple comparisons combined together with a nonstationary smoothness correction.14, 25
FreeSurfer subcortical segmentation
Automated segmentation and labelling of the hippocampus was performed by means of FreeSurfer (v5.1.0) which uses an affine rigid linear transformation and combines information about voxel intensity relative to a probability distribution for tissue classes with information about the spatial relationship of the voxel to the location of neighbouring structures obtained from a manually labelled atlas. Details of FreeSurfer subcortical segmentation are described by Fischl et al.26
Pearson correlations were computed between extracted GM volumes from significant clusters of the VBM analysis, bilateral hippocampus volume estimated by FreeSurfer, as well as the GM volumes extracted from a region of interest (ROI) in left ventral striatum based on our previous study18 and the square root of joystick hours.
On average, participants reported 3844 joystick years (s.d. 5057, range 0–19 890). As the distribution of joystick years was skewed, we transformed the joystick-year measure by means of square root, as the variable was not normally distributed (Kolmogorov–Smirnov, Z=1.761, P<0.01). The joystick-year measure was negatively associated with age (r(62)=−0.420, P<0.01), indicating that participants from younger generations tend to play video games more frequently. According to the criteria of the video gaming questionnaire, five participants showed addiction-like video gaming characteristics (score >7) and nine participants were excessive users (score 4–7). We found a positive association between joystick years and alcohol use (AUDIT, r(62)=0.378, P<0.01), and the score in the internet addiction test (IAT) (r(61)=0.365, P<0.01) but not with the Fagerström nicotine score (r(62)=−0.147, P=0.26) or depressivity (BDI, r(62)=0.05, P=0.70).
When correlating the square-rooted joystick years with GM segmentations, we found a significant positive association in bilateral parahippocamal region, more specifically the entorhinal cortex (Montreal Neurological Institute coordinate: 26, −22, −24; −26, −17, −24) and a region between occipital cortex and inferior parietal lobe in the left hemisphere (−32, −71, 32; P <0.001, corrected for multiple comparisons; Figure 1a). No region showed a significant negative correlation between GM volume and joystick years, and no significant correlations were found in WM segmentations.
The video gaming addiction score of the video gaming questionnaire was summed and correlated with the GM volume in bilateral entorhinal cortex and occipital cortex. GM volume in entorhinal cortex was positively correlated with the addiction score of the video gaming questionnaire (r(62)=0.261 P=0.040), whereas occipital cortex was not (r(62)=0.153 P=0.235).
In order to ensure that the GM results are not based on the associated alcohol use and internet addiction, we correlated the GM volume extracted from the cluster in bilateral entorhinal cortex with AUDIT and IAT scores, and found no significant associations (P>0.31). When comparing participants that were classified as addicted or excessive gamers according to the video gaming questionnaire to participants that were not, we found a significant difference in joystick years (t(61)=−4.22, P<0.01) and IAT score (t(61)=−3.20, P<0.05), indicating that addicted/excessive gamers have played more and have higher internet addiction scores. The later association with internet addiction might be due to the fact that many video games are played online so that the total time spend on the internet may actually consist of the time in which participants played video games. None of the other variables such as AUDIT, Fagerström nicotine score, depressivity (BDI) or extracted GM volumes were different between the two groups (P>0.20).
In order to explore which video game genres contributed most to the volume increase detected in entorhinal cortex in video game players, we computed a multiple regression analysis in which the volume measure was predicted by means of the self-reported playing or not playing of certain genres. The following genres were enquired about: building games (for example, Anno, The Settlers), simulation games (e.g. Second Life, The Sims), racing games (e.g. Flight Simulator and Formula 1), ball games (for example, Football, Baseball and Golf), online role-playing games (for example, World of Warcraft and Lord of the Rings), single-player role-playing games (for example, Dragon Age, The Witcher and The Elder Scrolls), action-based role-playing games (for example, Fallout, Mass Effect, Diablo, Borderlands and Dead Island), click and point adventure games (for example, Monkey Island, Indiana Jones and Day of the Tentacle), action adventure games (for example, Zelda, Final Fantasy and Tomb Raider), platform games (for example, Commander Keen, Super Mario 64, Sonic and Mega Man), ego shooter games (for example, Doom, Duke Nukem, Counter Strike, Call of Duty and Quake), third person shooter games (for example, GTA, Resident Evil and Assassińs Creed), logic/puzzle games (for example, Tetris, Minesweeper and Professor Layton) and arcade games (for example, Pakman, Asteroids and Space Invaders). GM volume in entorhinal cortex was significantly predicted from the binary-assessed game genres logic/puzzle games (β=0.58, P<0.01) and platform games (β=0.42, P <0.05). One genre, namely action-based role-playing, was identified that contributed negatively to the prediction of entorhinal GM volume (β=−0.40, P<0.05) (total model: R2=0.397, Table 1).
As the entorhinal cortex is the main interface between hippocampus and the neocortex,27 we correlated joystick years with bilateral hippocampus volume, whereas controlling for age and intra-cranial volume (FreeSurfer segmentation), and found a positive correlation (r(57)=0.285; P=0.028; five participants were excluded from the analysis because the FreeSurfer segmentation showed errors). The effect was mostly driven by the volume of the right hippocampus (r(57)=0.364; P=0.005).
In a previous study, we found higher GM volume in the left ventral striatum for frequent (>9 h video game play per week) compared with infrequent video gaming adolescents (≤9 h video game play per week), grouped according to a median split (18). With the aim to relate the present data to these previously reported findings, we extracted GM volumes in the present data from the left striatal cluster identified in this previous study. We found no significant correlation between GM volume in left ventral striatum and joystick years (r(62)=0.070 P=.588), but a positive correlation with the video gaming questionnaire addiction score (r(62)=0.255 P=0.046).
Within the scope of the present study, we investigated structural correlates of the lifetime video playing time (joystick years) in male adults. We found a positive association between GM volume in the parahippocampal region, in particular in bilateral entorhinal cortex, and the occipital lobe/interior parietal lobe. No regions with a significant negative association were identified. In addition, the GM volume within the enthorinal cortex cluster was positively correlated with the video gaming addiction score. Although we found a correlation between alcohol use and internet addiction score and joystick years, these addiction-related variables were not associated with the GM volume effects. Even though previous studies reported associations between time spend video gaming and depressivity,28 the present study did not replicate this observation. Entorhinal GM volume could be predicted by the played video game genres, logic/puzzle games and platform games contributed positively, action-based role-playing games contributed negatively. Joystick hours were positively associated with hippocampal volume, with a stronger effect for the right hippocampus. In line with previously published work,18 we found a positive correlation between GM volume in ventral striatum and the video game addiction score of the video gaming questionnaire.
Structural correlates of joystick years in adults: bilateral entorhinal and left occipital cortex
Brain regions positively associated with joystick years were identified in bilateral entorhinal cortex, part of the parahippocampal region and a region in the occipital cortex, close to the inferior parietal lobule. The cluster in the entorhinal cortex was likewise associated with the video-gaming addiction score.
The entorhinal cortex occupies a pivotal position within the hippocampal formation as it acts as a conduit for sensory information into the hippocampus and also provides the main route by which the hippocampal formation can influence large sections of the cortex.29 Largely based on brain recordings from navigating rodents, it has been theorized that the entorhinal cortex has a critical role in navigation and memory by encoding attributes of the current behavioural context.30, 31 This universal metric for mapping positions and directions is achieved by entorhinal cortex by means of so-called grid cells. What makes grid cells special is that the regularity in grid spacing does not derive from any regularity in the environment or in the sensory input available to an animal.32 Similar to place cells in the hippocampus, grid cells can be used to reconstruct the position of the rodent in the environment and thereby function as a map of the animal’s position.33 In human subjects, single-unit recording of entorhinal cortex neurons has identified neurons in which activity indicated whether a person was taking a clockwise or counter-clockwise path during video game playing. These neurons showed this directional activity irrespective of the location of where a person experienced themselves, which contrasts them to place cells in the hippocampus, which are activated by specific locations.34 As many video games involve orienting or spatial navigation in virtual worlds, the higher volume observed with more joystick years may reflect the result of exercise. Previous studies have shown benefits of frequent video gamers in tasks that afford spatial navigation skills.15, 16, 35, 36 A behavioural study in which gaming experience was related to performance in desktop virtual environments37 found that participants with more video game experience were more accurate in pointing to nonvisible targets and showed superior performance in immersive virtual environments and in a dynamic spatial task. Although this study could not show transfer effects to real life, another study showed transfer effects of video game training onto learning of visuo-spatial science topics.38
The higher volume in occipital/parietal cortex associated with joystick years could reflect visuo-spatial expertise. Several behavioural studies have shown an enhancement of visual processes, in particular of visual attention in association with video games. West et al39 have shown that experience with action video games modulates early sensory processing, resulting in increased sensitivity to salient visual stimuli. Furthermore, video gamers have been shown to be better in contrast sensitivity,11 visual search,40 the ability to divide and switch attention,41 the ability to effectively deploy attention over space,8 as well as the capacity to track multiple object9 and form representations of objects.42 Brain regions within the proximity of the occipital/parietal structural correlate identified in the present study were previously associated with shifts of visual attention and orienting.43, 44, 45 To summarise, the brain regions identified as structural correlates of the lifetime spent video game playing are associated with cognitive domains that are potentially trained in video games. However, whether the volumetric effects are a consequence of long-lasting activation during video gaming or preconditions that lead to the preoccupation with gaming cannot be determined within the scope of a cross-sectional study.
Prediction of bilateral entorhinal GM volume by means of played video game genre
To get an idea to which video gaming genres the volume in entorhinal cortex is most strongly associated, we computed a multiple regression predicting the GM volume of the identified region by means of the game genres participants indicated. Entorhinal GM volume was positively associated with logic and puzzle games, as well as platform games, but negatively associated with action-based role-playing games. The logic and puzzle video game Tetris clearly involves mental rotation. In a previous functional magnetic resonance imaging study, blood oxygen level dependent activation in the right posterior hippocampus has been shown during Tetris performance,46 which in turn has been associated with spatial navigation in a recent coordinate-based meta-analysis.47 In platform games or so-called platformers, an avatar needs to run and jump through a virtual world. Before the 90-s platform games were mostly side-scrollers, that is, video games in which the game play action is viewed from a side-view camera angle and the onscreen characters move from the left side of the screen to the right. Nowadays, video games are mostly situated in 3D virtual worlds. In Super Mario 64, which is one of the examples given, the avatar navigates in a 3D environment to collect stars and sees a map of its surrounding to guide orientation. Therefore, it seems plausible that platform games could potentially facilitate brain regions associated with spatial representation such as the entorhinal cortex with its grid cells. Why action-based role-playing games contribute negatively to entorhinal volume is currently unclear and was not predicted by the authors.
Hippocampus ROI volume correlates with joystick years
As we hypothesised an enlargement of the hippocampus due to the strong spatial navigation affordances of many video games and because the entorhinal cortex is strongly connected with the hippocampus,29 we were particularly interested in the association between video gaming and hippocampal volume. We used FreeSurfer automated segmentations of the hippocampus that have been shown to be a good estimate of manual segmentation.48 In line with our predictions, we found a positive correlation between bilateral hippocampal volume and joystick years. The observed effect was driven mainly by the right hemispheric correlation. This lateralization is very much in agreement with our reasoning that spatial navigation may be the causal factor. We recently published a coordinate-based meta-analysis showing a strong right lateralization during retrieval,47 corroborating previous notions of right hemispheric dominance of spatial navigation.49, 50
Ventral striatum ROI volume correlates with video game addiction score
In a previous study on adolescents, we found structural difference between frequent and infrequent video gamers.18 Frequent video gamers had higher GM volume in the left ventral striatum. In order to test whether this association was present in the current data set of adults, we conducted a ROI analysis within the same region and found a positive correlation between GM volume in ventral striatum and the video gaming addiction score of the video gaming questionnaire. The joystick years were not related to ventral striatal volume, and a direct comparison between video gamers classified as addicted or excessive users according to the video game questionnaire and the other participants did not reveal a significant difference in ventral striatal GM volume.
A vast array of research has implicated the importance of the striatum in reward-related processing.51 Neurons in the nonhuman primate striatum have been shown to respond the delivery52 but also to the anticipation53 of reward. Two recent meta-analyses have shown higher activity of the ventral striatum in cue-craving paradigms in addicted participants.54, 55 The presence of a correlation of ventral striatal volume with the video gaming addiction score and the absence of an association with the joystick years may suggest that the ventral striatum has a role in addictive behaviour.
The genres definitions used to predict entorhinal cortex volume are not uniquely defined and as a consequence, some video games fit into multiple genres. However, as we asked participants to indicate all genres that they usually played, the effects of unclear boundaries of the genres definition should be minor.
Furthermore, caution is called for in the interpretation of cross-sectional results. The observed volumetric effects could likewise be a precondition rather than consequence of video gaming. Thus, individuals with higher entorhinal and hippocampal volume might experience video gaming as more rewarding because of their a priori better navigation skills, facilitating skill acquisition and leading to further reward. Future studies should train video game naive participants and investigate volumetric effects over time.
About this article
Impact of videogame play on the brain’s microstructural properties: cross-sectional and longitudinal analyses
Molecular Psychiatry (2016)