Skip to main content
NCBI home page
Journal of Behavioral Addictions logoLink to Journal of Behavioral Addictions
. 2025 Aug 16;14(3):1304–1314. doi: 10.1556/2006.2025.00061

Neural responses to kills/deaths in real MOBA games are associated with addiction-related psychological traits and subjective pleasant/unpleasant experiences

Dong-Dong Zhou 1,†,*, Hong-Zhi Li 1,†, Jia-Jia Yang 2, Zhen Lv 1, Li-Yang Wan 1, Wo Wang 1, Li Kuang 1,3,**
PMCID: PMC12486282  PMID: 40818061

Abstract

Background and aims

Players of Multiplayer Online Battle Arena (MOBA) games are at a heightened risk of developing Internet Gaming Disorder (IGD). We aimed to investigate the neural responses triggered by kills and deaths during real MOBA gameplay and explore their association with addiction-related psychological traits and subjective pleasant or unpleasant experiences.

Methods

We developed an experimental protocol to capture moments of kills and deaths during real MOBA gameplay. Game players who frequently play “Honor of Kings” or “League of Legends” on mobile phones for at least 12 months were recruited. All participants completed the IGD-20 assessment and played six matches while concurrently recording electroencephalography with a 64-channel setting.

Results

Among males, the amplitude of P300 induced by kills and deaths showed a significant negative correlation with IGD-20 scores. Additionally, the post-death beta-band event-related synchronization (ERS) and theta-band event-related desynchronization (ERD) were significantly positively correlated with IGD-20 scores in males, while no significant correlation was found among females. The amplitude of P300 following deaths was significantly negatively correlated with subjective unpleasant experiences in males, yet positively correlated in females. Regardless of gender, the amplitude of post-kill P300 and beta-ERS were negatively correlated with pleasant experiences.

Discussion and Conclusions

This study establishes an ecologically-embedded paradigm that successfully captures real-time neural signatures of valenced in-game events in MOBA games. We offer a novel methodological framework and theoretical perspective for investigating IGD and will inspire further research into the neural mechanisms underlying IGD in naturalistic gaming environments.

Keywords: MOBA game, internet gaming disorder, ERP, neural oscillation

Introduction

In recent years, video gaming has gained immense popularity worldwide. Among the various genres, Multiplayer Online Battle Arena (MOBA) games such as “Honor of Kings” and “League of Legends” have emerged as the predominant favorites (Nuyens et al., 2016). While MOBA games are highly engaging and offer several benefits, including stress relief, mood modification, and improved executive functions (Li, Huang, Li, Wang, & Han, 2020), there are concerns that MOBA gamers might be more susceptible to problematic gaming behaviors and at a higher risk of developing Internet Gaming Disorder (IGD) (Han, Jeong, Jo, Son, & Yim, 2020; Männikkö, Ruotsalainen, Tolvanen, & Kääriäinen, 2019).

The reported average prevalence of IGD among adolescents and young adults is 9.9% (Gao, Wang, & Dong, 2022). However, this figure may be inflated due to significant heterogeneity observed across IGD prevalence studies, which is largely driven by methodological variations in measurement tools and sampling biases (Fam, 2018; Gou, Zhang, Tang, Zhang, & He, 2024; Zhou et al., 2024). Notably, the choice of diagnostic criteria, especially when comparing the ICD-11 and DSM-5 frameworks, and the selection of assessment scales have been shown to produce substantial discrepancies in reported prevalence rates (Zhou et al., 2024). Both diagnostic systems emphasize three core clinical features: a) loss of control over gaming behavior, b) prioritization of gaming over other life activities, and c) persistent engagement in gaming despite negative consequences, while requiring documented functional impairment in personal, familial, or occupational domains. A critical distinction lies in their nosological status: The DSM-5 currently categorizes IGD as a condition requiring further study, whereas the ICD-11 formally recognizes gaming disorder as a distinct behavioral addiction (American Psychiatric Association, 2013; World Health Organization, 2019).

Given the widespread engagement with online games and their captivating features, it is crucial to delve deeper into the neural mechanisms behind players' intense involvement. Previous research has highlighted significant findings, noting changes in brain structure and function in individuals with gaming disorder, primarily in areas related to impulse control, reward processing, and emotional regulation (Weinstein, Livny, & Weizman, 2017; Yao et al., 2017). However, these results have been inconclusive and exhibited considerable heterogeneity (Palaus, Marron, Viejo-Sobera, & Redolar-Ripoll, 2017), potentially due to the diversity of games played and methodologies used across studies.

Previous research has utilized experimental paradigms like delayed discounting tasks, card-guessing tasks, and gambling tasks to study reward processing in individuals with IGD (Raiha et al., 2020; J. Zhang, Hu, Wang, Wang, & Dong, 2020), often employing monetary rewards to explore the neural mechanism involved (Balodis & Potenza, 2015). However, the neural reactions to monetary rewards or punishments likely differ significantly from those elicited by rewards or punishments encountered during real gameplay. An intriguing study demonstrated that the neural responses of the striatal reward circuitry are more vigorously activated during win and loss events when participants are actively engaged in gameplay compared to passively watching games (Kätsyri, Hari, Ravaja, & Nummenmaa, 2013). This finding underscores the importance of active engagement in gameplay for understanding the neural underpinnings of gaming disorder, as the neural responses to winning and losing differ markedly between active participation and mere observation. Furthermore, another study introduced a semi-natural gaming design that allowed participants to experience in-game rewards, potentially providing insights into the neural responses under game-specific rewards rather than monetary incentives (Duven, Müller, Beutel, & Wölfling, 2015). However, this design did not specifically target the games to which the individuals with IGD were devoted, possibly failing to capture the neural responses to reward stimuli within their addictive games. Thus, previous studies have generally exhibited lower ecological validity. As a result, the precise neural responses of gamers to rewards or punishments in the games to which they are devoted, which are crucial for understanding the neural mechanisms of deep gaming engagement, remain unclear.

The rewarding stimuli during real MOBA gameplay include each ‘kill’ of an enemy and ultimate victories, while the punishing stimuli include each ‘death’ and ultimate defeats. In this study, we aim to examine the neural responses to in-game kills and deaths and their associations with addiction-related psychological traits and subjective pleasant/unpleasant experiences.

Methods

Participants

MOBA gamers actively playing either Honor of Kings or League of Legends on mobile platforms for a minimum of 12 months, with an average of at least 5 gaming sessions per week, were enrolled. Exclusion criteria comprised: a) history of cranial trauma or neurosurgical procedures, b) current or lifetime diagnosis of neurological/psychiatric disorders, and c) presence of chronic systemic diseases.

Assessments

Two trained researchers (Hong-Zhi Li, M.Psych.; Zhen Lv, M.D. candidate) conducted initial screening using the Mini International Neuropsychiatric Interview (MINI, Chinese Version 5.0.0). Subsequent evaluation employed the 20-item Internet Gaming Disorder Test (IGD-20) to examine gaming-related psychological traits (Pontes, Király, Demetrovics, & Griffiths, 2014). The IGD-20 inventory measures six dimensions: salience, mood modification, tolerance, withdrawal symptoms, conflict, and relapse. Participants rated each item on a 5-point Likert scale (1 = Strongly disagree; 5 = Strongly agree), yielding total scores ranging 20–100. Higher IGD-20 total scores reflect stronger tendencies toward gaming-related impairments, with scores exceeding 71 points (per established cutoffs) suggesting potential clinical diagnosis of IGD (Pontes et al., 2014).

Procedure

All participants used the same designated Android phone (Xiaomi 11 Pro) to play the game. Each participant logged into their gaming account for play. Participants primarily using iOS devices received Android-based gaming accounts with equivalent rank tiers to maintain gameplay parity across platforms. During the experiment, each participant played six matches of MOBA games. The game audio was played aloud to immerse participants and replicate a typical gaming scenario as closely as possible. The gaming phone used a 5G network provided by Telecom.

In this study, we developed an experimental protocol that enabled precise marking of in-game events (Fig. 1), such as kills and deaths, during real MOBA gameplay, along with synchronous electroencephalography (EEG) recording (refer to Fig. 1a and b). We mirrored the game screen from the phone to the EEG recording computer using EVScreenMirror (Hunan Yiwei Information Technology Co., Ltd) via a USB connection. The mirrored display operated at 60 frames per second (FPS). A real-time display of the mirroring delay was shown in the top-left corner of the mirrored window. The mean timing error is estimated at 6.84 ms (see Fig. 1c and d). Using EVCapture (Hunan Yiwei Information Technology Co., Ltd), we simultaneously recorded both the EEG and game screens. These recordings were set at a frame rate of 60 FPS and saved in the “.avi” format. Screen recording was initiated first, followed by EEG recording, and then the participant began playing the game. After completing a match, EEG recording was stopped before screen recording was halted. Participants were allowed short breaks between matches.

Fig. 1.

Fig. 1.

Open in a new tab

Our proposed experimental scheme and its time precision for marking. (a) Schematic of the experimental setup; (b) Phone screen mirroring window; (c) Schematic of estimating time delay errors; (d) Results of 100 tests estimating time delay errors

We recruited a total of 52 MOBA players. For the first 29 players recruited, they were only required to play the game during the experiment without being asked any questions. Prior to gameplay initiation, the subsequent cohort of 23 participants received standardized instructions regarding real-time affective valence assessment. Specifically, players were required to numerically quantify their subjective experiences following each in-game kill/death event using an 11-point digital rating scale as illustrated in Fig. S5. The scale endpoints were operationally defined as: 0 = no feeling at all, 10 = extreme pleasure or displeasure. To minimize disruption to the gaming experience and the neural responses to key events, we asked these questions 3 s after a kill/death and instructed participants to respond only when asked, without the need to report unprompted. If intense gameplay might lead to another kill or death happening shortly thereafter, the questions were asked after the intense gameplay had subsided, thus researchers needed to have a certain ability to assess the situation in MOBA games. These questions were conducted by Hong-Zhi Li, a seasoned MOBA game player on our team.

Upon completing the experiment, we meticulously examined the recorded video frame-by-frame using Wondershare Filmora (Wondershare Technology Co., Ltd), with each frame corresponding to 16.67 ms. Utilizing the interface of Curry 8 software, we established the onset of the EEG recording as the time zero reference point. This setup allowed us to precisely determine the timings of key in-game events, including the start and end of the game, as well as kills and deaths. During the frame-by-frame video review, we identified kills and deaths based on the game's embedded real-time match statistics, such as “6 kills, 0 deaths, and 0 assists” (Fig. 1b). The term Assist in MOBA gameplay denotes cooperative combat participation wherein a player contributes to kill an opponent without the last hit. Any changes in these statistics marked the timing of the corresponding events. We recorded all instances of kills and deaths throughout the gameplay and saved this data in a text document for subsequent analysis. This method enables us to capture the neural responses associated with kills and deaths in real MOBA games, thus providing highly ecologically valid data.

EEG recording and preprocessing

EEG data were collected using the Neuroscan 64-channel recording system, ensuring that the impedance of each electrode remained below 10 kΩ. The online reference electrode was positioned between Cz and CPz. The system employed an online filtering range of 0.05–400 Hz and a sampling rate of 1,000 Hz. EEG offline processing was conducted using EEGLAB version 2023.0 in MATLAB R2019a (Delorme & Makeig, 2004). Initially, we imported the key event markers and their corresponding latencies from the previously mentioned txt document into the EEG data file. We then removed unused channels. The EEG signals were band-pass filtered between 0.5 and 80 Hz, with an additional notch filter applied between 48 and 52 Hz to eliminate power frequency interference. Subsequently, all EEG signals were re-referenced to an infinity reference using the Reference Electrode Standardization Technique (REST) (L. Dong et al., 2017). To accelerate data processing speed and reduce memory usage, we downsampled the EEG data to 256 Hz. We manually removed any bad epochs and interpolated any bad channels before running an independent component analysis (ICA). Each component was classified automatically using ICLabel (Pion-Tonachini, Kreutz-Delgado, & Makeig, 2019). Based on the key in-game events (kills and deaths), the EEG signals were segmented to include one second before and two seconds after each event. Epochs where kill or death events occurred within 2 s of each other were excluded. We manually removed artifact components, primarily those associated with muscle activity and eye movements, before subsequent analysis.

Data analysis

Time domain analysis

The epochs were re-segmented to include 200 ms before and 1,000 ms after each event. We removed the baseline from epochs ranging from −200 to 0 ms and applied a low-pass filter at 30 Hz. To obtain the average P300 amplitude and single-trial P300 amplitude, we utilized multiple linear regression with a dispersion term (MLRd) to enhance the signal-to-noise ratio and to automatically detect P300 amplitude within a time window of 100 ms around the peak latency of the grand average P300 (Hu et al., 2011; Hu, Mouraux, Hu, & Iannetti, 2010). We extracted P300 amplitudes both at the FCz (where the EEG topography showed the greatest amplitude) and as a global response averaged across all electrodes. Finally, for in-game kills and deaths separately, we conducted correlation analyses between these P300 amplitudes and addiction-related psychological traits. We also analyzed the correlations between single-trial P300 amplitudes and subjective pleasure experiences (following kills) and subjective unpleasant experiences (following deaths).

Time-frequency analysis

We conducted a Short-time Fourier Transformation (STFT) on the EEG epochs, using a window size of 400 ms and a frequency range from 1 to 80 Hz. The average power between −800 ms and −200 ms served as the baseline and was subtracted from the derived power values. Initially, we plotted the time-frequency distribution at the midline brain electrodes and then selected regions of interest, such as delta-band event-related synchronization (ERS) and theta-band event-related desynchronization (ERD). Next, we plotted topographic maps of these time-frequency responses and identified the electrodes where these responses were most prominent for subsequent correlation analyses. Similarly, we calculated the global average of time-frequency responses and extracted the respective regions of interest. For in-game kills and deaths separately, we conducted correlation analyses between these time-frequency responses of interest and addiction-related psychological traits. We also analyzed the correlation between these time-frequency responses and subjective pleasant experiences (following kills) and unpleasant experiences (following deaths) on a single-trial basis.

Statistical analysis

When calculating the correlation between single-trial neural responses and subjective experiences, we used both the original subjective ratings and normalized ratings, which were derived by subtracting the mean score of each participant for both kills and deaths, separately. All correlation analyses were performed using Spearman's correlation. Statistical significance was set at a two-tailed p < 0.05. Initially, we analyzed the data without distinguishing between genders; subsequently, we performed separate correlation analyses for male and female players. All data analyses were conducted in SPSS version 25.

Ethics

The study procedures were carried out in accordance with the Declaration of Helsinki. The Institutional Review Board of the University-Town Hospital of Chongqing Medical University approved the study. All subjects were informed about the study and all provided informed consent. Parental consent was sought for those younger than 18 years of age.

Results

General characteristics

Utilizing our proposed experimental protocol, the average timing error was estimated at 6.84 ms (Fig. 1c and d). Due to excessive EEG artifacts, one participant was excluded. Ultimately, 51 MOBA players (30 males and 21 females) were included for subsequent analysis. Among these participants, 23 were asked to rate their subjective pleasant or unpleasant experiences following each kill and each death event. Participants had an average age of 20.55 ± 3.22 years and an average IGD-20 score of 50.29 ± 12.17. Using the standardized IGD-20 cutoff score of 71 (Pontes et al., 2014), 4 participants (7.8%) met diagnostic criteria for IGD, while the majority (92.2%, n = 47) comprised a non-clinical sample exhibiting varying levels of addiction-related psychological traits. The number of valid epochs was 20.88 ± 8.21 for the death condition, and 28.18 ± 12.59 for the kill condition.

Time domain analyses

After in-game deaths, a P300 component with an amplitude of 11.62 ± 4.40 μV was observed at 372.47 ± 30.33 ms at FCz, which we tentatively term the “death-related potential” (DRP) (Fig. 2a). For the global response, the amplitude was 1.92 ± 0.85 μV at 383.88 ± 33.40 ms (Fig. 2b). The P300 amplitude at FCz showed a significant negative correlation with relapse, conflict, and IGD-20 total scores for males, but only with relapse for females (Figure S1, 2d, and 2e). In single-trial analysis, the unpleasant rating was significantly negatively associated with P300 amplitude at FCz in males, but positively in females (Fig. 2f and S2). The normalized unpleasant rating was significantly positively associated with P300 amplitude at FCz in females, but no significant correlation was found in males (Fig. 2g). For the global response after in-game deaths, the P300 amplitude showed a significant negative correlation with conflict and IGD-20 total scores in males, absent in females (Fig. 2h and i). However, a significant positive correlation between global P300 amplitude and withdrawal was observed only in females (Fig. S2). Single-trial analysis indicated a significant negative correlation between global P300 amplitude and unpleasant rating in males, and a positive correlation in females (Fig. 2j and k).

Fig. 2.

Fig. 2.

Open in a new tab

Event-related potential results and their correlation analysis following deaths in real games. (a) Event-related potentials at FCz electrode post-death; (b) Average whole-brain event-related potentials post-death; (c) Topographical maps in specific time windows post-death; (d–e) Correlations between P300 amplitude (FCz) and IGD-20, relapse scores post-death; (f–g) Correlations between raw and normalized (participant's raw score minus their average score) subjective unpleasant experiences and P300 amplitude post-death; (h–i) Correlations between average whole-brain P300 amplitude and IGD-20, conflict scores post-death; (j–k) Correlations between raw and normalized subjective unpleasant experiences and average whole-brain P300 amplitude post-death. Correlations indicated by Spearman's rho, *p < 0.05, **p < 0.01, ***p < 0.001

After in-game kills, a P300 component with an amplitude of 5.21 ± 2.46 μV was observed at 333.72 ± 38.40 ms at FCz, which we term the “kill-related potential” (KRP) (Fig. 3a). The global response amplitude was 1.02 ± 0.58 μV at 390.85 ± 59.56 ms (Fig. 3b). The P300 amplitude in KRP at FCz was negatively correlated with IGD-20 total score, conflict, and age in males, but not in females (Fig. 3d–f and S3). Single-trial analysis showed that pleasant rating was negatively associated with P300 amplitude at FCz in males (Fig. 3g and S4). For the global response, there was a significant negative correlation between amplitude and mood modification in males, and between amplitude and age in females (Fig. 3h–i). Single-trial analysis indicated a significant negative correlation between P300 amplitude and pleasant rating after in-game kills; no significant correlation was found for normalized rating (Fig. 3j and S4).

Fig. 3.

Fig. 3.

Open in a new tab

Event-related potential results and their correlation analysis after kills in real games. (a) Event-related potentials at FCz electrode post-kill; (b) Average whole-brain event-related potentials post-kill; (c) Topographical maps in specific time windows post-kill; (d–f) Correlations between P300 amplitude (FCz) and IGD-20, conflict, age post-kill; (g) Correlation between raw subjective reward experience and P300 amplitude post-kill; (h–i) Correlations between average whole-brain P300 amplitude and emotion, age post-kill; (j) Correlation between raw subjective reward experience and average whole-brain P300 amplitude post-kill. Correlations indicated by Spearman's rho, *p < 0.05, **p < 0.01, ***p < 0.001

Time-frequency analysis

Following in-game deaths, from 400 to 1,800 ms, there was evident beta-ERS in parietal regions (Fig. 4a and g). Correlation analysis showed beta-ERS was positively correlated with IGD-20 total score, withdrawal, conflict, unpleasant rating, and normalized unpleasant rating in males, which was absent in females (Fig. 4b–f and S1). Similar results were observed for global beta-ERS (Fig. 4g–k). We also observed delta-ERS and theta-band ERD mainly in the centro-frontal region (Fig. 5a). The theta-ERD was positively correlated with IGD-20 total scores, salience, relapse, and conflict in males, but negatively with conflict in females (Fig. 5b–e). The delta-ERS was negatively correlated with salience and conflict in both genders, but only significant in males (Fig. 5f–g). For global neural responses, no significant correlation was observed for theta-ERD and delta-ERS. Single-trial analysis showed that alpha-ERD was significantly negatively correlated with normalized unpleasant rating in males (Figs S2 and 5h–i).

Fig. 4.

Fig. 4.

Open in a new tab

Beta-band neural oscillations and their correlation analysis following deaths in real games. (a) Beta-ERS post-death; (b–f) Correlations between beta-ERS and psychological traits related to game addiction and subjective unpleasant experience post-death; (g) Average whole-brain response of beta-ERS post-death; (h–k) Correlations between average whole-brain beta-ERS and psychological traits related to game addiction and subjective unpleasant experience post-death. Correlations indicated by Spearman's rho, *p < 0.05, **p < 0.01, ***p < 0.001

Fig. 5.

Fig. 5.

Open in a new tab

Delta, theta, and alpha-band neural oscillations and their correlation analysis following deaths in real games. (a) Delta-ERS and theta-ERD post-death; (b–g) Correlations between delta-ERS and theta-ERD with psychological traits related to game addiction post-death; (h) Alpha-ERD post-death; (i) Correlation between alpha-ERD and subjective unpleasant experience post-death. Correlations represented by Spearman's rho, *p < 0.05, **p < 0.01, ***p < 0.001

After in-game kills, from 300 to 1,800 ms, a noticeable ERS in the beta band was observed in the centro-parietal region (Fig. S6a). Although no significant correlation was observed between beta-ERS and addiction-related psychological traits, there was a negative correlation between beta-ERS and pleasant rating after in-game kills in both males and females. For normalized ratings, the correlation coefficients became larger and significant in females, while smaller and non-significant in males (Fig. S6b–c). For the global response, similar correlations were observed in both genders (Fig. S6d–f). For gamma-ERS after in-game kills, negative correlations were observed for IGD-20 total score, mood modification, relapse, and tolerance in males, however, an inverse relationship was observed in females (Fig. S6g–m). Single-trial analysis showed a significantly negative correlation between normalized rating and gamma-ERS in females, but not in males (Fig. S6n).

Discussion and conclusions

In this study, we developed an experimental protocol that enables precise marking of in-game events, including kills and deaths during real MOBA gameplay, along with synchronous EEG recording. With this setup, we successfully captured the neural responses associated with kills and deaths in MOBA games, providing highly ecologically valid data. To our knowledge, this is the first study to reveal the characteristics of neural responses such as event-related potentials (ERPs), neural oscillations (ERS and ERD), following in-game kills and deaths during real MOBA gameplay. Through time-domain analysis, time-frequency analysis, and single-trial analysis, we identified several previously unreported correlations between these neural responses and addiction-related psychological traits and subjective pleasant or unpleasant experiences. This fills a current gap in the relevant field, enhancing our understanding of the underlying neural mechanisms that keep MOBA gamers so deeply engaged. Our results highlight the importance of neural responses to rewarding and punishing events such as in-game kills and deaths in the study of game addiction.

The P300 component, a late positive waveform in event-related potentials (ERPs), typically emerges 300–600 ms post-stimulus. As a robust neurophysiological index, it is strongly associated with higher-order cognitive functions, including attentional resource allocation, response inhibition, and reward/aversion processing (Dores et al., 2025; Polich, 2007; Y. Zhang, Ou, Yuan, & Sun, 2021). Furthermore, the amplitude and latency characteristics of the P300 have established clinical utility as biomarkers for a range of neurological and psychiatric disorders (Duncan et al., 2009). As expected, both DRP and KRP predominantly exhibited late components (P300 in our study), with scarcely any discernible early components. Considering there were no abrupt visual or auditory stimuli immediately after kills or deaths, the game-related visual and auditory stimuli persisted throughout the epoch. This resulted in a lack of apparent bottom-up early components, with a predominance of top-down, late components. We found that in male players, the P300 amplitude following kills was significantly negatively correlated with the IGD-20 total score, implying that a lower P300 amplitude is associated with a higher addiction tendency. Simultaneously, we observed that the P300 amplitude following kills was also significantly negatively correlated with the subjective pleasure experienced post-kill, indicating that lower P300 amplitudes are linked to stronger subjective pleasure after kills. Combining these findings, we speculate that male players with higher addiction tendencies experience stronger subjective pleasure after kills, suggesting that male players who derive stronger pleasure from kills in MOBA games are more prone to addiction. This perspective aligns with the game addiction model based on reinforcement learning theory, where real-time pleasurable experiences in games can positively reinforce gaming behavior (Xi & Hu, 2022). However, this hypothesis still needs further verification by future research. Similarly, previous studies have reported that individuals with IGD showed smaller P300 amplitudes (Q. Li, Wang, et al., 2020). This study used the Go/NoGo paradigm, thus the P300 amplitude was considered related to inhibitory control. However, the P300 that arises after kill events in a natural setting may be associated more with top-down reward processing or emotional regulation, rather than inhibitory control. Therefore, directly comparing P300 from Go/NoGo paradigms with P300 from this study may not be appropriate.

Compared with previous studies, the rewarding and punishing stimuli in this research were not associated with money but were instead familiar in-game events to players, such as kills and deaths in real MOBA games. This could potentially limit the comparability of our results with those of previous studies. One study employing a semi-natural gaming design reported that individuals with IGD exhibited reduced P300 amplitude following reward stimuli compared to casual gamers (Duven et al., 2015). This finding is somewhat comparable to our results, and the scalp topography distribution of the P300 amplitude, primarily centered around the frontal-central region, as well as the latency and temporal evolution of the scalp distribution, are similar to our findings. The authors offer two explanations: first, addicted players might have reduced sensitivity to natural rewards, similar to findings in substance dependence research (Parvaz et al., 2012). However, other studies suggest that individuals with internet addiction have increased sensitivity to rewards and decreased sensitivity to punishments (G. Dong, Hu, & Lin, 2013; G. Dong, Huang, & Du, 2011). If a reduced P300 amplitude is considered indicative of decreased reward sensitivity (Wu & Zhou, 2009), our results might also support the notion that players with higher addiction tendencies have reduced sensitivity to rewards. Another explanation given is the tolerance effect (Duven et al., 2015), which should not be confused with the tolerance described in the IGD-20. This tolerance effect refers to the decreased sensitivity to rewarding stimuli due to frequent exposure, thus leading to a tolerance towards these stimuli. However, our single-trial analysis, showing that the P300 amplitude post-kill is inversely related to the subjective pleasant experience, appears to contradict this view. Since the ERP is a composite waveform of many components (Dien, 2010), the P300 we observe likely includes both positive and negative components. It's possible that the reactivity to rewards is represented by the negative components, suggesting that the observed decrease in P300 amplitude among players with higher addiction tendencies could be due to an increase in the negative components that represent reward reactivity.

This study offers several insights for future research. As Duven et al. (2015) suggested, studying the neural mechanisms of reward processing in ecologically valid contexts is crucial for a deeper understanding of the complex symptomatology of IGD (Duven et al., 2015), and our study was conducted based on this principle. Although our experimental protocol was developed for research on MOBA games, it can be seamlessly adapted to other types of games to precisely mark the timing of key events in real gameplay, thereby facilitating the study of neural responses to key events in other game types and advancing research on the neural mechanisms of IGD in natural settings. Moreover, the neural responses to kills and deaths that are closely linked to addiction characteristics may serve as targets for neuromodulation, potentially improving the current challenges in treating IGD (Gorowska, Tokarska, Zhou, Gola, & Li, 2022; Zajac, Ginley, & Chang, 2020). Additionally, in the field of artificial intelligence (AI), while AI models can now rival top human MOBA players through supervised learning (Ye et al., 2022), they fall short in replicating human emotional responses, reward experiences, and motivational cravings. The findings on human neural reactions to in-game events may offer valuable insights for developing AI with more human-like attributes.

There are several limitations. First, the method we proposed cannot pinpoint the latency of key in-game events to the accuracy of 1 ms. This limitation primarily arises from the frame rate of the projection. In this study, the frame rate of projection was 60, indicating that the finest time resolution for the game display is 16.67 ms. Given our primary interest in the late components associated with higher-level cognitive functions, a time resolution of 16.67 ms may not be a significant issue. However, accurately capturing the latency of key in-game events remains a technical challenge worth exploring. Second, we only identified in-game kills and deaths as key events. Ideally, this should also include events like the end of a match due to a loss or victory. Since each match lasts about 20 min and game outcomes are unpredictable, obtaining a sufficient number of epochs for events like victories or losses becomes challenging. This often requires participants to play a substantial number of matches and spend extended periods on data collection. Third, during EEG signal preprocessing, we excluded epochs where the time interval between consecutive kills or between a kill and a death was less than 2 s, to prevent interference from overlapping neural responses following key events. However, this approach meant we couldn't capture the neural responses induced by double kills, triple kills, or even pentakills, which often provide players with a more intense reward experience. Finally, while our ecologically-embedded paradigm successfully demonstrates the feasibility of naturalistic research approaches, the relatively low proportion of IGD cases (7.8%) in our sample limits our ability to draw definitive conclusions regarding disorder-specific neural correlates. Nevertheless, the current findings establish a robust methodological foundation for future case-control studies, marking a crucial advancement toward elucidating the neurobiological mechanisms underlying IGD in naturalistic contexts.

The innovative method introduced in this study enables the acquisition of neural responses with remarkable ecological validity following kills and deaths in real MOBA games. We have delineated the characteristics of neural responses to kills and deaths in naturalistic settings of real MOBA games and provided preliminary evidence of a close relationship between these neural responses and addiction-related psychological traits. The observed neural reactivity patterns to in-game events demonstrate significant associations with addiction-relevant psychological dimensions, providing novel insights into neurocognitive mechanisms underlying MOBA gaming engagement. This study offers a new methodology and perspective for investigating IGD and will inspire further research into the neural mechanisms of IGD in naturalistic contexts.

Supplementary material

jba-14-1304-s001.pdf (3.3MB, pdf)
Open in a new tab

Funding Statement

Funding sources: This study was supported by the Chongqing Municipal Health Commission (Chongqing Medical Youth Top Talent Project) and the Program for Youth Innovation in Future Medicine at Chongqing Medical University (W0180).

Footnotes

Authors' contribution: Study concept and design: DZ, HL, LK. Analysis and interpretation of data: DZ, HL, WW, JY. Obtained funding: DZ, WW. Statistical analysis: JY, ZL, LW. Study supervision: DZ, WW, LK. Writing – original draft: DZ, HL, JY. Writing – review & editing: DZ, HL, JY, ZL, LW, WW, LK. All authors had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest: The authors declare no conflict of interest.

Contributor Information

Dong-Dong Zhou, Email: zhoudongdong@cqmu.edu.cn.

Hong-Zhi Li, Email: lihongzhi97@163.com.

Jia-Jia Yang, Email: yangjiajia9980@163.com.

Zhen Lv, Email: 673823208@qq.com.

Li-Yang Wan, Email: 422954131@qq.com.

Wo Wang, Email: wangwo@cqmu.edu.cn.

Li Kuang, Email: kuangli0308@163.com.

Data availability

All derived data and code are publicly available on the Dryad (https://doi.org/10.5061/dryad.zpc866th0).

References

  1. American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders: Diagnostic and statistical manual of mental disorders (5th ed.). [Google Scholar]
  2. Balodis, I. M., & Potenza, M. N. (2015). Anticipatory reward processing in addicted populations: A focus on the monetary incentive delay task. Biological Psychiatry, 77(5), 434–444. 10.1016/j.biopsych.2014.08.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Delorme, A., & Makeig, S. (2004). EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1), 9–21. 10.1016/j.jneumeth.2003.10.009 [DOI] [PubMed] [Google Scholar]
  4. Dien, J. (2010). The ERP PCA Toolkit: An open source program for advanced statistical analysis of event-related potential data. Journal of Neuroscience Methods, 187(1), 138–145. 10.1016/j.jneumeth.2009.12.009 [DOI] [PubMed] [Google Scholar]
  5. Dong, G., Hu, Y., & Lin, X. (2013). Reward/punishment sensitivities among internet addicts: Implications for their addictive behaviors. Progress in Neuro-psychopharmacology & Biological Psychiatry, 46, 139–145. 10.1016/j.pnpbp.2013.07.007 [DOI] [PubMed] [Google Scholar]
  6. Dong, G., Huang, J., & Du, X. (2011). Enhanced reward sensitivity and decreased loss sensitivity in internet addicts: An fMRI study during a guessing task. Journal of Psychiatric Research, 45(11), 1525–1529. 10.1016/j.jpsychires.2011.06.017 [DOI] [PubMed] [Google Scholar]
  7. Dong, L., Li, F., Liu, Q., Wen, X., Lai, Y., Xu, P., & Yao, D. (2017). MATLAB toolboxes for reference electrode standardization technique (REST) of scalp EEG. Frontiers in Neuroscience, 11, 601. 10.3389/fnins.2017.00601 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dores, A. R., Peixoto, M., Fernandes, C., Geraldo, A., Griffiths, M. D., & Barbosa, F. (2025). Neurophysiological correlates of near-wins in gambling: A systematic literature review. Journal of Gambling Studies, 41(1), 5–35. 10.1007/s10899-024-10327-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Duncan, C. C., Barry, R. J., Connolly, J. F., Fischer, C., Michie, P. T., Näätänen, R., … Van Petten, C. (2009). Event-related potentials in clinical research: Guidelines for eliciting, recording, and quantifying mismatch negativity, P300, and N400. Clinical Neurophysiology, 120(11), 1883–1908. 10.1016/j.clinph.2009.07.045 [DOI] [PubMed] [Google Scholar]
  10. Duven, E. C., Müller, K. W., Beutel, M. E., & Wölfling, K. (2015). Altered reward processing in pathological computer gamers--ERP-results from a semi-natural gaming-design. Brain and Behavior, 5(1), 13–23. 10.1002/brb3.293 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Fam, J. Y. (2018). Prevalence of internet gaming disorder in adolescents: A meta-analysis across three decades. Scandinavian Journal of Psychology, 59(5), 524–531. 10.1111/sjop.12459 [DOI] [PubMed] [Google Scholar]
  12. Gao, Y. X., Wang, J. Y., & Dong, G. H. (2022). The prevalence and possible risk factors of internet gaming disorder among adolescents and young adults: Systematic reviews and meta-analyses. Journal of Psychiatric Research, 154, 35–43. 10.1016/j.jpsychires.2022.06.049 [DOI] [PubMed] [Google Scholar]
  13. Gorowska, M., Tokarska, K., Zhou, X., Gola, M. K., & Li, Y. (2022). Novel approaches for treating internet gaming disorder: A review of technology-based interventions. Comprehensive Psychiatry, 115, 152312. 10.1016/j.comppsych.2022.152312 [DOI] [PubMed] [Google Scholar]
  14. Gou, S., Zhang, W., Tang, Y., Zhang, J., & He, Q. (2024). Prevalence of internet gaming disorder among Chinese adolescents: A systematic review and meta-analysis. Asian Journal of Psychiatry, 102, 104257. 10.1016/j.ajp.2024.104257 [DOI] [PubMed] [Google Scholar]
  15. Han, H., Jeong, H., Jo, S. J., Son, H. J., & Yim, H. W. (2020). Relationship between the experience of online game genre and high risk of internet gaming disorder in Korean adolescents. Epidemiology and Health, 42, e2020016. 10.4178/epih.e2020016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hu, L., Liang, M., Mouraux, A., Wise, R. G., Hu, Y., & Iannetti, G. D. (2011). Taking into account latency, amplitude, and morphology: Improved estimation of single-trial ERPs by wavelet filtering and multiple linear regression. Journal of Neurophysiology, 106(6), 3216–3229. 10.1152/jn.00220.2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hu, L., Mouraux, A., Hu, Y., & Iannetti, G. D. (2010). A novel approach for enhancing the signal-to-noise ratio and detecting automatically event-related potentials (ERPs) in single trials. Neuroimage, 50(1), 99–111. 10.1016/j.neuroimage.2009.12.010 [DOI] [PubMed] [Google Scholar]
  18. Kätsyri, J., Hari, R., Ravaja, N., & Nummenmaa, L. (2013). Just watching the game ain't enough: Striatal fMRI reward responses to successes and failures in a video game during active and vicarious playing. Frontiers in Human Neuroscience, 7, 278. 10.3389/fnhum.2013.00278 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Li, X., Huang, L., Li, B., Wang, H., & Han, C. (2020). Time for a true display of skill: Top players in league of legends have better executive control. Acta Psychologica (Amst), 204, 103007. 10.1016/j.actpsy.2020.103007 [DOI] [PubMed] [Google Scholar]
  20. Li, Q., Wang, Y., Yang, Z., Dai, W., Zheng, Y., Sun, Y., & Liu, X. (2020). Dysfunctional cognitive control and reward processing in adolescents with internet gaming disorder. Psychophysiology, 57(2), e13469. 10.1111/psyp.13469 [DOI] [PubMed] [Google Scholar]
  21. Männikkö, N., Ruotsalainen, H., Tolvanen, A., & Kääriäinen, M. (2019). Psychometric properties of the internet gaming disorder test (IGDT-10) and problematic gaming behavior among Finnish vocational school students. Scandinavian Journal of Psychology, 60(3), 252–260. 10.1111/sjop.12533 [DOI] [PubMed] [Google Scholar]
  22. Nuyens, F., Deleuze, J., Maurage, P., Griffiths, M. D., Kuss, D. J., & Billieux, J. (2016). Impulsivity in multiplayer online battle arena gamers: Preliminary results on experimental and self-report measures. Journal of Behavioral Addictions, 5(2), 351–356. 10.1556/2006.5.2016.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Palaus, M., Marron, E. M., Viejo-Sobera, R., & Redolar-Ripoll, D. (2017). Neural basis of video gaming: A systematic review. Frontiers in Human Neuroscience, 11, 248. 10.3389/fnhum.2017.00248 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Parvaz, M. A., Maloney, T., Moeller, S. J., Woicik, P. A., Alia-Klein, N., Telang, F., … Goldstein, R. Z. (2012). Sensitivity to monetary reward is most severely compromised in recently abstaining cocaine addicted individuals: A cross-sectional ERP study. Psychiatry Research, 203(1), 75–82. 10.1016/j.pscychresns.2012.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pion-Tonachini, L., Kreutz-Delgado, K., & Makeig, S. (2019). ICLabel: An automated electroencephalographic independent component classifier, dataset, and website. Neuroimage, 198, 181–197. 10.1016/j.neuroimage.2019.05.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Polich, J. (2007). Updating P300: An integrative theory of P3a and P3b. Clinical Neurophysiology, 118(10), 2128–2148. 10.1016/j.clinph.2007.04.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Pontes, H. M., Király, O., Demetrovics, Z., & Griffiths, M. D. (2014). The conceptualisation and measurement of DSM-5 internet gaming disorder: The development of the IGD-20 test. Plos One, 9(10), e110137. 10.1371/journal.pone.0110137 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Raiha, S., Yang, G., Wang, L., Dai, W., Wu, H., Meng, G., … Liu, X. (2020). Altered reward processing system in internet gaming disorder. Frontiers in Psychiatry, 11, 599141. 10.3389/fpsyt.2020.599141 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Weinstein, A., Livny, A., & Weizman, A. (2017). New developments in brain research of internet and gaming disorder. Neuroscience and Biobehavioral Reviews, 75, 314–330. 10.1016/j.neubiorev.2017.01.040 [DOI] [PubMed] [Google Scholar]
  30. World Health Organization (2019). ICD-11 for mortality and morbidity statistics (Version: 01/2023). Retrieved from https://icd.who.int/browse11/l-m/en. [Google Scholar]
  31. Wu, Y., & Zhou, X. (2009). The P300 and reward valence, magnitude, and expectancy in outcome evaluation. Brain Research, 1286, 114–122. 10.1016/j.brainres.2009.06.032 [DOI] [PubMed] [Google Scholar]
  32. Xi, W., & Hu, Y. Z. (2022). Internet gaming disorder in adolescents: Review and prospect. Chinese Journal of Applied Psychology, 28(1), 3–19. 10.3969/j.issn.1006-6020.2022.01.001 [DOI] [Google Scholar]
  33. Yao, Y. W., Liu, L., Ma, S. S., Shi, X. H., Zhou, N., Zhang, J. T., & Potenza, M. N. (2017). Functional and structural neural alterations in internet gaming disorder: A systematic review and meta-analysis. Neuroscience and Biobehavioral Reviews, 83, 313–324. 10.1016/j.neubiorev.2017.10.029 [DOI] [PubMed] [Google Scholar]
  34. Ye, D., Chen, G., Zhao, P., Qiu, F., Yuan, B., Zhang, W., … Huang, L. (2022). Supervised learning achieves human-level performance in MOBA games: A case study of honor of kings. IEEE Transactions on Neural Networks and Learning Systems, 33(3), 908–918. 10.1109/tnnls.2020.3029475 [DOI] [PubMed] [Google Scholar]
  35. Zajac, K., Ginley, M. K., & Chang, R. (2020). Treatments of internet gaming disorder: A systematic review of the evidence. Expert Review of Neurotherapeutics, 20(1), 85–93. 10.1080/14737175.2020.1671824 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Zhang, J., Hu, Y., Wang, Z., Wang, M., & Dong, G. H. (2020). Males are more sensitive to reward and less sensitive to loss than females among people with internet gaming disorder: fMRI evidence from a card-guessing task. BMC Psychiatry, 20(1), 357. 10.1186/s12888-020-02771-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Zhang, Y., Ou, H., Yuan, T. F., & Sun, J. (2021). Electrophysiological indexes for impaired response inhibition and salience attribution in substance (stimulants and depressants) use disorders: A meta-analysis. International Journal of Psychophysiology, 170, 133–155. 10.1016/j.ijpsycho.2021.10.008 [DOI] [PubMed] [Google Scholar]
  38. Zhou, R., Morita, N., Ogai, Y., Saito, T., Zhang, X., Yang, W., & Yang, F. (2024). Meta-analysis of internet gaming disorder prevalence: Assessing the impacts of DSM-5 and ICD-11 diagnostic criteria. International Journal of Environmental Research and Public Health, 21(6). 10.3390/ijerph21060700 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All derived data and code are publicly available on the Dryad (https://doi.org/10.5061/dryad.zpc866th0).

Articles from Journal of Behavioral Addictions are provided here courtesy of Akadémiai Kiadó

RESOURCES