(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Investigating the mediating role of emotional intelligence in the relationship between internet addiction and mental health among university students [1]

['Girum Tareke Zewude', 'Department Of Psychology', 'Wollo University', 'Dessie', 'Derib Gosim', 'Seid Dawed', 'Tilaye Nega', 'Getachew Wassie Tessema', 'Department Of Information System', 'Amogne Asfaw Eshetu']

Date: 2024-12

Abstract Introduction The widespread use of the internet has brought numerous benefits, but it has also raised concerns about its potential negative impact on mental health, particularly among university students. This study aims to investigate the relationship between internet addiction and mental health in university students, as well as explore the mediating effects of emotional intelligence in this relationship. Objective The main objective of this study was to examine whether internet addiction (dimensions and total) negatively predicts the mental health of university students, with emotional intelligence acting as a mediator. Methods To address this objective, a cross-sectional design with an inferential approach was employed. Data were collected using the Wong Law Emotional Intelligence Scale (WLEIS-S), Internet Addiction Scale (IAS), and Keyes’ Mental Health Continuum-Short Form (MHC-SF). The total sample consisted of 850 students from two large public higher education institutions in Ethiopia, of which 334 (39.3%) were females and 516 (60.7%) were males, with a mean age of 22.32 (SD = 4.04). For the purpose of the study, the data were split into two randomly selected groups: sample 1 with 300 participants for psychometric testing purposes, and sample 2 with 550 participants for complex mediation purposes. Various analyses were conducted to achieve the stated objectives, including Cronbach’s alpha and composite reliabilities, bivariate correlation, discriminant validity, common method biases, measurement invariance, and structural equation modeling (confirmatory factor analysis, path analysis, and mediation analysis). Confirmatory factor analysis was performed to assess the construct validity of the WLEIS-S, IAS, and MHC-SF. Additionally, a mediating model was examined using structural equation modeling with the corrected biased bootstrap method. Results The results revealed that internet addiction had a negative and direct effect on emotional intelligence (β = –0.180, 95%CI [–0.257, –0.103], p = 0.001) and mental health (β = –0.204, 95%CI [–0.273, –0.134], p = 0.001). Also, Internet Craving and Internet obsession negatively predicted EI (β = –0.324, 95%CI [–0.423, –0.224], p = 0.002) and MH (β = –0.167, 95%CI [–0.260, –0.069], p = 0.009), respectively. However, EI had a significant and positive direct effect on mental health (β = 0.494, 95%CI [0.390, 0.589], p = 0.001). Finally, EI fully mediated the relationship between internet addiction and mental health (β = –0.089, 95%CI [–0.136, –0.049], p = 0.001). Besides The study also confirmed that all the scales had strong internal consistency and good psychometric properties. Conclusion This study contributes to a better understanding of the complex interplay between internet addiction, emotional intelligence, and mental health among university students. The findings highlight the detrimental effects of internet addiction on mental health, and the crucial mediating role of emotional intelligence. Recommendations The findings discussed in relation to recent literature have practical implications for practitioners and researchers aiming to enhance mental health and reduce internet addiction among university students. Emotional intelligence can be utilized as a positive resource in interventions and programs targeting these issues.

Author summary This study examined the psychometric properties and relationships between internet addiction, emotional intelligence, and mental health among Ethiopian university students. The measures used ‐ the Wong Law Emotional Intelligence Scale (WLEIS-S), Internet Addiction Scale (IAS), and Keyes’ Mental Health Continuum-Short Form (MHC-SF) ‐ were found to be reliable and valid in the Ethiopian context. The findings indicate that higher levels of internet addiction are associated with lower emotional intelligence and poorer mental health. Conversely, emotional intelligence was positively correlated with mental health. The study determined that internet addiction negatively predicts both emotional intelligence and mental health, while internet craving directly predicts emotional intelligence and indirectly and negatively predicts mental health through emotional intelligence. These results highlight the detrimental effects of excessive internet use on the mental well-being of university students. The authors recommend interventions to raise awareness, promote healthy digital habits, and enhance emotional intelligence as a means of supporting students’ mental health and preventing internet addiction.

Citation: Zewude GT, Gosim D, Dawed S, Nega T, Tessema GW, Eshetu AA (2024) Investigating the mediating role of emotional intelligence in the relationship between internet addiction and mental health among university students. PLOS Digit Health 3(11): e0000639. https://doi.org/10.1371/journal.pdig.0000639 Editor: Calvin Or, The University of Hong Kong, HONG KONG Received: February 29, 2024; Accepted: September 11, 2024; Published: November 20, 2024 Copyright: © 2024 Zewude et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: All data are in the manuscript and/or supporting information files. Funding: The author(s) received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist.

2. Research method 2.1. Research design The current study employed a quantitative research design with an associational and cross-sectional approach, deemed well-suited to achieve the stated objectives. 2.2. Study setting The study’s target population consists of undergraduate students enrolled in public universities in the Amhara Regional State of Ethiopia. 2.3. Sample and sampling The study was conducted in the Amhara Regional State of Ethiopia, focusing on two randomly selected public universities: Wollo University and Woldia University. The choice of this area was based on the researchers’ extensive 16-year experience working in the region and the convenience of accessing accurate and efficient data due to the proximity of the study area to their workplace. The sampling process involved dividing the sampling frame into subsections that represented characteristic groups. From each stratum, a random sample was selected. Initially, 889 university students were randomly chosen and invited to participate in the surveys. However, due to missing information, mistakes, or carelessness in data entry, 39 participants were excluded. This resulted in an effective response rate of 95.6%. The final data consisted of 850 students, randomly split into two separate studies. The first study focused on validating the adapted instruments, while the second explored the mediation model through a complex structural equation modeling (SEM) analysis. In the first sample, there were 199 male and 101 female students. The second sample comprised 317 male and 233 female students. In total, the participants included 516 male students (60.7%) and 334 female students (39.3%). The mean age of the participants was 22.32 years, with a standard deviation of 4.04. To ensure comprehensive and reliable data, the researchers grouped the students’ year/batch into three levels. Initially, the students were stratified by year and gender. From the total sample of 850 students, the following numbers were selected using a simple random sampling technique: (a) Freshman (1st year): 233 students (124 male and 109 female), (b) Sophomore (2nd year): 260 students (160 male and 100 female), and (c) Senior (>2nd year): 357 students (232 male and 125 female). The sample size was determined based on established guidelines. A small sample size is typically around 100, a medium size is approximately 150, and a large size is considered to be greater than 200. To ensure statistically stable estimates and minimize sampling errors, it is recommended to have a sample size of 200 or more [44]. Therefore, in line with these recommendations and considering the power of the test, the sample size in this study was determined. 2.4. Instruments 2.4.1. Socio-demographic information. Undergraduate students self-reported their gender, batch, age, and university. 2.4.2. Internet Addiction Scale (IAS). The IAS developed by [23] and later adapted in Ethiopian context by [6] used to assess an individual’s excessive and compulsive internet use that interferes with daily functioning using a 17-item scale. The IAS was a seven-item response option ranging from Very Strongly disagree (1) to Very Strongly agree (7). IAS has four main dimensions with very good psychometric properties: Internet craving (IC; Five Items; α = 0.771, CR = 0.836), Internet compulsive disorder (ICD; Four items; α = 0.776, CR = 0.857), Addictive behaviour (AB; Four Items; α = 0.741, CR = .802), Internet obsession (IO; Four Items; α = 0.663, CR = 0.862) and the total scale (α = 0.878). In this study, Cronbach’s alpha coefficient, composite reliability (CR), construct validity, convergent and discriminant validity was acceptable based on global cut-off points/rule of thumb. 2.4.3. Emotional Intelligence Scale (EIS-16). EI was assessed using the 16-item EIS-16 [45] based on the Salovey Mayer EI framework [46]. Respondents rated each item on a 7-point Likert scale, ranging from 1 (very strongly disagree) to 7 (very strongly agree). EIS includes four main dimensions: Self-Emotion Appraisal (SEA); Others’ Emotion Appraisal (OEA); Use of Emotions (UOE, and Regulation of Emotions (ROE), each of which are measured on four items [47]. The reliability and construct validity of the Amharic version tested in this study was acceptable (see details under the measurement model). In this study, the Cronbach’s alpha, CR, and construct validity were found to be acceptable for the Ethiopian cultural context. 2.4.4. Mental Health Continuum-Short Form [MHC-SF]. MHC-SF is the most widely used instrument designed to measure the status of mental health of an adolescent. The Keyes’ Mental Health Continuum-Short Form [MHC-SF] was used to measure mental health of the participants [25]. The MHC-SF instrument was used to assess ‘frequency of happiness, social belongingness to a community and managing the psychological functioning of the daily life [6,25]. Three major clusters: social, emotional, and psychological well-being were used to assess the healthy functioning of adolescents’ mental health. Respondents The scale covered three dimensions: social well-being (5 items; α = 0.74), emotional well-being (3 items; α = 0.85), and psychological well-being (6 items; α = 0.84) and comprised 14 items [48]. The overall scale of MHC-SF reliability was α = 0.91. Respondents rate each item on a 7-point Likert scale, ranging from 1 (Very strongly disagree) to 7 (Very strongly agree). This scale possessed excellent construct validity and reliability [25]. Besides, the reliability and validity of the Amharic version have been proven in the Ethiopian context [6]. 2.5. Statistical data analysis IBM SPSS 29.0 and Smart-PLS 4.1.0.3 were used to perform the analyses. The psychometric properties and the complex mediation analyses were two essential aspects of this study. To test an instrument for psychometrically suitability, it is recommended to apply several methods and follow a scientific procedure in their assessment in study one followed by the complex mediation analysis using bootstrapping corrected bias. The main rationale of testing psychometric properties was because of cross-cultural validation is threatened by methodological difficulties, including those stemming from the translation of the questionnaire and the measurements of other instruments [33]. Therefore, in this study, validation was done following the guidelines proposed by [49]: (a) initial translation/forward translation, (b) translation synthesis, (c) back translation, (d) expert/translator review, and (e) administration and validation. In addition, the instruments were validated based on the recommendation of [30, 32]. Overall, the validation and the mediation findings were obtained through four processes. Multi-collinearity. VIF and tolerance were used to identify multi-collinearity in statistical data, following the recommendations of [32, 50]. In addition, the Harman single-factor test was used to examine common method variance bias. Evidence of reliability. CR and the Cronbach’s alpha coefficient were used to test the internal consistency of the subscales. Excellent internal consistency is shown by values over 0.90, good internal consistency is indicated by values between 0.80 and 0.90, and acceptable internal consistency is demonstrated by values between 0.70 and 0.80 [47,51–53]. Confirmation of construct validity through convergent, divergent, and discriminant validity. Average variance extracted and maximum shared variance were used to assess convergent and discriminant validity. AVE values greater than 0.5 are indicative of good convergent validity in a factor. Additionally, variables with a sufficient level of discriminant validity have an MSV value is lower than their AVE value [32]. Measurement invariance. We used CFA to examine the psychometric equivalence of the variables across distinct groups for measurement invariance (MI) [29]. In this work, a single-group CFA and multi-group CFA with four MI phases were used in accordance with accepted scientific practices [29]. Stage 1 involved conducting a configural invariance to create a baseline model that could be used for all groups without restriction, with the tested construct being the same in each group. Stage 2 of the analysis looked at the metric measurement invariance (MMI), which observed how indicators were reacted to by various groups using the same constrained factorial loadings. Stage 3 involved scalar MI, often known as strong invariance (SMI). In this test, factor loadings and indicator intercepts were limited uniformly across groups. In the fourth stage, strict invariance (RMI), or residual measurement invariance (RMI), was assessed. RMI represents the similarity of metric and scalar invariant items’ residuals [29]. Following [29,32]. Using multi-group CFA, the MI four sequential-staged analysis in the current study produced the following recommendation criteria. For metric, scalar, and residual invariance, the CFI and TLI ranged from 0 (perfect) to 0.01 (acceptable), and 0.015 (RMSEA) [29,31–32]. Structural Equation Modeling (SEM). This study investigated how well construct validity and causal relationship could explain the study variables [32]. Hence, SEM the best statistical Tanique to fit with the research purpose used to analyze the relationships between observed and latent (unobserved) variables [32]. We used three types of SEM in this study for testing causal relationship between the exogenous and endogenous variables of which there are three types: measurement model, structural model and path analysis [32]. Measurement Model is used to specifies how the observed variables are related to the underlying constructs they are intended to measure done through Confirmatory Factor Analysis (CFA). CFA tests a measurement theory by providing evidence of the validity of individual measures using the model’s overall fitness and other evidence of construct validity [31–32,47]. B) Structural model is specifying the hypothesized causal relationships among the latent variables and how they influence each other; (C) path analysis used to test and estimate complex relationships among variables currently used in our study [32]. We evaluated the goodness-of-fit using normed chi-square (χ2/df), the Tucker Lewis Index (TLI), Comparative Fit Index (CFI), and Root Mean Squared Error of Approximation (RMSEA). Measurement and structural models are generally deemed to exhibit excellent and sufficient fit if χ2/df is below 3 or 5, RMSEA and SRMR are below 0.08 and 0.01, and TLI and CFI are more than above 0.95 and 0.90, respectively [54]. The hypothesized model described in Figs 1 and 2 was examined using the 95% bias-corrected confidence intervals to examine indirect effects using the bootstrap method and 5000 resamples. 2.6. Procedures and informed consent statement The questionnaire applied incorporated 49 questions, measuring the EI (16 questions), the Internet addiction Scale (17 questions), Mental health Continuum Short Form (14 questions) and three socio-demographic factors. Paper and pencil were used by every participant to complete the surveys. The American Psychological Association’s ethical guidelines and standards, the Wollo University, the Internal Review Board, and standard data collection process were all followed. Participation was voluntary, and the researchers assured the participants that their data would be anonymized. The 1964 Helsinki Declaration items 21 CFR 56 (Institutional Review Boards, IRB) and 21 CFR 50 (Protection of Human Subjects) were adhered in this study. This study received an ethical approval letter of IRB from Wollo University, Institute of Teachers Education and Behavioral Sciences (certificate number: Ref. 419/2023).

4. Discussion This study investigated two important issues: the psychometric properties of the three measures and the mediation analysis. Following the correlation among the constructs in the first research hypothesis, the second hypotheses aimed at testing the psychometric properties of the three measures namely, the Wong Law Emotional Intelligence Scale (WLEIS-S), Internet Addiction Scale (IAS) and Keyes’ Mental Health Continuum-Short Form (MHC-SF) of Amharic version in Ethiopian context. Hence, the Ethiopian Amharic version of IAS, WLEIS-S, MHC-SF were checked Cronbach alpha’s and composite reliability, convergent, discriminant and construct validity and proved the three measures were reliable and valid. The AVE is greater than 0.05, and AVE was less than MSV, squared correlation of the sub constructs which we used with a sample of Ethiopian university students. In the second goal of the study as seen in the hypothesis 5 (see Fig 7) that internet addiction (IA) predicts mental health (MH) as the dependent variable, with Emotional Intelligence (EIQ) serving as a mediating variable. The study also examined the relationship between IA and MH, with EI as a mediating variable. The results showed a negative correlation between IA, EIQ and MH, indicating that higher levels of IA were associated with lower levels of EIQ and MH. However, EIQ and MH were positively correlated, suggesting that higher levels of EI were associated with better MH. The negative prediction of internet addiction on both EIQ and mental health highlights the detrimental effects of excessive internet use on the mental well-being of university students. Internet addiction is a growing concern in today’s digitally connected world, with students being particularly vulnerable due to the extensive use of online platforms for study, communication, and entertainment. In addition, Internet Craving directly predicted EIQ and indirectly and negatively predicted mental health through EIQ. However, Addictive behavior directly and positively predicted EIQ and indirectly predicted MH. This finding needs further exploration cross culturally. These findings are consistent with several studies suggesting a link between Emotional Intelligence (EIQ) and Mental Health (MH). The hypothesis supported by the previous scientific evidence that internet addiction has a negative impact on mental health, showing a strong relationship between them [3,6,10–14,17]. Additionally, [57] conducted a study on "gender-linked personality and mental health: The role of trait emotional intelligence" and found a significant negative relationship between Emotional Intelligence and mental health. Other studies have also found that higher levels of internet addiction are associated with lower Emotional Intelligence and poorer mental health. For example, [6,57–58], found that higher levels of internet addiction were linked to lower Emotional Intelligence and poorer mental health. Similarly, [6,59] found that internet addiction was associated with lower moral values and psychological well-being, while [49,52–53] discovered a negative influence on emotional intelligence. Furthermore, [26–28] found that addiction to social networking sites (SNS) and higher levels of internet addiction were connected to higher levels of moral disengagement and poorer mental health. These findings have significant implications for interventions and preventive measures aimed at addressing internet addiction and promoting mental health among university students. Efforts should be directed towards raising awareness about the potential risks associated with excessive internet use and providing strategies for developing healthy digital habits. In addition, interventions should focus on enhancing emotional intelligence by promoting activities that facilitate the appraisal of others’ emotions, regulation of emotions, self-appraisal of emotions, and effective use of emotions. This finding is consistent with previous research that highlights the protective role of these factors in promoting mental well-being [6,60–63]. It suggests that interventions aimed at improving emotional intelligence can positively impact mental health, particularly among university students who often face various stressors and challenges during their academic journey. University counseling services and educational programs can play a vital role in supporting students in developing coping skills, managing stress, and maintaining a healthy balance between online and offline activities. By providing resources and guidance, these services can contribute to the overall well-being of students and help them navigate the demands of both their online and offline lives. The second main goal of the study was the mediating role of Emotional Intelligence in the relationship between internet addiction and mental health. There are several possible explanations for the observed mediation effect of Emotional Intelligence. First, excessive internet use can lead to diminished engagement in offline activities, such as social interactions, physical exercise, and face-to-face communication, which are crucial for developing and maintaining better emotional intelligence capabilities. Second, internet addiction may contribute to feelings of isolation, loneliness, and decreased self-esteem, which can further deplete Emotional Intelligence and negatively impact mental health. Third, excessive internet use might disrupt sleep patterns, leading to fatigue and impaired cognitive functioning, which can directly influence Emotional Intelligence and mental health as well as indirectly affect mental health of university students. The study suggests that positive interventions focusing on using Emotional Intelligence, along with the development of healthy digital system, improves the mental health of university students and their daily functioning. These interventions may lead to personal and organizational development and growth. This line of reasoning is supported by prior studies that have found Emotional Intelligence to be the best predictors of mental health and positive outcomes on students’ life, reducing stress, and fostering healthy digital functioning [10,13,17,61–63]. Moreover, Emotional Intelligence has been identified as a preventive resource that can be used to improve mental health and lower internet addiction, and empirical evidence has shown a negative relationship with internet addiction and a positive relationship with mental health, which supports the hypotheses of this study. It is important to note that this study focused specifically on university students, and further research is needed to generalize these findings to other populations. Additionally, the study relied on self-report measures, which may be subject to biases and limitations. Future research could employ longitudinal designs to examine the causal relationships between Internet Addiction, Emotional Intelligence, and Mental Health, as well as explore potential moderators and other mediating factors that may contribute to the observed associations. In conclusion, this study provides valuable insights into the impact of internet addiction on mental health in university students. The findings underscore the importance of addressing internet addiction and promoting Emotional Intelligence as key factors in enhancing mental well-being among this population. These findings contribute to understanding the relationships between IA, EI, and MH and highlight the importance of positive interventions and the better mental health and healthy internet usage and positive EI for promoting mental well-being among university students. By recognizing and addressing these issues, universities and relevant stakeholders can contribute to the overall holistic development and mental health of their students.

5. Conclusion The purpose of the current study was to test the psychometric suitability of the three measures namely IAS, EIS-16 and MHC-SF and examine the mediation role of emotional intelligence (EIQ) between internet addiction (IA) and Mental Health (MH). The results showed that IA had a significant and negative direct effect on Emotional Intelligence and MH. Additionally, Emotional Intelligence had a positive direct impact on MH and fully mediated the relationship between IA and MH. This indicates that higher levels of Internet Addiction among university students are associated with lower levels of Emotional Intelligence, which encompasses positive emotional resources such as others’ emotion appraisal, regulation of emotions, self-emotion appraisal and use of emotions. Furthermore, lower Emotional Intelligence is associated with poorer mental health outcomes. The finding that Emotional Intelligence fully mediates the relationship between internet addiction and mental health suggests that the detrimental effect of internet addiction on mental health is primarily driven by its impact on Emotional Intelligence resources. However, emotional intelligence partially mediated the relationship among dimensions of IA and mental health. In other words, when university students have higher levels of internet addiction, it leads to lower Emotional Intelligence, which, in turn, contributes to poorer mental health outcomes. These results emphasize the importance of addressing internet addiction in university students, as it not only directly affects mental health but also indirectly influences mental health through its impact on Emotional Intelligence. These findings underscore the importance of implementing interventions and preventive measures that target both internet addiction and the promotion of Emotional Intelligence resources among university students. By reducing internet addiction and fostering Emotional Intelligence, we can enhance mental well-being and mitigate the negative effects of internet addiction on mental health. Practitioners and researchers should consider incorporating strategies that aim to reduce excessive internet use and enhance Emotional Intelligence skills in their programs and interventions. This may involve providing educational resources, counseling services, and workshops that focus on improving Emotional Intelligence competencies such as self-awareness, self-regulation, empathy, and relationship management. By addressing both internet addiction and Emotional Intelligence, we can support the mental well-being of university students and promote healthier online and offline behaviors. Overall, the findings emphasize the need for a comprehensive approach to addressing the interplay between internet addiction, Emotional Intelligence, and mental health among university students. By targeting these factors simultaneously, we can develop effective strategies to enhance mental health outcomes and promote overall well-being in this population.

6. Implications and future research suggestions 6.1. Implications for the researchers The psychometric properties of the measures (Wong Law Emotional Intelligence Scale, Internet Addiction Scale, and Keyes’ Mental Health Continuum-Short Form) have been established in the Ethiopian Amharic context. Researchers can confidently use these measures in future studies involving university students. The study confirms the negative impact of internet addiction on both Emotional Intelligence (EIQ) and mental health (MH) among university students. This highlights the need for interventions and preventive measures to address internet addiction and promote mental well-being. The study highlights the mediating role of Emotional Intelligence between internet addiction (dimensions and general factor) and mental health. Researchers can explore further the underlying mechanisms and processes through which Emotional Intelligence influences the relationship between internet addiction and mental health. The study emphasizes the importance of raising awareness about the potential risks associated with excessive internet use and providing strategies for developing healthy digital habits among university students. Researchers can collaborate with educational institutions to design and implement interventions targeting digital well-being and emotional intelligence enhancement. 6.2. Suggestions for future research Generalization. While this study focused on university students, future research should aim to generalize these findings to other populations, such as different age groups or socio-cultural contexts. This would provide a more comprehensive understanding of the relationships between internet addiction, Emotional Intelligence, and mental health across diverse populations. Longitudinal Studies. Employing longitudinal designs would allow researchers to explore the causal relationships between internet addiction, Emotional Intelligence, and mental health. Longitudinal studies would provide insights into the temporal dynamics and potential directionality of these relationships. Moderators and Mediators. Future research could investigate potential moderators and other mediating factors that may influence the associations between internet addiction, personality traits, positive psychological capital, emotional intelligence, mindfulness and mental health [6,10,13,17,63]. For example, variables such as social support, social capital, self-esteem, or coping strategies could be examined to better understand the complex interplay between these constructs. Intervention Studies. Conducting intervention studies that focus on improving Emotional Intelligence and promoting healthy digital habits among university students would provide valuable insights into effective strategies for enhancing mental well-being. These studies can evaluate the impact of interventions on reducing internet addiction and improving mental health outcomes. Comparative Studies. Comparing the impact of different types of online activities (e.g., social media use, gaming, browsing) on Emotional Intelligence and mental health could help identify specific areas of concern and inform targeted interventions. Use of Objective Measures. While self-report measures are commonly used, future research could consider incorporating objective measures (e.g., behavioral observations, physiological indicators) to complement self-report data and provide a more comprehensive understanding of the relationships between internet addiction, Emotional Intelligence, and mental health. By addressing these research implications and pursuing future studies based on the data, researchers can further advance the knowledge in this field, contribute to the development of effective interventions, and promote the well-being of university students in the digital age.

Supporting information S1 File. All confirmatory Factor Analysis Results and Structural Equation Modelling outputs. https://doi.org/10.1371/journal.pdig.0000639.s001 (ZIP)

[END]
---
[1] Url: https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000639

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.