Dietary patterns and smartphone use in adolescents: A large-sample, cross-sectional study

Background: This study aimed to examine whether the frequency of consuming specic foods was associated with smartphone usage time and the problems caused by smartphone overuse in adolescents. Methods: Both routines (food consumption and smartphone use) were investigated in a large sample of Korean adolescents using a nationwide self-report survey. Food intake was assessed using a seven-point scale (“never” to “1, 2, and 3 or more times per day”) for nine items: fruits, vegetables, milk, soda, energy drinks, sweetened beverages, fast food, instant noodles, and snacks. Smartphone usage time and problematic usage were determined using self-report items. The study took place in Korea. In total, 62,276 students aged 12–18 years participated. Results: Most respondents (66.5%) used smartphones for over two hours per day. Higher rates of consuming fruits, vegetables, and milk were associated with signicantly lower smartphone usage; in contrast, higher consumption of soda, energy drinks, sweetened beverages, fast food, instant noodles, and snacks was associated with higher smartphone use. Conclusions: Our ndings provide useful clinical information regarding adolescents’ mental health. Future studies should investigate underlying mechanisms and examine the ecacy of adopting dietary interventions for adolescents with excessive smartphone use.


Introduction
Adolescence is a period of signi cant emotional, cognitive, and behavioral developmental changes, and these changes are strongly in uenced by various social factors at the individual, family, community, and national levels [1]. Notably, the mental health of adolescents is associated with various daily healthrelated behaviors, such as sleep [2], physical activity [3], diet [4], screen time [5], substance abuse [6], and school violence [7].
Since the unveiling of the iPhone in 2007 [8], the smartphone has rapidly penetrated the daily life of both adolescents and adults, and it has now become an important daily routine in the lives of adolescents. In 2015, 92% of teens were reported to go online daily, and only 8% of teens were online less than once per day [9]. In 2018, 95% of U.S. adolescents were reported to have their own smartphones or access to one, and 45% of teens were online on a near-constant basis [10]. Importantly, previous research has suggested that excessive smartphone use is associated with unfavorable outcomes for adolescents' mental health, such as depression and anxiety [11] [12].
Notably, the problematic or maladaptive use of smartphones by adolescents has rapidly become a widespread social concern [13]. Researchers have not yet reached a consensus regarding smartphone addiction in adolescence, which has led to the use of various methodologies and concepts in the study of this phenomenon, including addiction, dependence, problematic use, and abuse [13]. Despite the lack of conceptual delimitation, several studies have reported substantial concerns for smartphone addiction with a wide prevalence range of 0.4-64% [13]. Among the various biopsychosocial factors related to the addictive disorders, impulsivity and sensation seeking are considered as the most important individual characteristics predicting smartphone abuse [14][15] [16].
Dietary habits are another important daily routine associated with adolescents' mental health. Although eating was originally considered a hedonic behavior stimulating the brain reward circuit, there is controversy around framing addictive-like eating problems as "eating addiction" versus "food addiction" [17] [18]. Whereas the framework of "eating addiction" suggests a behavioral addiction triggering an addictive-like response in susceptible individuals, the "food addiction" model focuses on the questions of whether certain foods provoke more addictive-like eating behavior than other foods. Despite this controversy, multiple studies have reported that some kinds of food might affect neurobiological changes, provoking addictive-like eating more than other foods [19]. For instance, frequent consumption of highly processed food with high fat, salt, and sugar content, such as soda, pizza, and snacks, increases the individual's impulsivity and craving for food [20][21] [22] [23]. On the other hand, the Mediterranean diet, which includes ber-rich fruit and vegetables, has been reported to decrease impulsivity [24].
Interestingly, it has been suggested that the overuse of smartphones and the high consumption of some speci c dietary contents share a neurobiological mechanism that might be associated with addictive-like behavior. However, to our knowledge, there is a lack of studies on the direct associations between smartphone overuse and speci c dietary patterns. Thus, our study aimed to examine whether the frequency of consuming speci c foods is associated with smartphone usage time and the problems caused by smartphone overuse.

Participants
Our study analyzed data from the 2017 Korean Youth Risk Behavior Web-based Survey (KYRBS), which is a nationwide, self-report, cross-sectional study conducted by the Korean Centers for Disease Control and

Measures
To assess food intake frequency, a self-rated questionnaire was used that included nine items: fruits, vegetables, milk, soda, caffeinated energy drinks, sweetened beverages, fast food (i.e., pizza, hamburgers, and fried chicken), instant noodles, and snacks. Consumption of each item was rated on a seven-point rating scale (never, 1 to 2, 3 to 4, and 5 to 6 times per week; and 1, 2, and more than 3 times per day).
Smartphone usage time was assessed with the following item: "How many hours and minutes per day have you used a smartphone in the last 30 days?" Smartphone usage time was categorized into weekdays and weekends, and the average daily hours of use on weekdays and weekends were used in the subsequent analysis.
The problems related to con icts with the family and disturbances in school performance caused by smartphone use were assessed using one item for each problem as rated on a four-point Likert scale, with responses from 1 = strongly disagree to 4 = strongly agree. The items were "I have experienced severe con icts with my family because of smartphone use" and "Due to the use of smartphones, there are di culties in carrying out my studies." In our study, the high-risk group for problems caused by smartphone use was de ned based on responses of agree or strongly agree to each item.

Procedure
The students were provided with detailed instructions about the study by trained teachers before completing the survey. Students who agreed to participate responded to an anonymous, selfadministered, web-based questionnaire presented on a computer screen. Before participation in the survey, the students were provided with an online informed consent form which was signed before completing the questionnaire. The steering committee representing the O ce of Education from 17 provinces of Korea oversaw the survey.

Statistical analysis
Smartphone usage time and problems caused by smartphone use were compared among the groups according to the consumption frequency of each food. For the comparison of smartphone usage time, the groups of "1, 2, and more than 3 times per day" were merged into one group of "more than 1 time," given the small proportion of participants in those frequency groups. Differences in smartphone usage time among the food intake frequency groups were analyzed by three models. Model 1 was analyzed using analysis of variance, with the consumption frequency of each kind of food as the independent variables and smartphone usage time as the dependent variable, while Model 2 was analyzed using analysis of covariance, including sex, age, and socioeconomic status as the covariates. Model 3 included parental educational levels as covariates in addition to those in Model 2. Post-hoc analyses were conducted using Tukey's honest signi cant difference test in Model 2 due to many missing values for parental educational level. The odds ratio for the high-risk group of con ict with family and disturbances in school performance caused by excessive smartphone use was analyzed using a logistic regression model with the categories of foods as the independent variables. The statistical analyses were conducted using the software package SPSS 25.0 for Windows (IBM Corp., Armonk, NY, USA).

Demographic characteristics
The demographic characteristics of the participants are shown in Table 1. Among the 62,276 participating youths, 31,624 (50.8%) were male. A small proportion of adolescents (12.3%) had not used a smartphone in the past 30 days. Two-thirds (66.5%) used a smartphone for over 2 hours per day and 5.3% used a smartphone for more than 10 hours per day.

Association between food intake frequency and smartphone usage time
Food intake frequency is shown in Table 2. Responses of "1 to 2 times per week" had the most prevalent frequency for fruits, milk, soda, sweetened beverages, fast food, instant noodles, and snacks; "3 to 4 times per week" was most prevalent for vegetables, and "never" was most prevalent for caffeinated energy drinks. Responses of "2 times per day" and "more than 3 times per day" for soda, caffeinated energy drinks, sweetened beverages, fast food, instant noodles, and snacks were very low (under 2%). Table 3, Figure 1, and Figure 2 show the association between the food intake frequency and smartphone usage time. Higher rates of consuming fruits (F = 151.8; p < 0.001), vegetables (F = 119.9; p < 0.001), and milk (F = 33.0; p < 0.001) were associated with signi cantly lower smartphone usage. In contrast, higher consumption of soda (F = 292.5; p < 0.001), energy drinks (F = 24.0; p < 0.001), sweetened beverages (F = 224.8; p < 0.001), fast food (F = 192.1; p < 0.001), instant noodles (F = 196.2; p < 0.001), and snacks (F = 131.6; p < 0.001) were associated with higher smartphone use.
Association between food intake frequency and problems caused by smartphone use The odds ratios for the high-risk groups of con ict with family and disturbances in school performance caused by excessive smartphone use increased with more frequent consumption of soda, caffeinated energy drinks, sweetened beverages, fast food, instant noodles, and snacks ( Table 4). The odds ratios also increased with less frequent intake of vegetables. However, a higher intake frequency of fruit was more strongly associated with an increased odds ratio in con ict with family, whereas milk had no signi cant associations with the ratio for con ict with family.
The scores of the items for problems caused by smartphone use were signi cantly different among the groups of food intake frequencies (Table S1). Higher levels of consuming fruits, vegetables, and milk, and lower levels of consuming soda, caffeinated energy drinks, sweetened beverages, fast food, instant noodles, and snacks were signi cantly associated with fewer problems caused by smartphone use.
However, the average scores of problems caused by smartphone use items were under two ("disagree") in all the food intake frequency groups.

Discussion
The present study investigated the association between excessive smartphone use and dietary habits in Korean adolescents using a cross-sectional design with a large sample. Most participating adolescents (87.7%) reported that they had used a smartphone in the last 30 days, and 66.5% reported using a smartphone for over two hours per day. These ndings indicate the pervasive use of smartphones by adolescents and are comparable with U.S. studies that found that 92% of teens reported going online daily and 45% were online "almost constantly" [10] [9].
The major ndings of the present study are the signi cant associations between speci c dietary patterns and smartphone usage time. Adolescents who consume more soda, sweetened beverages, drinks with a high level of caffeine, fast food, instant noodles, and snacks, which could be labeled as addictive food, spent more time using a smartphone. In contrast, frequent intake of fruits and vegetables was linearly associated with lower levels of smartphone usage. Intake of milk was also associated with lower smartphone usage, although the association was less strong compared to fruits and vegetables. To our knowledge, our study presents novel ndings, in that there have been no previous studies of the association between smartphone use and dietary habits in adolescents. In particular, the linearity of the associations between the consumption frequency of speci c foods and smartphone usage time is interesting, although the effect size of the associations is small. Consistent with our ndings, previous studies have reported associations between psychiatric symptoms and dietary patterns in disorders characterized by impulsivity and vulnerability to addiction, such as Attention-De cit/Hyperactivity Disorder (ADHD) [26][27] [28], binge-eating disorder [29], and gambling [30]. For instance, Kim et al reported linear associations between scores on ADHD symptoms (hyperactivity and inattention) and the consumption frequency of speci c foods in the general population of school-aged children [27].

The fth edition of the Diagnostic and Statistical Manual of Mental Disorders has included internet
gaming disorder in the section "Conditions for Further Study" [31], and gaming disorder was listed as a formal diagnosis in the International Classi cation of Diseases 11th Revision, released in 2018 [32].
Although excessive smartphone use has not yet been agreed upon as a formal mental disorder, the addictive potential of excessive smartphone use should be studied carefully, considering that internet gaming is one of the major uses of smartphones and that smartphones are widely accessible at any time. Impulsivity and altered reward circuit activation in the brain are the shared neurobiological features in the disorders vulnerable to addiction, such as ADHD and substance abuse [33] [34]. These are also the major predictive dimensions of excessive smartphone use [35] [13]. Dietary patterns are also known to affect impulsivity and reward circuit activation [18].
Although the concept of food addiction is a controversial topic and there is no consensus on whether humans show addictive-like eating [37], previous animal studies indicate that high consumption levels of high-fat and high-sugar foods induce neurobiological changes, such as dysfunction of the reward system and downregulation of dopamine receptors observed in other addictive disorders [37][38] [18]. Human studies have also suggested evidence for the addictive potential of hyperpalatable food via neural activation in addictive-like eating behavior and substance dependence, such as altered activation in the reward circuit [39] [40][37] [19].
In contrast to addictive high-fat, high-sugar, and highly processed foods, the Mediterranean diet, which is rich in fruits and vegetables, has been reported to be negatively associated with impulsivity and sensation-seeking [41]. Food addictions are also associated with a lower intake of the Mediterranean diet [36]. These results are consistent with the ndings of the present study. Despite multiple studies suggesting a protective association of fruit and vegetable intake with impulsivity and vulnerability to addiction, the underlying mechanism has not yet been identi ed clearly. However, studies have hypothesized some mechanisms, including phytochemicals, short chain fatty acids, and the antioxidant properties in fruit and vegetables, and have reported positive evidence for their neuroprotective effect [42].
Even though dietary patterns had a signi cant linear association with smartphone usage time in our study, dietary patterns do not necessarily cause smartphone "addiction" directly. For "addiction," in addition to excessive use, other elements (such as social impairment, tolerance, and failure to control behavior) also must be present [13]. Our study found relatively inconsistent associations for the odds ratios for the problems caused by smartphone use among the food intake frequency groups, compared to a linear association between dietary habit and smartphone usage time. These ndings might suggest that the effect of dietary patterns on smartphone use is not so strong as to elicit the label of smartphone "addiction"; however, it may be that diet can contribute to the vulnerability to addiction by changing the underlying neural mechanism of the reward circuit activation. Previous studies also have reported consistent ndings regarding the addictive potential of foods. Twenty-one studies and twelve studies included in a systematic review about "food addiction" reported evidence of brain reward system dysfunction and impaired control according to food intake, respectively, which indicates the existence of a change in neurobiological function caused by dietary habits [19]. However, only two studies reported evidence for social impairment caused by "food addiction" [19]. These studies suggest that addiction involves more social factors in addition to neurobiological change, which is in line with our ndings.

Limitations
There are some limitations to this study that should be noted. First, it used a cross-sectional design, which limits the ability to determine causal relationships. Children and adolescents who are vulnerable to addiction and whose sensation-seeking and impulsive temperament are high might share the common features of smartphone overuse and frequent intake of addictive food. However, previous studies reported that dietary intervention can change the core symptoms in children with ADHD. In addition, Lumley et al. reported that greater impulsivity is associated with consuming a Western-style diet, with bidirectional causation [43]. Considering these previous studies, smartphone usage might be in uenced by long-term dietary patterns. However, longitudinal or experimental studies are needed to con rm the exact causal relationship.
Second, our study was based only on an anonymous, self-reported questionnaire completed by participants, and it is possible that adolescents inaccurately reported their dietary patterns and smartphone use. In addition, although we included items to evaluate the problems caused by smartphone use, validated tools to assess the severity of smartphone abuse in terms of addiction criteria were not adopted in our study. Thus, future studies that include multi-informant reports and objective measures assessing children's behavior are needed to con rm the ndings of our study. Importantly, a previous study examining the correspondence between adolescent and parent reports for psychiatric symptoms found that the adolescents' reports for substance abuse were more accurately correlated with the real diagnosis than the parents' reports [44]. This suggests that reports from adolescents can be viewed as reliable.
Finally, our study was performed using a typically developing population, which limits the generalizability of the ndings to the populations with addictive psychopathology, such as ADHD, substance abuse, and eating disorders. To con rm the present ndings, future studies in groups with psychiatric disorders are warranted.

Conclusion
Diet is a major but modi able daily routine of adolescents, and dietary interventions could be improved by knowing the psychological characteristics related to consuming each kind of food. Despite several limitations, our study is the rst to reveal a relationship between dietary patterns and smartphone overuse with a large sample of adolescents. The present ndings of a linear relationship between dietary patterns and smartphone overuse are interesting. The smartphone is an important environmental factor today that induces great concerns for adolescents' mental health at the family and societal level. Although dietary habits are not a critical contributor to smartphone addiction, future studies investigating interventions with diet and smartphone use could help improve adolescents' mental health.

Declarations
Ethics approval and consent to participate The Institutional Review Board of the Dankook University Hospital (DKUH 2019-07-022) approved the current study.

Consent for publication
Not applicable.

Availability of data and materials
The use of the data from the 2017 Korean Youth Risk Behavior Web-based Survey (KYRBS repository, https://yhs.cdc.go.kr) was permitted by the Korea Centers for Disease Control and Prevention.

Competing interests
The authors declare no con ict of interest.

Funding
This research did not receive any speci c grant from funding agencies in the public, commercial, or notfor-pro t sectors.  Model 1 is a crude model with the consuming frequency of each kind of food as independent variable and smartphone use time as a dependent variable. Model 2 is an adjusted model including sex, age, and socioeconomic status as covariates in addition to model 1. Model 3 included parental educational levels as covariates in addition to model 2 post-hoc : a (never), b (1-2 times/week), c (3-4 times/week), d (5-6 times/week), e (1 or more than 1 time/day) * p < 0.05; ** p < 0.001 Table 4. Odds ratios for the high-risk group of problems caused by smartphone use according to the food intake frequency.

Figure 1
Smartphone usage time among the groups of food intake frequency (hours per day).

Figure 2
Smartphone usage time among the groups of food intake frequency (hours per day).