Elsevier

Computers in Human Behavior

Volume 89, December 2018, Pages 230-236
Computers in Human Behavior

Full length article
Succumb to habit: Behavioral evidence for overreliance on habit learning in Internet addicts

https://doi.org/10.1016/j.chb.2018.08.030Get rights and content

Highlights

  • Internet addiction should be considered as a transition process from goal-directed to habit-based.

  • In instrumental learning, prolonged training rendered Internet addicts less sensitive to outcome devaluation.

  • It is hard for Internet addicts to adjust behavior during goal reevaluation.

  • Internet addicts have an enhanced propensity to develop stimulus-driven habits and overly rely on it.

Abstract

Studies of Internet addiction (IA) under the cognitive-behavioral framework have suggested that IA derived from the excessive expectation of positive outcomes, which was a goal-directed behavioral progress. However, according to the dual-system account under the framework of associative learning theory, IA should be considered as a transition process from goal-directed to habit-based. It can be explained in terms of aberrant learning processes, where Internet addicts apparently succumb to habit with decreased sensitivity to reward devaluation. To test this hypothesis, we implemented an instrumental learning paradigm to investigate the difference of reliance on habit between Internet addicts and non-addicts. A total of 21 Internet addicts and 23 normal control subjects were trained in the first experimental phase and learned to associate stimuli with rewarding outcomes by pressing correct keys. In the subsequent outcome devaluation test and in the stimulus-response (S-R) habit test, subjects were required to adjust responses to changes in the value of current outcomes. Results revealed that (1) all subjects were able to gradually acquire the correct behavioral responses in the training phase, however, (2) extensive training rendered Internet addicts less sensitive to outcome devaluation in the outcome devaluation test, and (3) Internet addicts insisted on responding to previous stimuli in the S-R habit test, regardless of whether their responses were rewarded. Thus, we conclude that Internet addicts have an enhanced propensity to develop stimulus-driven habits and overly rely on it, leading to the failures to adjust behavior during goal reevaluation.

Introduction

Think back to the first time you accessed the Internet and recall how it felt when you began to contact or play games with others online. During that time, an Internet-connected computer was like the door to a new world that excited you like nothing in your life had ever done before and had taken you a lot of time to indulge in it. Nevertheless, this situation usually did not last too long and then you might get everything back on the track. Unfortunately, a small proportion of people got stuck in the cyber world and lose the ability to control Internet use, which represented a syndrome known as Internet addiction (IA) and would result in preoccupation and feelings of restlessness when attempting to cut back (Young, 1998, Young, 2004).

Several plausible models have been proposed to explain and predict IA, the most influential of which is the cognitive-behavioral model developed by Davis (2001). In this model, maladaptive cognitions (such as low self-efficacy, low self-esteem, and negative self-evaluation) and the reinforcement an individual receives from various Internet activities play critical roles in the occurrence and maintenance of IA. Davis's theory has “inspired an upsurge of psychometric cognitive-behavioral studies” (Griffiths, Kuss, Billieux, & Pontes, 2016). With the application of neuropsychological and neuroimaging techniques in the field of IA, some similar but more neurobiological-related cognitive-behavioral models have been proposed (Brand et al., 2014, Dong and Potenza, 2014).

Under the cognitive-behavioral framework, IA is considered the result of the reciprocal influence between individuals' maladaptive cognition or characteristics and the reinforcement of the Internet (Davis, 2001). On one hand, individuals with various cognitive problems have strong motivations to seek novelty and/or escape from the “bitter” reality (Cheng and Li, 2014, Choi et al., 2014, Dong et al., 2013). They then are easier to immerse in the online virtual world to achieve the “profit and avoid harm” utilitarian purpose. On the other hand, the Internet caters to this need and the rewarding effect of Internet activities could in turn reinforce the use of Internet (Brand et al., 2014, Davis, 2001). For instance, using social-networking sites can satisfy the sense of belonging, leading individuals who feel isolated in the offline world to addict to social networking (Gao et al., 2017, Griffiths, 2013). Therefore, according to the cognitive-behavioral framework, IA arises due to the excessive expectation of positive outcomes. It is such a process that can be described as a special explicit motivation-driven, or more succinctly, goal-directed behavioral progress. However, the question then becomes whether the Internet addicts' motivation in the use of Internet has always been goal-directed.

According to the dual-system account under the framework of associative learning theory (Dickinson and Balleine, 1993, de Wit and Dickinson, 2009), behaviors are jointly regulated by goal-directed and habitual brain systems. When the goal-directed system plays a dominant role, behaviors are performed to achieve desirable goals or to avoid undesirable outcomes (Gillan et al., 2011). It makes us sensitive to the changing behavioral outcome values and appropriately adjusts our actions to maximize the benefit. Nevertheless, it is less efficient because of the higher consumption of cognitive resources. On the contrary, after constant repetitions, an action belongs to the stimulus-response (S-R) process and the habitual system can occupy a dominant position. This is fast in responding without the need for information processing about consequences, although inflexible in adapting strategy to new conditions (Bai and Zheng, 2011, Behrens et al., 2013, Behrens et al., 2011). These two systems both have strong adaptive advantages and usually work together in a balanced state to prompt efficient and flexible behaviors. The bulk of extant research have suggested that the disruption of this balance contributed to the development and maintenance of addiction (Belin et al., 2013, Ersche et al., 2016, Everitt and Robbins, 2005, Everitt and Robbins, 2016, Hogarth et al., 2013, McKim and Boettiger, 2015). Drug addiction researchers defined addiction as a transition process of addictive behavior from goal-directed to habit-based, and it can be regarded as a maladaptive S-R habit (Everitt and Robbins, 2005, Everitt and Robbins, 2016). That is, the shift from casual to compulsive drug use despite negative consequences is based on increased habit formation at the expense of goal-directed control.

Similarly, an obsession with the Internet is a learned behavior initially directed toward seeking enjoyment or avoiding discomfort. It should be also considered as a transition process from goal-directed to habit-based. In the early days of exposing to the Internet, Internet addicts, like most people, went online based on the expectations of rewards (no matter appetitive or avoidant). Then, with increased duration of Internet access and repetitive use of Internet applications, the nature of Internet addicts' motivation has changed significantly. At this time, if the Internet-related signals emerge, they can trigger compulsive automatic Internet use (Ko et al., 2013, Niu et al., 2016), even when the Internet addicts face or are aware of the extremely adverse consequences. Therefore, excessive Internet use may gradually “deteriorate into a stimulus-driven habit that is elicited by antecedent stimuli and is thus performed regardless of any goals” (Ersche et al., 2016).

The current supposition implicates that, for Internet addicts, the explicit motivation to achieve the “profit and avoid harm” purpose on the Internet that cognitive-behavioral framework has underlined is gradually degrading, namely, “reward devaluation” (Horstmann et al., 2015), while the behavior becomes increasingly automatic, inflexible and habitual. Some existent evidences indirectly supported this hypothesis. First, for example, a cross-lagged panel questionnaire survey showed that excessive Internet use would increase the feelings of loneliness over time but not vice versa (Yao & Zhong, 2014), suggesting that Internet addicts' Internet seeking behaviors do not depend on positive expectations, and even, go against the expectations to reduce discomfort. In fact, using the dual-system perspective in the online context, several intriguing questionnaire studies have recently pointed out that mobile phone use and the experience of momentary social media addiction symptoms were driven by tug-of-war between social self-regulation and habit (Osatuyi and Turel, 2018, Soror et al., 2015, Turel and Qahri-Saremi, 2016, Turel, 2015). Second, an experimental study using a switching paradigm task demonstrated that Internet addicts presented decreased mental flexibility and inhibition towards Internet-related stimuli (Zhou, Yuan, & Yao, 2012). Finally, neurological research has found that activation within the dorsal striatum was associated positively with the duration of IA (Kim et al., 2011, Lu et al., 2017), while the dorsal striatum is believed to be involved in the process of habit building (Bai and Zheng, 2011, Everitt and Robbins, 2005).

Diverging from the previous cognitive-behavioral framework, we conceptualize IA as the transition from initial voluntarily Internet use to habitual, compulsive Internet use. It can be explained in terms of aberrant learning processes, where Internet addicts apparently succumb to habit with decreased sensitivity to reward devaluation, that is, overly rely on inflexible habit system (Sjoerds et al., 2013, Yin and Knowlton, 2006).

In the present study, we observed the differences in learning processes between Internet addicts and non-addicts to directly examine our view. We adopted an instrumental learning paradigm initially designed to distinguish between goal-directed and habit-based learning (de Wit, Niry, Wariyar, Aitken, & Dickinson, 2007). This paradigm has been used in several fields, such as obsessive-compulsive disorder (Gillan et al., 2011), drug dependence (Ersche et al., 2016, Sjoerds et al., 2013), and Parkinson disease (de Wit, Barker, Dickinson, & Cools, 2011). This paradigm consists of three sequential phases as depicted in Fig. 1. In the training phase, subjects learn to respond (R) to stimuli (S) to collect outcomes (O). Then in the outcome devaluation test, subjects need to use their knowledge of the R-O associations to make correct response between a devalued outcome and a still-valuable outcome. Higher error rate would indicate more reliance on the habitual S-R associations regardless of outcomes during the previous training phase. Finally, the S-R habit test can create conditions in which the goal-directed and habitual systems directly compete for behavioral control. Responses to the stimuli that are no longer associated with valuable outcomes indicated decreased sensitivity to reward devaluation, implying that the habitual system occupies a dominant position.

Based on the literature review, we predicted that Internet addicts would be more likely to adopt the S-R habitual learning strategy in behavior acquisition, that is, more dependent on the habitual system. Thus, we proposed the following hypotheses.

Hypothesis 1

During the training phase, both Internet addicts and non-addicts would be able to gradually acquire correct behavioral responses.

Hypothesis 2

During the outcome devaluation test, Internet addicts would exhibit a higher error rate than non-addicts.

Hypothesis 3

During the S-R habit test, (a) Internet addicts would make more responses to stimuli with devalued outcomes than non-addicts, and (b) the degree of IA would be positively related to habitual responses.

Section snippets

Subjects

All subjects were recruited by advertisements at local universities and participated in initial screening through the Chen Internet Addiction Scale (CIAS) online or offline. The CIAS is a widely used criterion for IA in Mainland China and Taiwan. The CIAS contains 26 items rated on a four-point scale (e.g., “I cannot control my impulsivity to use the Internet”). A total of 566 individuals completed this scale. The coefficient alpha of CIAS in the present study was 0.92. Using the recommended

Training phase

The response accuracy of IA group and NC group on three discriminations during the training phase was shown on Fig. 2, revealing that learning performance improved steadily in all subjects over 12 training blocks, F (11, 462) = 24.92, p < .01, partial η2 = 0.37. In addition, there was a significant main effect of discrimination, F (2, 84) = 13.59, p < .01, partial η2 = 0.25. Bonferroni post hoc comparisons further showed that subjects performed the worst under the incongruent discrimination

Discussion

The current study examined the roles of and goal-directed and habitual systems in behavior acquisition among Internet addicts and non-addicts. In a set of tasks using the instrumental learning paradigm (de Wit et al., 2007), we found evidence that Internet addicts were more likely to adopt S-R habitual learning strategy, that is, more dependent on the habitual system, consistent with the research hypotheses. Both Internet addicts and non-addicts were able to use feedback to gradually acquire

Declarations of interest

None.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grant umber 71771102). We are very grateful to Dr. Yun Tang for her help in writing and revising the manuscript in English, as well as for the assistance she has always provided.

References (53)

  • A. Horstmann et al.

    Slave to habit? Obesity is associated with decreased behavioural sensitivity to reward devaluation

    Appetite

    (2015)
  • C.H. Ko et al.

    The association between internet addiction and psychiatric disorder: A review of the literature

    European Psychiatry

    (2012)
  • C.H. Ko et al.

    Screening for internet addiction: An empirical study on cut-off points for the chen internet addiction scale

    The Kaohsiung Journal of Medical Sciences

    (2005)
  • J. Nie et al.

    Impaired inhibition and working memory in response to internet-related words among adolescents with internet addiction: A comparison with attention-deficit/hyperactivity disorder

    Psychiatry Research

    (2016)
  • G.-F. Niu et al.

    Cue-induced craving for Internet among Internet addicts

    Addictive Behaviors

    (2016)
  • B. Osatuyi et al.

    Tug of war between social self-regulation and habit: Explaining the experience of momentary social media addiction symptoms

    Computers in Human Behavior

    (2018)
  • E. Pool et al.

    Measuring wanting and liking from animals to humans: A systematic review

    Neuroscience & Biobehavioral Reviews

    (2016)
  • T.E. Robinson et al.

    The neural basis of drug craving: An incentive-sensitization theory of addiction

    Brain Research Reviews

    (1993)
  • M.Z. Yao et al.

    Loneliness, social contacts and internet addiction: A cross-lagged panel study

    Computers in Human Behavior

    (2014)
  • B.A. Anderson et al.

    Attentional bias for nondrug reward is magnified in addiction

    Experimental and Clinical Psychopharmacology

    (2013)
  • Y. Bai et al.

    The role of the cortex-dorsal striatum pathway in morphine-induced compulsive drug seeking and taking behaviors

    Advances in Psychological Science

    (2011)
  • T. Behrens et al.

    Actions, action sequences and habits: Evidence that goal-directed and habitual action control are hierarchically organized

    PLoS Computational Biology

    (2013)
  • T. Behrens et al.

    Speed/accuracy trade-off between the habitual and the goal-directed processes

    PLoS Computational Biology

    (2011)
  • M. Brand et al.

    Prefrontal control and internet addiction: A theoretical model and review of neuropsychological and neuroimaging findings

    Frontiers in Human Neuroscience

    (2014)
  • C. Cheng et al.

    Internet addiction prevalence and quality of (real) life: A meta-analysis of 31 nations across seven world regions

    Cyberpsychology, Behavior, and Social Networking

    (2014)
  • S. Chen et al.

    Development of a Chinese Internet addiction scale and its psychometric study

    Chinese Journal of Psychology

    (2003)
  • Cited by (8)

    • The imbalance between goal-directed and habitual systems in internet gaming disorder: Results from the disturbed thalamocortical communications

      2021, Journal of Psychiatric Research
      Citation Excerpt :

      Instrumental learning paradigms have provided a theoretical framework of impaired behavioral control for addiction by assessing the dynamic interactions between the goal-directed system and the habitual systems (Everitt & Robbins, 2005, 2016). Our previous study showed that, in the habit test of the instrumental learning paradigm, IGD patients insisted on responding to previous reward stimuli even when the incentive outcome did not follow (B. Zhou, Wang, Zhang, Li and Nie, 2018). This insistence suggests that IGD patients have an enhanced propensity to develop reward-driven habits and tend to be over-reliant on them, thus failing to adjust behavior during goal reevaluation (S. de Wit et al., 2007).

    View all citing articles on Scopus
    View full text