Games as an early intervention after trauma : can a visuospatial computer game played on a smartphone after experimental trauma reduce the build-up of image-based intrusive memories?

(1)

1 Graduate School of Psychology

Games as an Early Intervention after Trauma: Can a Visuospatial

Computer Game Played on a Smartphone after Experimental Trauma

Reduce the Build-up of Image-Based Intrusive Memories?

Elze Landkroon (6184294) July 2016

Master Thesis Research Master’s Psychology University of Amsterdam

Research conducted at Medical Research Council Cognition and Brain Sciences Unit, Cambridge, United Kingdom

Supervisors

Prof. dr. Merel Kindt (University of Amsterdam)

Dr. Ella James and prof. dr. Emily Holmes (Medical Research Council Cognition and Brain Sciences Unit)

(2)

Abstract

Intrusions are recurrent, involuntary mental images of a traumatic event, and a hallmark

symptom of post-traumatic stress disorder. Research showed that playing a visuospatial computer game (Tetris) after seeing a traumatic film, an experimental analogue for trauma, reduced

subsequent visual intrusions to the film. It was hypothesised that visuospatial tasks, which are of the same modality as the visual images of the film, reduce intrusions by interfering with the visual images of the film in working memory, while tasks of a different modality cannot reduce

intrusions. Given the importance of easily accessible, early interventions after trauma, it was investigated whether playing another seemingly visuospatial game on a smartphone after an analogue trauma also reduced subsequent intrusions. In addition, a word puzzle game was used as an active control condition to investigate whether visuospatial aspects of the game were necessary to reduce intrusions. Using the established trauma film paradigm with new clips from eyewitness media, healthy volunteers (N = 50) played a visual-based game (Candy Crush) or a

word puzzle game (WordBrain) on their smartphones or performed no task after viewing a traumatic film. Contrary to what was expected, participants in the Candy Crush group did not experience fewer intrusions than the other groups. WordBrain reduced intrusions over the week compared to the no-task group, but only when baseline intrusions were taken into account. The modality and workload of these games was not studied here and should be investigated in future research to disentangle the importance of modality and workload of games to reduce intrusive memories. Nevertheless, the study showed that certain games, even when played on a

smartphone, could reduce the build-up of intrusive memories over the course of a week, indicating that game play might be applied as an early intervention after trauma.

Keywords: Intrusive memory, involuntary memory, smartphone game play, trauma film, mental

(3)

Table of Content Abstract p. 2 Introduction p. 4 Methods p. 9 Results p. 16 Discussion p. 24 References p. 31

Appendix A: Description film clips trauma film p. 35

Appendix B: Instructions competing cognitive tasks p. 44

(4)

Introduction

Many people endure trauma during life, after which post-traumatic stress disorder (PTSD) can develop (American Psychiatric Association [APA], 2013). Around 8% of the population develops PTSD at some point in life. Intrusions are the most characteristic symptom of this disorder (APA, 2013) and are described as involuntary, recurrent mental images of the traumatic event, which can occur in all modalities but are most common as visual mental images (Ehlers, Hackmann, Steil, Clohessy, Wenninger, & Winter, 2002). Intrusions are associated with significant distress (Grey & Holmes, 2008), and have been shown to be a precursor of PTSD (O’Donnell, Elliot, Lau, & Creamer, 2007). Currently, Cognitive Behavioural Therapy (CBT) is the most effective treatment for PTSD (National Institute for Health and Clinical Excellence, 2005). However, CBT is only administered at least 4 weeks after trauma, when PTSD can be diagnosed (National Institute for Health and Clinical Excellence, 2005). Presently, no clear evidence-based early interventions exist for people who endured trauma that are short-term and easily applied (Roberts, Kitchiner, Kenardy, & Bisson, 2010; Kearns, Ressler, Zatzick, & Rothbaum, 2012; Litz, Gray, Bryant, & Adler, 2002). Early interventions preventing PTSD or post-traumatic stress symptoms would benefit individuals and society as a whole, since it would reduce both individual’s suffering and healthcare costs. Therefore, there is a pressing need for early interventions that are widely available and easily accessible. Since intrusions can be so debilitating and are common following a traumatic event, research into developing preventative interventions to reduce intrusions is highly needed. The main objective of the present study is to investigate a feasible early intervention to reduce image-based intrusions after trauma with the use of smartphones, which are widely used devices and thus easily accessible.

An example where such an early intervention procedure would be worthwhile and easily applied is in the world of journalism. Many journalists have to watch hours of news footage shot by people with their mobile phones all over the world (Dubberley, Griffin, & Bal, 2015). Examples of this ‘eyewitness’ footage include scenes from war zones, terrorist attacks and car crashes. Journalists watch such footage to assess whether the footage is authentic and to decide whether it should be spread through media channels to the public. Furthermore, these journalists do not know what the footage will contain before they watch it. Twenty-two percent of journalists have to watch over six hours of this ‘eyewitness’ footage a day (Dubberley et al., 2015). About forty percent of journalists state that viewing distressing eyewitness media has a negative impact on their personal lives, for example because they develop intrusive memories as a result. This is supported by the addition in the DSM-5 that viewing traumatic media footage as part of professional work is recognised as a ‘traumatic event’ (APA, 2013). Therefore, journalists

(5)

would benefit from an easy accessible early intervention to reduce intrusions.

Several clinical models try to conceptualize PTSD and its symptoms, such as intrusive memories. Even though these models differ from one another in some respects, they all try to explain the development of intrusive memories by highlighting the information processes involved during encoding and consolidation of traumatic events (Ehlers & Clark, 2000; Brewin, Gregory, Lipton, & Burgess, 2010). Two forms of information processing are defined: conceptual processing and sensory perceptual processing. Conceptual processing refers to giving meaning to an event and placing it in context, while sensory perceptual processing refers to the processing of the perceptual details of an event (Ehlers & Clark, 2000; Brewin et al., 2010). During conceptual processing of a traumatic event, the perceptual details can be contextualized. For individuals experiencing trauma, a shift in processing balance might occur due to the emotional response in reaction to a traumatic event, which is thought to enhance perceptual memory while this does not enhance conceptual memory (Arntz, de Groot, & Kindt, 2005). When there is relatively more sensory perceptual processing of a traumatic event compared to conceptual processing, this might lead to the development of intrusions as the perceptual details are not properly contextualized and therefore intrusions can be perceived as reliving the event (Holmes & Bourne, 2008).

Hypothetically, reducing the number of sensory image-based intrusions after trauma can be achieved by reducing sensory perceptual processing of a traumatic event (Ehlers & Clark, 2000; Brewin et al., 2010). Visuospatial tasks performed during or after a traumatic event are hypothesized to reduce image-based intrusions of a traumatic event because both the visuospatial task and the visual aspects of a trauma memory rely on the limited resources of Working Memory (WM) and therefore reduces sensory perceptual processing of the traumatic event (Baddeley, 2003, 2012). WM is a system that can maintain, store and manipulate information. It consists of a central executive to integrate information and to control two slave systems, the phonological loop to maintain auditory information, and the visuospatial sketchpad (VSSP) to process and manipulate visual and spatial information. Research showed that the VSSP has limited capacity, since performance on a visuospatial task can be disrupted by a concurrent dual task of the same modality, but to a lesser extent by tasks of a different modality (Logie, Zucco, & Baddeley, 1990). New information can be stored from WM in short-term memory (STM), where it is stored but cannot be manipulated, and lasts seconds to hours. These newly formed memory traces are transiently labile and vulnerable to disruption and become stable and consolidated over time in long-term memory (LTM), which lasts days to weeks (Nader, 2003; McGaugh, 2000). Following an event a window of at least six hours seems to exist in which memories are labile and become

(6)

consolidated into LTM. Epinephrine seems to modulate memory consolidation by stimulating β-adrenergic receptors, which are located peripherally and project to the brainstem, which again influences release of norepinephrine in the amygdala. In rats, stimulation of β-adrenergic receptors in the amygdala enhances memory consolidation through a process of protein synthesis. Studies in humans indicate that just as in rats, β-adrenergic receptors in the amygdala influence memory consolidation (McGaugh, 2002). Memory for emotional events are enhanced compared to neutral events (Cahill, Prins, Weber, & McGaugh, 1994). Blocking these β-adrenergic receptors has been shown to impair the enhanced memory for emotional events in humans, suggesting that a higher stimulation of β-adrenergic receptors takes places during emotional events to enhance memory (Cahill, et al., 1994). During emotional events information is processed via a direct route to the amygdala instead of the indirect route involving the hippocampus (LeDoux 1996). The amygdala has been shown to be important in implicit fear memory, while the hippocampus has been shown to play a key role in declarative memory, e.g. the contextualization of events (LaBar & Cabeza, 2006). The more pronounced role of the amygdala compared to the hippocampus during emotional events might lead to the increased sensory perceptual processing compared to conceptual processing and thus lead to the development of intrusions (Brewin et al., 2010).

The modality-specificity hypothesis states that a modality-specific cognitive task performed during or after trauma within the memory consolidation time-window may reduce image-based intrusions, because the task competes with the visual aspects of the trauma memory for the limited resources of the VSSP in WM (Holmes, James, Coode-Bate, & Deeprose, 2009; Holmes, James, Kilford, & Deeprose, 2010). Therefore, there is less capacity in WM to process the visual aspects of the trauma film and therefore will be less sensory perceptually processed.

Previous research on early interventions to reduce image-based intrusions used the trauma film paradigm to experimentally model trauma (for a review see James, Lau-Zhu, Clark, Visser, Hagenaars & Holmes, 2016). The trauma film paradigm involves showing a film with scenes depicting actual or threatened death and serious injury (see Figure 1). Afterwards, participants record intrusive memories of the film in a diary for one week. Earlier research repeatedly showed that the film footage used in the trauma film paradigm can be used to induce intrusive memories, and that it is a useful analogue for real-life trauma (James et al., 2016).

(7)

Concurrent dual task methodology was used in earlier research with the trauma film paradigm in which participants performed tasks while viewing the trauma film. When participants performed a visuospatial pattern-tapping task (tapping keys in a specified sequence) during encoding of the film, this led to a reduction of subsequent intrusions over a week, while participants who performed a verbal task (counting backwards) reported an increase of intrusions compared to participants who did not perform a concurrent dual task (Holmes, Brewin, & Hennessey, 2004). Importantly, deliberate recall of the film one week post-film showed no difference between groups, indicating that deliberate memory is intact. Another study showed again that while visuospatial tasks performed concurrently with the trauma film viewing reduced later visual intrusions of the film, verbal tasks can increase the build-up of intrusions (Bourne, Frasquilho, Roth, & Holmes, 2010). Altogether, these findings suggest that it is not the distracting aspect of a task that accounts for reducing intrusions; rather that task-modality is important. This is in line with theories stating that an unbalance in conceptual versus sensory perceptual processing might increase intrusive memories when there is relatively more conceptual processing than sensory perceptual processing, while the opposite unbalance in information processing might reduce intrusive memories (Ehlers & Clark, 2000; Brewin et al., 2010).

To be able to use cognitive tasks as an early intervention after trauma, they should also reduce intrusions when performed after viewing the trauma film. Tasks administered directly post-film and 30 minutes post-film showed that an established visuospatial cognitive task (complex pattern tapping) reduced intrusions compared to an established verbal-based control task (counting backwards) or no task (Deeprose, Zhang, Dejong, Dalgleish, & Holmes, 2012). Again, there were no differences between groups on recognition memory given one-week after film viewing, suggesting deliberate memory was not affected by task completion post-film (Deeprose et al., 2012). To make this procedure more easily applicable as an early intervention after trauma, this procedure should also work with computer games instead of laboratory tasks.

Computer games have also been used instead of laboratory tasks. For example, performing a visuospatial computer game (Tetris) for 10 minutes half an hour after seeing a trauma film reduced intrusions over the subsequent week compared to no task (Holmes et al., 2009). In a follow-up study, performing a verbal-based computer game (Pub quiz) 30 minutes after seeing a trauma film increased intrusions one week later compared to no task (Holmes et al.,

(8)

2010), while the visuospatial Tetris computer game reduced intrusions of the film. Again,

indicating that a task of the same modality as the trauma film competes with the visual aspects of the film for limited resources of the VSSP, thereby reducing intrusions, while a task of a different modality does not. A second experiment showed similar results even when tasks were performed four hours after viewing the trauma film, which was hypothesised to be still within the

consolidation window of six hours. Since people are not able to perform a cumbersome

laboratory-based cognitive task immediately after experiencing a trauma, these studies with game play in an extended time frame show promise for an early intervention after trauma.

In these abovementioned experimental studies, playing verbal games during or after trauma film viewing, might have led to reduced conceptual processing and therefore to an unbalance in information processing and thus a higher number of intrusions compared to participants who did nothing during or after trauma film viewing. However, not all verbal games showed this increase in intrusion frequency, indicating that not all verbal games interfere with conceptual processing (James et al., 2016). Originally, conceptual processing was linked to verbal processing, but recently this is linked to abstract, context independent representations (Brewin et al., 2010). Therefore, not all verbal tasks will interfere with conceptual processing. On the other hand, performing a visuospatial task during or after viewing the trauma film might have reduced sensory perceptual information processing of the trauma film and thus led to a lower number of intrusions compared to participants who did nothing during or after trauma film viewing.

In sum, previous research showed that a visuospatial computer game played on a

computer after seeing a trauma film reduced subsequent intrusions. However, people who endure trauma are more likely to have a smartphone at hand than a computer, e.g. when people are waiting in an emergency room. Immersion, a deep mental involvement, of a game is affected by screen size (Thompson, Nordin, & Cairns, 2012): smaller screens lead to less immersion.

Immersion is positively related to game performance (Cheng, She, & Annetta, 2014). However, it is unclear whether game performance and immersion could lead to less competition with the trauma film memory. The current study investigates whether the use of smartphones after viewing a trauma film is also effective to reduce intrusions. Furthermore, to provide a further understanding of the modality-specificity hypothesis, the present study researches whether the visuospatial aspects of a task are necessary to reduce intrusions by using different games.

In the current study it is hypothesised that compared to an active control word puzzle game (WordBrain) played on a smartphone and a no-task control group, participants who play a visual-based computer game (Candy Crush) on a smartphone soon after viewing a traumatic film will experience fewer image-based intrusive memories of the film over the subsequent week. In

(9)

addition, one week post-film participants in the Candy Crush group will report lower scores on the intrusion provocation task (a laboratory-based intrusion monitoring task) and the intrusion subscale of the Impact of Event Scale – Revised. Both the active control group and the no-task control group will yield similar results on the outcome measures, thereby providing further evidence for the modality-specificity hypothesis. A similar effect size is expected between the experimental condition and the control conditions as has previously been found in studies using a computer instead of a smartphone.

Methods

Participants

Fifty participants (29 females, 21 males) were recruited from the volunteer panel of the Medical Research Council (MRC) Cognition and Brain Sciences Unit. Participants’ age ranged from 18 to 64 (M = 34.52, SD = 14.45). Forty-four % of participants were students, 38 % were employed, 5

% were unemployed and 8 % were retired. Participants were 70 % white Caucasian, 8 % mixed, 14 % Asian and 8 % Chinese. All participants provided written informed consent after being informed about the nature of the experiment and the possibly distressing film. Ethical approval has been given by the University of Cambridge Council of the school of the Biological Sciences Cambridge Psychology Research Ethics Committee (ethics reference: pre.2013.149).

Trauma film Trauma film

The 20-minute film consisted of 25 different scenes, which have content of the type listed as traumatic in the DSM-5, i.e. actual or threatened death and serious injury (APA, 2013). Eleven clips were used in previous research (Holmes et al., 2009; Holmes et al., 2010; James et al., 2015) and fourteen new clips were added. Thirteen of these new clips were provided by Thompson Reuters (multinational mass media and information firm) and as such were scenes from real media footage, ‘eyewitness media’. The remaining new clip was a scene from the movie Irréversible. A description of the film clips is available in appendix A.

(10)

Film reminder task

The film reminder task existed of 25 stills from the film, one from each of the film clips. These stills were selected from parts of the film that were as neutral as possible. Each still was shown for 3 seconds on a computer screen in a fixed random order, and stills automatically moved on to the next one.

Competing Cognitive Tasks Computer Games for Smartphone

All games were downloaded from the Google Play Store, Apple store or Windows Phone store depending on smartphone of the participant.

Candy Crush. Candy Crush is a game, which was presumed to be a visuospatial-based game.

Participants saw a matrix filled with brightly coloured candies of different shapes and had to swap the candies in the matrix in order to make rows or columns of the same kind of candy. Difficulty increased as the participants completed more levels. Participants were instructed to focus on how a certain move would change the subsequent configuration of the brightly coloured candies on the screen to ensure they used visuospatial imagery. See appendix B for the protocol instructions.

WordBrain. WordBrain is a word puzzle game where participants saw a grid with scrambled letters

and they had to make specific words by swiping over the screen in the order of the letters that spelled out the word. The number of letters in the grid increased as participants completed more levels. Participants were instructed to use verbal strategies to generate words and to ‘focus verbally on the meanings of the words’ in the grid. This was in order for the participants to use verbal strategies

instead of mental imagery (Holmes, Lang, & Shah, 2009). See appendix B for the protocol instructions.

No-task control. Participants in the no-task control group sat quietly for 10 minutes, an equivalent

period of time to the Candy Crush and WordBrain group. During this time period participants were not allowed to talk or look at their smartphone, but were told they could think about anything with no restrictions.

Baseline game play

Participants were asked how often they played video games in general, ranging from 0 (never) to 8 (every day). They were also asked whether they played Candy Crush and WordBrain before and in case they had, how often, ranging from 0 (never) to 8 (every day). Also participants were

(11)

asked the approximate duration every time they played the game, ranging from 0 (never) to 8 (more than six hours a day). They were also asked on what device they usually played the games, ranging from 1 (computer) to 2 (smartphone) and 3 (other). Furthermore, they were asked to rate their skill for these games from 0 (not at all) to 5 (extremely skilled). Also, participants were asked what smartphone they had and what kind of mobile operating system. With this information the screen size of their smartphone could be determined.

Smartphone Game Ratings

Game play scores. Game scores for Candy Crush and WordBrain were obtained and also the

number of levels completed, failed and retried.

Game questions. After game play (both Candy Crush and WordBrain) or no-task control (doing

nothing for ten min), the participants indicated whether they followed the instructions and how much mental imagery and verbal strategies they used during that time. The participants who played either of the games also rated their enjoyment of the game, the attention they paid to the game, the difficulty of the game and whether the game distracted them. All ratings used visual analogue scales (VAS) ranging from 0 (not at all) to 10 (extremely).

Game immersion. Participants that played a game during experimental allocation also filled in the

Immersive Experience Questionnaire (IEQ). The IEQ is a 31-item questionnaire on a 5-point Likert scale, which measures game immersion (Jennett, Cox, Cairns, Dhoparee, Epps, Tijs, & Walton, 2008). An example question of the IEQ is ‘To what extent did you notice events taking place around you?’ As well as a total immersion score which ranges from 31 to 155, the questionnaire

also calculates five subscale immersion factors: cognitive involvement (range 9-45), emotional involvement (range 6-30), real world dissociation (range 7-35), control (range 5-25) and challenge (range 4-20). Higher scores reflect a higher level of immersion. The questionnaire also composes a single question ‘How immersed did you feel?’ as a convergent rating, which should be in accordance

with the IEQ total score. This question was rated on a scale from 1 to 10, again higher scores reflecting a higher level of immersion. The construct validity of the IEQ is good (Angelides & Agius, 2014).

Measures

Baseline measures

(12)

Beck Depression Inventory-II (BDI-II) is a 21-item self-report questionnaire that assesses

depressive symptoms and their severity (Dozois, Dobson, & Ahnberg, 1998). An example is: ‘Sadness: 0) I do not feel sad, 1) I feel sad much of the time, 2) I am sad all of the time, 3) I am so sad or

unhappy that I can't stand it.’ Every item is scored on a scale of 0 to 3, with a range from 0 to 63. A

higher score represents more depressive symptomatology. The BDI-II has a high internal consistency of α = .93 among students and α = .92 among outpatients, a good validity and has diagnostic discrimination (Beck et al., 1996, cited in Dozois et al., 1998).

Spielberger State-Trait Anxiety Inventory (STAI-T) was used to assess trait anxiety

(Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), with a 20-item self-report measure. An example is ‘I feel nervous and restless’. Every item is scored on a scale of 1 (almost never) to 4

(almost always), with scores ranging from 20 to 80. Higher scores indicate higher levels of trait anxiety. The internal consistency is satisfactory (α = .89) and also the test-retest reliability is acceptable (ρ = .88) (Barnes, Harp, & Jung, 2002).

Traumatic Experience Questionnaire (TEQ) was adapted from the Criterion A list of the

Posttraumatic Diagnostic Scale (PDS; Foa, 1995), and was used to assess trauma history. Participants indicated whether they have seen or experienced 12 different forms of traumatic events. An example is ‘Have you experienced or witnessed a serious accident, fire, or explosion? (E.g. an industrial, farm, car, plane or boating accident.)’ ‘Yes’ answers were summed and ranged from 0 (no

traumatic events) to 12 (all types of traumatic events witnessed or experienced). The PDS has high internal consistency (α = .92), test-retest reliability (kappa = .74, 87% agreement) and a high diagnostic agreement with the Structured Clinical Interview (kappa = .65, 82% agreement) (Foa, Cashman, Jaycox, & Perry, 1997).

Spontaneous Use of Imagery Scale (SUIS) is a 12-item questionnaire that measures trait tendency

to use spontaneous imagery in daily life (Reisberg, Pearson, & Kosslyn, 2003). Every item is scored on a scale of 1 (never appropriate) to 5 (completely appropriate), with scores ranging from 12 to 60. Higher scores indicate a higher level of spontaneous imagery in daily life. An example of the SUIS is ‘If I am looking for new furniture in a store, I always visualize what the furniture would look like in particular places in my home’. Internal consistency is very good (α =.98; Reisberg, Pearson, &

Kosslyn, 2003). Also the reliability and convergent validity of the SUIS is acceptable (Nelis, Holmes, Griffith, & Raes, 2014).

Pre – Post film mood ratings

Six VAS were used to rate participants’ mood pre- and post-film for sadness, hopelessness, depressiveness, fearfulness, horror and anxiousness on scales from 0 (not at all) to 10 (extremely)

(13)

(James et al., 2015). A composite mood score was calculated by summing participants’ ratings on the six scales. Higher scores reflect worse mood (James et al., 2015).

Baseline intrusions

After watching the film, participants were asked to sit quietly for 5 minutes with their eyes closed. Participants were told they could think about anything they wanted. During this time participants were instructed to make a mark on a piece of paper every time an image of the film popped into their minds spontaneously.

Intrusive Memory Measures Intrusion diary

Participants received a pen-and-paper diary to record intrusive memories for a week (Holmes et al., 2009). Intrusive memories were defined as scenes of the film that appeared spontaneously in their mind. Participants were instructed not to include memories that were deliberately recalled. Furthermore, a distinction between mental images and verbal thoughts was made. Participants were asked to make a mark in the diary on the corresponding day and time period (morning, afternoon or evening) immediately after they experienced an intrusion. Also, participants were instructed to write down the content of their intrusions to ensure that the intrusions matched one of the film clips. If participants experienced no intrusions during a certain time period, they had to write that down in the diary as well. Only the intrusive mental images were scored and the total number of image-based intrusions was calculated.

Diary compliance. Participants were asked how accurate they think the diary they completed

is on a scale from 1 (not at all accurate) to 10 (extremely accurate). Intrusion Provocation Task (IPT)

Twenty-five blurred static visual images were presented, one image from every scene of the trauma film. After viewing the images, participants were instructed to sit quietly for 2 minutes with their eyes closed. Again, participants were told they could think about anything they wanted. Participants were instructed to make a mark on a piece of paper for every image-based intrusive memory triggered by the images. Total frequency is the IPT-intrusion score.

The Impact of Event Scale – Revised (IES-R)

(14)

Participants indicated for 22 items how distressing each item had been during the past week regarding the traumatic film they had watched. Every item is scored from 0 (not at all) to 4 (extremely). The intrusion subscale exists of 8 questions. The score on this scale was the mean of the 8 questions, thus ranging from 0 (not at all) to 4 (extremely). Higher scores indicate a higher level of re-experiencing symptoms. The IES-R has high internal consistency (α = .96) (Creamer, Bell, & Failla, 2003).

Manipulation checks Film ratings

Distress. After watching the film, participants rated how distressing they found the movie on a

scale from 0 (not at all) to 10 (extremely).

Attention. Participants rated how much attention they paid to the movie on a scale from 0 (not at

all) to 10 (total attention). Past-week game play

Participants were asked whether they played Candy Crush or WordBrain during the week between the sessions. If they played either of the games they were asked on how many days they played the game and how much time they spend playing the game(s) each time, again ranging from 0 (never) to 8 (more than six hours a day).

Memory tests

Verbal recognition memory test. Sixty-two true/false statements were provided and participants had to

indicate whether each statement was true or false. Scores were calculated by summing correct responses.

Visual recognition memory test. In total 50 still images were displayed one at a time for 5 seconds

each before automatically moving on to the next image. Twenty-five still images were images of the trauma film and the other 25 images were foil images, one for each of the scenes in the film. Foils were either from the original full film clip but edited out of the shown trauma film or they were visually similar. The stills were presented in a fixed random order. On a sheet provided participants indicated whether they recognized the stills as being from the trauma film viewed the week before. Scores were calculated by summing correct responses.

(15)

Demand ratings

The Perceived Awareness of the Research Hypotheses Scale (PARH) investigates whether participants are aware of the goal of the experiment (Rubin, Paolini, & Crisp, 2010). Participants indicated for all four items whether they agreed with each statement on a scale of 1 (strongly disagree) to 7 (strongly agree). The mean on these items was taken as the score. Scores higher than 4 indicate that participants are aware of the goal of the experiment. The scale has a satisfactory internal consistency (α = .77 and .81).

Procedure

A between groups, two sessions design with three groups was used. In the first session

participants completed the baseline measures, see Figure 2. Then all participants, including the participants in the no-task control group, practiced both games for 1.5 minutes to ensure that they would be immediately engaged when playing the game after watching the film. After rating their mood pre-film, they saw the trauma film on a computer. Participants were told to immerse themselves, pay close attention to the film and imagine they were at the scene as a bystander. Participants put on headphones and the lights were turned off as the experimenter exited the room for the duration of the film. After watching the film, participants rated their mood again and their attention paid to the movie and their distress due to the movie. Then the baseline intrusions were measured. Directly after participants viewed the film reminder task and were then

Figure 2. Design of the study. IES-R: Intrusion subscale of the Impact of Event Scale – Revised, IPT: Intrusion Provocation Task.

(16)

allocated to play either Candy Crush, WordBrain or no-task for ten minutes. Participants then answered the game questions and the IEQ (no-task group filled out only the first game questions). Then, participants were instructed how to keep the diary over the week between sessions. Diary instructions were given verbally and these were also written down in the diary. One week later, participants returned to the lab with their diary and performed two recognition memory tests, filled out the IES-R, assessed their diary accuracy and answered questions about game play in the week between sessions. Also, participants performed the IPT, gave demand ratings and their diary entries were checked. Lastly, participants were debriefed, thanked and paid.

Results

Data Analysis

One participant did not complete the study and as such was not included in the data analysis. A total of 50 participants were included in the analysis. Assumptions of independence of observations were met for all participants.

All data were examined for potential outlying values, ≥ 3 SD of the mean in either direction. All outliers were examined for their influence on the results. If the outlier did not influence the results, no adjustment was made. Outliers are reported in Appendix C. Furthermore, the assumptions of normality were checked using Kolmogorov-Smirnov test and by

looking at histograms of the data. Homogeneity of variances was checked for all data using

Levene’s Test of Homogeneity of Variances. When the assumptions were met, means between

conditions were compared with one-way ANOVA’s. Significant results were followed-up with post-hoc pairwise comparisons corrected with Bonferroni. However, when assumptions of normality and/or homogeneity of variances were violated, the Kruskal-Wallis Test was used. The

test statistics of violated assumptions can be found in Appendix C.

Repeated-measures ANOVA was used to compare mood from pre- to post-film and to

compare intrusion frequency per day across groups. The p-values were corrected with

Greenhouse-Geisser when sphericity was violated. Nominal data were analysed using the chi-square test. Relationships between variables were analysed using a Pearson’s product-moment correlation. All post-hoc pairwise comparisons were corrected with Bonferroni. An alpha level of .05 was used for all statistical tests. IBM SPSS statistical package version 22 was used for all analysis.

(17)

Measures

Baseline measures

There were no significant differences at baseline for the demographic characteristics between groups in age, gender, employment and ethnicity, see Table 1. There were also no significant differences between groups in depression, trait anxiety, traumatic experiences and spontaneous use of imagery at baseline.

Table 1. Baseline Characteristics for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16) n (%) n (%) n (%) Female 7 (41.18) 11 (64.71) 11 (68.75) M (SD) M (SD) M (SD) Age 33.24 (13.61) 31.65 (12.67) 38.93 (16.73) Employment 2.35 (1.11) 2.29 (1.05) 2.00 (1.03) Ethnicity 1.65 (0.93) 1.88 (1.54) 1.50 (1.15) BDI-II 5.82 (6.56) 5.53 (5.79) 5.50 (7.98) STAI-T 35.47 (8.80) 35.06 (9.63) 36.50 (11.33) TEQ 1.17 (1.07) 1.11 (1.22) 1.13 (1.67) SUIS 37.71 (8.64) 39.35 (9.18) 39.00 (8.12) Attention to film 9.41 (0.71) 9.29 (0.77) 9.44 (0.73) Post-film distress 7.88 (1.17) 7.94 (0.68) 7.81 (1.22)

Mood rating pre-film 4.15 (4.43) 5.32 (9.49) 5.53 (7.42)

Mood rating post-film 25.67 (12.76) 25.57 (16.25) 23.88 (14.19)

Baseline intrusions 12.94 (9.36) 13.00 (11.35) 11.50 (5.90)

Note_._{BDI-II: Beck Depression Inventory-II, STAI-T: State-Trait Anxiety Inventory – Trait, TEQ: Traumatic}

Experience Questionnaire, SUIS: Spontaneous Use of Imagery Scale.

Pre – Post-film mood ratings

There was a significant effect of time on mood, Pillai’s Trace F (1, 47) = 116.83, p < .01, η2₌ .71, meaning mood significantly deteriorated from pre-film to post-film, see Table 1. This decline in mood did not differ between groups, as the interaction between mood and group was non-significant.

(18)

Baseline intrusions

The number of baseline intrusions was similar in all three groups, see Table 1.

Competing Cognitive Tasks

Baseline game play

There were no differences at baseline between groups on game play in general and on previous Candy Crush game play and WordBrain game play, see Table 2. Also, no differences were found in skill of the participants on both games between the three groups.

Table 2. Baseline Game Play Scores for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16) M (SD) M (SD) M (SD)

Frequency video games 2.76 (2.77) 2.41 (2.60) 1.94 (2.38)

Frequency video games on smartphone

2.41 (2.24) 1.82 (2.10) 1.00 (1.32)

Time playing video games 1.71 (1.61) 1.65 (1.32) 1.06 (1.00)

Played CC? * 10 (yes) 5 (yes) 6 (yes)

Frequency CC 1.71 (2.66) 0.65 (1.93) 1.38 (2.55)

Time playing CC .71 (.85) .53 (.87) .38 (.50)

Device CC * 8 (smartphone) 5 (smartphone) 4 (smartphone)

Skill CC .71 (.99) .42 (.80) .56 (.96)

Played WB? * 6 (yes) 6 (yes) 5 (yes)

Frequency WB .65 (1.37) .76 (1.39) .63 (1.31)

Time playing WB .35 (.61) .76 (1.09) .31 (.48)

Device WB * 6 (smartphone) 6 (smartphone) 5 (smartphone)

Skill WB .36 (.70) .59 (.87) .19 (.54)

Note._{CC: Candy Crush, WB: WordBrain. *: Frequency of participants that gave the answer between (brackets).}

Smartphone Game Ratings

Game questions. The amount of following the instructions during experimental allocation was

significantly affected by group, H (2) = 6.12, p = .047, see Table 3. However, pairwise

(19)

were no differences between groups on how well participants followed the instructions during experimental allocation. There were no significant differences between groups on use of mental imagery during experimental allocation. The amount of verbal strategies used during

experimental allocation was significantly affected by group, H (2) = 14.31, p < .01. Pairwise

comparisons with adjusted p-values showed a significant difference in use of verbal strategies

between the Candy Crush group and the WordBrain group, p < .01, meaning that participants in

the WordBrain group used more verbal strategies than participants in the Candy Crush group. There were no differences between participants in Candy Crush and WordBrain group on game enjoyment, game attention, game difficulty and game distraction.

Game immersion. There were no differences between participants in Candy Crush and WordBrain

group regarding game immersion (total score, single score and the five subscales), see Table 3. Table 3. Game Questions for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16) M (SD) M (SD) M (SD) Followed instructions 8.76 (.97) 9.49 (.65) 9.42 (.78) Mental imagery 7.24 (1.96) 7.50 (3.09) 7.65 (1.82) Verbal strategies 3.30 (2.66) 7.59 (2.34) 4.60 (3.27) Game enjoyment 5.95 (1.83) 5.41 (2.98) - Game attention 7.60 (1.61) 8.38 (1.75) - Game difficulty 3.85 (2.70) 4.83 (2.92) - Game distraction 6.82 (2.85) 6.27 (3.48) - IEQ total 101.76 (14.53) 97.82 (10.85) - IEQ challenge 12.18 (2.00) 12.24 (2.17) - IEQ control 16.35 (2.67) 15.53 (3.04) -

IEQ real world dissociation 23.94 (4.44) 21.00 (4.21) -

IEQ emotional involvement 15.06 (4.18) 14.47 (4.93) -

IEQ cognitive involvement 34.24 (5.89) 34.59 (5.50) -

IEQ single immersion 7.00 (1.62) 6.53 (1.59) -

(20)

Intrusive Memory Measures

The assumption of linear relations between each pair of the dependent variables within each condition was not met. The correlations between the dependent variables were insignificant, except the correlation between the IPT and diary frequency in the no-task group, see Table 4. The scatterplots also did not show a clear pattern. Therefore, three separate one-way ANOVA’s were conducted on the three dependent variables rather than a MANOVA.

Table 4. Correlations between Outcome Measures for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16)

IPT IES-R IPT IES-R IPT IES-R

Diary .27 0.10 .11 .43 .59* .30

IPT - .16 - -.25 - .35

Note._{*: Significant at p < .05.} Intrusion diary

Diary compliance. The Kruskal-Wallis test showed that there was no difference in diary compliance

between the three groups, see Table 8.

Table 5. Intrusive Memory Measures for Each Experimental Condition

Intrusion diary frequency 8.18 (5.60) 4.47 (4.11) 9.00 (6.39)

IPT frequency 5.59 (3.20) 5.24 (3.09) 6.00 (4.03)

IES-R intrusion score .90 (.40) .96 (.66) 1.05 (.62)

Note._{IPT: Intrusion Provocation Task, IES-R intrusions: intrusion subscale of Impact of Event Scale-Revised.}

Intrusion diary. A one-way ANOVA was significant, F (2, 49) = 3.30, p = .045, η2_{= .12, see Table} 5. However, none of the comparisons with adjusted p-values were significant at follow-up. Candy

Crush did not differ significantly from WordBrain, p = .16, and not from no-task group, p = 1.00.

(21)

Exploratory analyses. The course of the frequency of intrusions over the week was studied, see

Table 6 and Figure 4. With a repeated-measures ANOVA it was shown that there was a significant effect of day, F (3.69, 173.61) = 22.33, p < .01, η2_{= .32, meaning there was a decline} in the number of intrusions over the days. There was no significant interaction between group and day, indicating that the decline of intrusions over the week was similar for all groups.

Table 6. Frequency of Intrusions in Diary per Day for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16) M (SD) M (SD) M (SD) Day 1 2.53 (2.12) 1.82 (1.81) 3.31 (2.70) Day 2 1.59 (1.28) 1.12 (1.54) 2.06 (2.2) Day 3 .88 (1.36) .35 (.61) 1.00 (1.03) Day 4 .65 (1.00) .24 (.56) .63 (1.15) Day 5 .53 (1.18) .29 (.59) .63 (1.09) Day 6 1.12 (1.76) .29 (.59) .44 (1.09) Day 7 .65 (.86) .18 (.39) .31 (.60) Day 8 .24 (.56) .18 (.39) .32 (.79)

(22)

The number of baseline intrusive images, measured immediately after film viewing in the laboratory, was significantly correlated to intrusive images in the diary, r = .33, p = .02. This

relationship turned out to be significant in the no-task group, r = .54, p = .03, but not in the

other two groups (Candy Crush, r = .34, p = .18; WordBrain, r = .39, p = .13). Therefore, the

number of baseline intrusions was included in the analysis on diary frequency as a covariate and was significantly related to the frequency of intrusions as reported in the diary, F (1, 46) = 7.58, p

< .01, η2_{= .14. There was also a significant effect of condition on the number of intrusions} reported in the diary after controlling for the effect of baseline intrusions in the laboratory, F (2,

46) = 4.05, p = .02, partial η2_{= .15. Pairwise comparisons with Bonferroni correction showed a} significant difference between the WordBrain group and the no-task group on frequency of intrusions in the diary when baseline intrusions were controlled for, p = .03. This suggests that

participants in the WordBrain group had fewer intrusions as measured by the diary than participants in the no-task group when baseline intrusions were taken into account, see Figure 5.

Figure 4. Number of intrusions as measured by the diary per day for each experimental condition.

Figure 5. Number of intrusions as measured by the diary over the week with number of baseline intrusions taken into account. *: Significance of pairwise comparisons.

(23)

Post-film distress, diary compliance, game scores, game questions and game immersion scores were not related to the intrusion diary frequency overall and also not related in each of the groups separately. In addition, screen size was unrelated to diary frequency and to game immersion.

Intrusion Provocation Task

There were no differences between the groups in the frequency of intrusions reported by participants on the intrusion provocation task given in the laboratory one week post-film viewing, F (2, 49) = .20, p = .82, η2_{= .00, see Table 5.}

Intrusion subscale of the Impact of Events Scale – Revised

There were no differences between groups on scores on the intrusion subscale of the IES-R, F

(2, 49) = .32, p = .73, η2_{= .01, see Table 5.} Manipulation checks

Film ratings

Distress. There were no significant differences between groups on post-film distress, see Table 1. Attention. There were no significant differences on attention paid to the film between the three

groups.

Past-week game play

There were no differences in the amount of game play of Candy Crush and WordBrain between the three groups in the time between session 1 and session 2, see Table 7.

Table 7. Past Week Game Play for Each Experimental Condition

Candy Crush (n = 17) WordBrain (n = 17) No-task control (n = 16)

Frequency yes Frequency yes Frequency yes

Played CC? 3 4 5

Played WB? 4 8 6

(24)

Memory tests

Verbal recognition memory test. There were no significant differences between the three groups on

the verbal recognition memory test, see Table 8.

Visual recognition memory test. There were no significant differences between groups on the visual

recognition memory test. Demand ratings

A one-way ANOVA was conducted and was non-significant, meaning there were no differences between groups on awareness of the hypothesis, see Table 8. Furthermore, for every group the mean score was compared with the mean of this scale. In the Candy Crush group the score was significantly lower than 4, t (16) = -2.67, p = .02, suggesting that this group was unaware of the

hypotheses of the experiment. In the WordBrain group the score was not significantly different from 4, meaning that this group stated they were not aware of the hypotheses of the experiment. In the no-task group the score was significantly lower than 4, t (15) = -2.87, p = .01, meaning that

this group stated they were unaware of the hypotheses of the experiment. Table 8. Manipulation Checks for Each Experimental Condition

Verbal recognition memory 39.88 (4.53) 41.88 (4.86) 39.88 (3.67)

Visual recognition memory 35.59 (3.94) 35.53 (3.95) 34.19 (4.58)

Demand PARH 3.16 (1.29) 3.42 (1.40) 3.03 (1.35)

Diary compliance 8.88 (0.86) 8.21 (1.69) 8.13 (1.20)

Note._{PARH: Perceived Awareness of Research Hypotheses}

Discussion

In this experiment the effects of game play on a smartphone after viewing a trauma film on subsequent image-based intrusive memories was studied. Contrary to the prediction, the results showed that participants who played Candy Crush did not have fewer intrusive memories compared to participants in the WordBrain (the active control game) group and the no-task control group, as measured by all three main outcome measures. Participants who played WordBrain did not have fewer intrusive memories than participants in the no-task control group,

(25)

as measured by all three outcome measures. However, when the number of intrusions reported directly after watching the trauma film was controlled for, the participants who played WordBrain did experience fewer intrusions reported on the diary than participants in the no-task control group. Even though the number of baseline intrusions was equal in all three groups, the relationship between the number of baseline intrusions and intrusions in the diary over the week differed between the groups. In the control group there was a positive relationship between the number of baseline intrusions and the number of intrusions in the subsequent week measured by the diary, while this relationship was not visible in the other two groups. The relationship between baseline intrusions and intrusions in the subsequent week might have been altered by game play. Importantly, deliberate recall of the trauma film did not differ between groups, indicating that game play did not interfere with deliberate recall, but only with involuntary intrusions of the trauma film memory.

Results are unlikely to be explained by differences between groups prior to allocation to experimental group. Participants in all groups were similar in demographic characteristics and on baseline measures of depression, anxiety, previous trauma, daily use of imagery, change in mood due to trauma film and baseline intrusions. Also, participants in all groups were equally distressed by the trauma film and paid the same amount of attention to the trauma film. One explanation for the findings could be that Candy Crush did not adequately recruit visuospatial properties and therefore was not taxing the VSSP sufficiently to reduce intrusions. Participants were instructed to use mental imagery, but it could be that the game was played without using mental imagery and rotation. Participants started at level one and difficulty increased as the game progressed. This was done to ensure that participants played at their own ability, which is thought to be important because a task that is either too easy or too difficult may not tax working memory sufficiently to exert dual task competing effects (Van den Hout & Engelhard, 2012). However, the first levels of the game might have been too easy, especially for participants who were already familiar with the game. Playing the game for a longer duration may mitigate this in future studies. Furthermore, it took time to load the games between levels. Thus, the game might not have been as engaging and taxing the VSSP as other games, such as Tetris, that has shown to reduce intrusive memories when played a comparable time to Candy Crush in this experiment (Holmes et al., 2009; Holmes et al., 2010). Furthermore, the word puzzle game WordBrain that was used as an active control might have been more visuospatial-based than anticipated. Even though participants rated that they used more verbal strategies than participants in the Candy Crush group, they also rated that they used a comparable amount of mental imagery. Participants were required to think how the letters in the grid could form a word, which may have led them to use mental imagery strategies.

(26)

Also, WordBrain is different to the verbal Pub quiz general knowledge game used in previous studies, where participants had to answer questions and use semantic knowledge (Holmes et al., 2010). Therefore, the word puzzle game might not interfere with conceptual processing, but with sensory perceptual processing instead (Ehlers & Clark, 2000; Brewin et al., 2010). Thus, it might be that WordBrain taxed the VSSP whereas Candy Crush taxed the VSSP to a lesser extent, leading WordBrain to interfere with sensory perceptual processing of the trauma film, while Candy Crush did not. Then the results would still be in line with the modality-specificity hypothesis. It was not in the scope of the current study to determine the modality of the games.

Another interpretation of the results is that the modality-specificity hypothesis is incorrect. Previous research showed that it may be not the modality of a task that is necessary to reduce intrusions, but that only a high executive load is important (Pearson & Sawyer, 2011). However, other research, as mentioned earlier, does not support this hypothesis, because a visuospatial task reduced intrusions, while a verbal task did not (Holmes et al., 2010). Furthermore, articulation reduced vividness and emotional intensity of auditory images more than eye movements, while it was the other way around for visual images (Kemps & Tiggemann, 2007). Also, if a secondary task is taxing working memory too much, it may not have an effect at all, since working memory is taking up all resources for the secondary task (Van den Hout & Engelhard, 2012). Together, this suggests that modality-specificity is an important consideration for tasks that are performed following trauma to reduce intrusions. However, in these studies it is unclear whether the tasks have the same amount of workload. These secondary tasks, including eye movements in the treatment of Eye Movement and Desensitization and Reprocessing (EMDR), could also just tax the central executive of WM and not necessarily the modality-specific resources (Gunter & Bodner, 2008). Therefore, it could also be argued that these tasks only differ in the amount they tax WM and that modality is not important (Pearson & Saywer, 2011). This suggests that perhaps not only modality, but also workload of tasks may be important to reduce intrusions. Future studies should try to disentangle the influence of workload and modality, by researching the workload of the tasks used and their modality.

In future studies the modality of both games should be investigated prior to investigating if the games can reduce intrusion frequency. Previous research with Functional Magnetic Resonance Imaging (fMRI) showed that the Blood Oxygen Level Dependent (BOLD) signal increased in the occipital-parietal cortex while playing Tetris (Price, Paul, Schneider, & Siegle, 2012), supporting the visuospatial properties of this task. However, such studies do not exist for the relatively new games used in this study. Further knowledge on this topic would provide understanding of the modality-specificity hypothesis. In addition, from this study alone it cannot

(27)

be concluded how much working memory is taxed during game play for WordBrain and Candy Crush. This should also be investigated in future research. The reaction time task could be used to infer how much working memory is taxed. To know whether the same amount of working memory is taxed by the different games, participants must perform a reaction time task A and then play either of the games concurrently. The increase in reaction time on task A due to the concurrent game played gives an indication of the amount of working memory needed by the concurrent task (Van den Hout & Engelhard, 2012). This would provide information as to whether the games are equally taxing working memory and whether this played a role in the findings from the current study.

On two of the three main outcome measures in this study no differences were found between groups, with effect sizes around zero. Previous studies that did show significant differences on intrusion frequency measured by the diary have shown inconsistent results regarding differences on the IES-R and the IPT, the two outcome measures in this study that showed no difference between groups (Holmes et al., 2009; Holmes et al., 2010; James et al., 2015). This may indicate that these two measures, a self-report questionnaire and a laboratory intrusion provocation task, might not be as sensitive to measure intrusions as a result of analogue trauma as a real-life measure of intrusions, namely the diary. This might explain the fact that no differences were found between groups on these two measures in this experiment.

This study was the first study to use smartphones for game play instead of desktop computers. Previously it has been shown that screen size affects immersion of a game and that immersion is related to game performance (Thompson et al., 2012; Cheng, She, & Annetta, 2014). In this study, immersion levels of games played on a smartphone compared to games played on a computer was not tested. Study results do suggest that smartphones can also be used to reduce image-based intrusive memories. Namely, screen size of the smartphone was not related to any of the outcome measures and screen size was also unrelated to immersion. Game performance was also unrelated to any of the outcome measures. Furthermore, WordBrain reduced the number of intrusions compared to the no-task group and this effect size (medium to large) was similar to effect sizes found in previous studies using computers for game play (Cohen’s d = .72 and Cohen’s d = .84 when baseline intrusions were controlled for, compared to Cohen’s d =

.91 in Holmes et al., 2009 and Cohen’s d = .70 in Holmes et al., 2010). This suggests that

smartphones can be used as a method to deliver games as a preventative intervention after trauma.

The trauma film used in this study differed from the trauma film used in previous studies (Holmes et al., 2009; Holmes et al., 2010; James et al., 2015). This film had 14 new clips added

(28)

and had a 9-minute longer duration. The number of intrusions in the no-task group is higher than in previous studies, likely due to the addition of new film clips. These clips were mostly footage from the media. Journalists watch distressing footage as part of their jobs and watching this footage is traumatic for many journalists (Dubberley et al., 2015). The footage used in the trauma film was not as distressing as the clips journalist view on a regular basis, however the trauma film used in this study is comparable to the footage they have to view. As such, the trauma film in this study may be especially well suited for studying the build-up of intrusions of the kind experienced in the work of journalists. In this study it was shown that playing WordBrain on a smartphone reduced the build-up of image-based intrusive memories over the week, when the amount of intrusions directly after trauma film viewing was controlled for. In the first few days after watching the trauma film participants experienced less intrusive memories when they had played WordBrain after the trauma film than the participants in the other two groups, while after a week there were no differences between groups anymore. Even though participants without the intervention had a higher number of intrusions at the beginning of the week, all participants experienced nearly zero intrusions at the end of the week. Due to a floor effect of intrusions after analogue trauma in this study, it is unclear whether the intervention has longer-lasting effects to reduce the number of image-based intrusions even after the first few days. Therefore, it is unclear whether this preventative intervention has lasting effects for individuals over longer periods of time. It is hard to interpret these results, because in this study the trauma film was used as an experimental analogue to study real-life trauma. A direct translation from these results to a clinical population is therefore not warranted. Individuals who suffered trauma and develop PTSD tend to experience intrusions over longer periods of time (APA, 2013) and perhaps in these individuals a comparable preventative intervention would be able to reduce the number of intrusions compared to no intervention over a longer time period. Currently, there are no studies researching the effect of game play as a preventative intervention after trauma in a clinical sample. Future studies should provide more insights as to whether the results from these experimental studies can be translated to the clinical population and whether the intervention has possibly longer-lasting effects in a clinical sample.

Even though the number of image-based intrusions are equal in all groups after a week, playing WordBrain after the trauma film did reduce the number of intrusions over the week, which is due to a difference in the first few days after trauma film viewing. Therefore, playing WordBrain might be able to alleviate symptoms in the first few days after trauma. As game play on a smartphone could be easily applied as an early intervention after trauma, it might be worthwhile in the future to ensure that journalists play this game on a smartphone after watching

(29)

distressing footage to reduce the build-up of intrusions directly after watching the footage. If a similar preventative intervention shows to reduce intrusions after real-life trauma, a simple smartphone intervention could also be used in other workplaces such as emergency rooms, where trauma is routinely experienced.

This study has several limitations. This study may have benefitted from a condition with the visuospatial game (Tetris) used in previous studies, so results could be interpreted further. Furthermore, lack of statistically significant effects might have been due to inadequate power. Based on previous studies a total of 66 participants should have been recruited instead of 50. Another limitation was that there was no direct comparison of smartphone game play with game play on a computer. A direct comparison between game play on a smartphone and on a computer can provide stronger evidence as to whether any differences or benefits exist between devices.

In the first period after trauma individuals tend to experience intrusions, even when they do not develop full-blown PTSD (Ehlers, 2015). Intrusions might therefore not be necessarily indicative of psychopathology. However, individuals who were diagnosed with PTSD a year after trauma, experienced more intrusions eight days after trauma than participants who were not diagnosed with PTSD after a year, indicating that intrusion frequency might predict the development of PTSD (O’Donnell et al., 2007). Another study showed that intrusion frequency can only account for 9% of the severity of PTSD symptoms, while the nowness, associated distress and lack of context of the intrusions were able to explain an additional 43% of the variance (Michael, Ehlers, Halligan, & Clark, 2005). Intrusion frequency is only moderately related to the distress caused by these intrusions and weakly to the severity of PTSD symptoms (Steil & Ehlers, 2000). Moreover, those intrusions that are experienced as distressing are particularly related to PTSD symptoms. Whether intrusions are experienced as distressing seems to depend on the meaning given to an intrusion. When individuals gave intrusions a dysfunctional meaning, e.g. ‘If after all of this time I react to memories of the event in such a strong way there must be something wrong with me!’ (Steil & Ehlers, 2000), this was associated with distress and severity of PTSD symptoms. Therefore, in future studies not only the frequency of intrusions should be studied but also the distress and the meaning associated with each intrusion as these aspects of intrusions might be more important in preceding PTSD than frequency of intrusions alone.

The current study investigated game play to reduce the build-up of intrusions soon after trauma. However, not in all situations after trauma it is possible to undertake this intervention directly after trauma. Examples are war, natural disasters or in case of serious medical injuries when surgical procedures are needed first. Playing a visuospatial computer game within the

(30)

reconsolidation window has shown to reduce intrusions (James et al., 2015), indicating that game play on smartphones may also be used over longer time frames to reduce image-based intrusions. Under certain conditions reactivated memories become malleable again and are therefore sensitive for modification when the memory is reconsolidated (Nader, 2003). In case when serious medical injuries need to be taken care of first, game play within the reconsolidation window may still reduce intrusions.

This study investigated the parameters that make game play effective in reducing image-based intrusive memories. Game play on a smartphone after experimental trauma can lead to fewer image-based intrusions and potential other benefits. Intrusions are a hallmark symptom of post-traumatic stress disorder and a precursor of PTSD (O’Donnell et al., 2007). Investigating methods such as game play on smartphones might aid in the prevention of the development of intrusions and PTSD and have implications for individuals and society as a whole.

(31)

References

Angelides, M.C., & Agius, H. (2015). Handbook of digital games. IEEE/John Wiley and Sons.

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (DSM- 5®). American Psychiatric Pub.

Arntz, A., de Groot, C., & Kindt, M. (2005). Emotional memory is perceptual. Journal of Behavior Therapy and Experimental Psychiatry, 36(1), 19-34. doi:10.1016/j.jbtep.2004.11.003

Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4(10), 829-839. doi:10.1038/nrn1201

Baddeley, A. (2012). Working memory: Theories, models, and controversies. Annual Review of Psychology, 63, 1-29. doi:10.1146/annurev-psych-120710-100422

Barnes, L. L., Harp, D., & Jung, W. S. (2002). Reliability generalization of scores on the

Spielberger State-Trait Anxiety Inventory. Educational and Psychological Measurement, 62(4),

603-618. doi:10.1177/0013164402062004005

Brewin, C. R., Gregory, J. D., Lipton, M., & Burgess, N. (2010). Intrusive images in psychological disorders: characteristics, neural mechanisms, and treatment implications. Psychological Review, 117(1), 210-232. doi: 10.1037/a0018113

Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). β-Adrenergic activation and memory for emotional events. Nature, 371(6499), 702-704. doi:10.1038/371702a0

Cheng, M. T., She, H. C., & Annetta, L. A. (2015). Game immersion experience: Its hierarchical structure and impact on game-based science learning. Journal of Computer Assisted Learning, 31(3), 232-253. doi: 10.1111/jcal.12066

Creamer, M., Bell, R., & Failla, S. (2003). Psychometric properties of the impact of event scale— revised. Behaviour Research and Therapy, 41(12), 1489-1496. doi:10.1016/j.brat.2003.07.010

Deeprose, C., Zhang, S., Dejong, H, Dalgleish, T., & Holmes, E. A. (2012). Imagery in the aftermath of viewing a traumatic film: Using cognitive tasks to modulate the development of involuntary memory. Journal of Behavior Therapy and Experimental Psychiatry, 43(2), 758-764.

doi:10.1016/j.jbtep.2011.10.008

Dozois, D. J., Dobson, K. S., & Ahnberg, J. L. (1998). A psychometric evaluation of the Beck Depression Inventory–II. Psychological Assessment, 10(2), 83-89.

Dubberley, S., Griffin, E., & Bal, H. M. (2015). Making secondary trauma a primary issue: A study of eyewitness media and vicarious trauma on the digital front line. Istanbul: Eyewitness Media Hub.

(32)

disorder. Journal of Psychology, 218, 141-145. doi:

http://dx.doi.org/10.1027/0044-3409/a000021

Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder. Behaviour Research and Therapy, 38(4), 319-345. doi:10.1016/S0005-7967(99)00123-0

Ehlers, A., Hackmann, A., Steil, R., Clohessy, S., Wenninger, K., & Winter, H. (2002). The nature of intrusive memories after trauma: The warning signal hypothesis. Behaviour Research and Therapy, 40(9), 995-1002. doi:10.1016/S0005-7967(01)00077-8

Foa, E. B. (1995). Post-traumatic stress diagnostic scale (PDS). Minneapolis: National Computer Systems.

Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The validation of a self-report measure of posttraumatic stress disorder: The Posttraumatic Diagnostic Scale. Psychological assessment, 9(4), 445-451.

Grey, N., & Holmes, E. A. (2008). ‘‘Hotspots’’ in trauma memories in the treatment of posttraumatic stress disorder: A replication. Memory, 16, 788–796.

doi:10.1080/09658210802266446

Gunter, R. W., & Bodner, G. E. (2008). How eye movements affect unpleasant memories: Support for a working-memory account. Behaviour Research and Therapy, 46(8), 913-931.

doi:10.1016/j.brat.2008.04.006

Holmes, E. A., & Bourne, C. (2008). Inducing and modulating intrusive emotional memories: A review of the trauma film paradigm. Acta Psychologica,127(3), 553-566.

doi:10.1016/j.actpsy.2007.11.002

Holmes, E. A., Brewin, C. R., & Hennessy, R. G. (2004). Trauma films, information processing, and intrusive memory development. Journal of Experimental Psychology: General, 133(1), 3-22.

doi:10.1037/0096-3445.133.1.3

Holmes E. A., James, E. L., Coode-Bate, T., & Deeprose, C. (2009). Can playing the computer game "Tetris" reduce the build-up of flashbacks for trauma? A proposal from

cognitive science. PLoS One, 4(1), e4153. doi:10.1371/journal.pone.0004153

Holmes, E. A., James, E. L., Kilford, E. J., & Deeprose, C. (2010). Key steps in developing a cognitive vaccine against traumatic flashbacks: Visuospatial Tetris versus verbal Pub Quiz. PloS one, 5(11), e13706. doi:10.1371/journal.pone.0013706

Holmes, E. A., Lang, T. J., & Shah, D. M. (2009). Developing interpretation bias modification as a" cognitive vaccine" for depressed mood: imagining positive events makes you feel better than thinking about them verbally. Journal of abnormal psychology, 118(1), 76-88.