Conceived and designed the experiments: EAH. Performed the experiments: EAH ELJ EJK CD. Analyzed the data: EAH ELJ EJK CD. Contributed reagents/materials/analysis tools: EAH. Wrote the paper: EAH ELJ EJK CD.
The authors have declared that no competing interests exist.
Flashbacks (intrusive memories of a traumatic event) are the hallmark feature of Post Traumatic Stress Disorder, however preventative interventions are lacking. Tetris may offer a ‘cognitive vaccine’
The trauma film paradigm was used as an experimental analog for flashback development in healthy volunteers. In both experiments, participants viewed traumatic film footage of death and injury before completing one of the following: (1) no-task control condition (2) Tetris or (3) Pub Quiz. Flashbacks were monitored for 1 week.
First, computer games can have differential effects post-trauma, as predicted by a cognitive science formulation of trauma memory. In both Experiments, playing Tetris post-trauma film reduced flashbacks. Pub Quiz did not have this effect, even increasing flashbacks in Experiment 1. Thus not all computer games are beneficial or merely distracting post-trauma - some may be harmful. Second, the beneficial effects of Tetris are retained at 4 hours post-trauma. Clinically, this delivers a feasible time-window to administer a post-trauma “cognitive vaccine”.
The psychological impact of trauma is a major challenge to human health worldwide. The majority of healthy individuals are liable to suffer a traumatic event at some point in their life time
We proposed the development of computerized, low intensity, intervention against PTSD flashbacks for use as a preventative mental health strategy
Human memory differentiates visual and verbal components
Pathological trauma flashbacks consist of sensory, visual images (i.e. vivid visual memories such as the sight of the blood-spattered body of a fellow soldier)
Cognitive science shows that visuospatial cognitive tasks compete for resources with visual images
The biology of memory consolidation suggests a 6 hour time frame post-trauma within which memories are malleable
Thus, visuospatial cognitive tasks given within 6 hours post-trauma will interfere with visual flashback memory consolidation, and reduce later flashbacks, as demonstrated in our previous study
In contrast, verbal tasks post-trauma will not reduce flashbacks as verbal tasks compete with verbal, conceptual processing of the event but not the visual images that make up flashbacks
Further, verbal tasks post-trauma will compete with the type of verbal-conceptual processing necessary to make sense of what has happened and from clinical models may serve to increase (rather then reduce) later trauma flashbacks
With regards to points 1–5 above, our previous paper in PLoS ONE
(2) Would all computer games have the effect of reducing flashbacks via distraction/enjoyment, or might some games even be harmful?
A critical limitation of our 2009 study
Interestingly, a counter argument to a “distraction account” is that according to the theory from cognitive science outlined in points 6–7 above, a comparison task such as a computer game requiring verbal processing should not only be ineffective but may even be harmful in increasing the development of later flashbacks. “Pub Quiz”
(1) Would the beneficial effects of playing Tetris on flashbacks be sustained if administered several hours post-trauma?
A second limitation of our 2009 study was that we only tested the intervention soon after trauma, whereas in the real world people may not have access to an intervention for several hours. Thus in Experiment 2, we considerably extended the time frame from 30 min to 4 hours post trauma to explore the impact of playing the computer games after a significant interval on the development of flashbacks.
In summary, we predicted that:
Experiment 1: playing the computer game Tetris 30 min after a traumatic film would significantly
Experiment 2: effects would be maintained when the games were played 4 hours post-trauma.
Sixty non-clinical participants watched a film of traumatic scenes of injury and death (n = 20 per group). Film viewing was followed by a 30-min structured interval before random assignment to one of three experimental conditions (
Participants completed the trauma film paradigm, a well established experimental analog for PTSD. All participants viewed a traumatic film followed by a 30-min structured break. Participants were then allocated to one of three experimental conditions [Tetris vs. no-task control vs. Pub Quiz] which they completed for 10 min. Afterwards participants in the computer game conditions rated their enjoyment of the game. Flashbacks (involuntary memories) were monitored for 1 week using an intrusion diary. After 1 week, diary compliance was checked and a test of voluntary memory (recognition memory test) for the trauma film was administered.
Measure | No-task (n = 20) | Tetris (n = 20) | Pub Quiz (n = 20) | ANOVA | |||
mean | sem | mean | sem | mean | sem | ||
Age | 29.05 | 2.68 | 25.90 | 2.11 | 26.75 | 2.24 | |
Beck Depression Inventory (BDI) | 5.35 | 0.91 | 4.90 | 1.04 | 5.85 | 1.03 | |
Trait Anxiety (STAI-T) | 37.20 | 1.98 | 39.10 | 1.95 | 36.60 | 2.10 | |
Traumatic Experiences Questionnaire (TEQ) | 1.60 | 0.39 | 1.50 | 0.22 | 1.90 | 0.47 |
No-task (n = 20) | Tetris (n = 20) | Pub Quiz (n = 20) | ANOVA | ||||||
mean | sem | mean | sem | mean | sem | Time | Group | Group* Time | |
Pre-film mood | 2.61 | 0.58 | 2.98 | 0.52 | 2.76 | 0.58 | |||
Post-film mood | 12.60 | 1.14 | 10.93 | 1.40 | 11.25 | 1.43 |
p<0.01.
Following the 30-min structured interval period in which participants completed standardized filler tasks, a brief reminder task for the trauma film (neutral static slides from film clips) was administered to all groups. Participants then either completed the visuospatial condition Tetris
No-task (n = 20) | Tetris (n = 20) | Pub Quiz (n = 20) | ANOVA | ||||
mean | sem | mean | sem | mean | sem | ||
Frequency of initial intrusions | 12.35 | 3.40 | 4.30 | 1.10 | 5.90 | 1.38 |
Tetris relative to no-task control p<0.05; Pub Quiz relative to no-task control = NS.
Measure | No-task (n = 20) | Tetris (n = 20) | Pub Quiz (n = 20) | ANOVA/t-test | |||
mean | sem | mean | sem | mean | sem | ||
Post-condition mood | 5.81 | 1.08 | 4.23 | 0.79 | 3.86 | 0.92 | |
Task compliance | 8.78 | 0.20 | 8.45 | 0.36 | 8.89 | 0.24 | |
Game enjoyment | - | - | 7.36 | 0.33 | 7.18 | 0.32 | t(38) = 0.37 (NS) |
Game difficulty | - | - | 3.68 | 0.49 | 3.69 | 0.43 | t(38) = 0.01(NS) |
After leaving the laboratory, participants then kept a daily structured diary in which they recorded their flashbacks (involuntary visual mental images) to the trauma film over a period of 1 week. Crucially, we found that participants in the visuospatial condition experienced significantly fewer flashbacks over the week than both the no-task (
Measure | No-task (n = 20) | Tetris (n = 20) | Pub Quiz (n = 20) | ANOVA | |||
mean | sem | mean | sem | mean | sem | ||
Diary compliance | 8.60 | 0.94 | 8.80 | 1.32 | 8.75 | 0.85 | |
Recognition memory score | 20.95 | 1.59 | 20.65 | 0.64 | 19.65 | 1.04 |
On a recognition memory test for the trauma film given after 1 week, performance was comparable in the Tetris, Pub Quiz and no-task conditions (
75 non-clinical participants took part in Experiment 2 with no baseline differences between the three conditions (
Measure | No-task (n = 26) | Tetris (n = 26) | Pub Quiz (n = 26) | ANOVA | |||
mean | sem | mean | sem | mean | sem | ||
Age | 24.62 | 1.89 | 22.04 | 1.03 | 22.23 | 1.11 | |
Beck Depression Inventory (BDI) | 6.23 | 1.21 | 6.12 | 1.08 | 5.65 | 4.28 | |
Trait Anxiety (STAI-T) | 36.58 | 2.12 | 36.81 | 1.78 | 37.58 | 1.60 | |
Traumatic Experiences Questionnaire (TEQ) | 1.46 | 0.32 | 0.92 | 0.33 | 1.08 | 0.25 |
No-task (n = 26) | Tetris (n = 26) | Pub Quiz (n = 26) | ANOVA | ||||||
mean | sem | mean | sem | mean | sem | Time | Group | Group* Time | |
Pre-film mood | 3.09 | 0.77 | 4.04 | 0.67 | 3.49 | 0.68 | |||
Post-film mood | 7.05 | 1.11 | 8.52 | 1.01 | 7.31 | 1.03 |
p<0.01.
After the film participants left the laboratory for 4 hours and were free to go about their daily business during this time. Upon return, intervention procedures were identical to that of the three conditions in Experiment 1 (
Participants completed the trauma film paradigm. All participants viewed a traumatic film followed by a 4-hr break where participants went about their daily business. Participants then returned to the laboratory and were allocated to one of three experimental conditions (as in experiment 1) which they completed for 10 min. Participants in the computer game conditions rated their enjoyment of the game. Flashbacks (involuntary memories) were monitored for 1 week using an intrusion diary. After 1 week, diary content and compliance was checked and a test of voluntary memory (recognition memory test) for the trauma film was administered.
No-task (n = 26) | Tetris (n = 26) | Pub Quiz (n = 26) | ANOVA | ||||
mean | sem | mean | sem | mean | sem | ||
Frequency of initial intrusions | 9.84 | 1.31 | 5.65 | 0.94 | 6.00 | 0.67 |
Tetris relative to no-task control p<0.05; Pub Quiz relative to no-task control p<0.05.
Measure | No-task (n = 26) | Tetris (n = 26) | Pub Quiz (n = 26) | ANOVA/t-test | |||
mean | sem | mean | sem | mean | sem | ||
Post condition mood | 4.76 | 1.22 | 3.78 | 0.65 | 4.62 | 0.87 | |
Task compliance | 8.50 | 0.38 | 8.51 | 0.26 | 8.71 | 0.35 | |
Game enjoyment | - | - | 6.60 | 0.28 | 6.46 | 0.28 | t(50) = 0.33 (NS) |
Game difficulty | - | - | 2.89 | 0.35 | 3.69 | 0.32 | t(50) = 0.10 (NS) |
Participants' week long diaries recording flashbacks to the trauma film show that those in the Tetris condition experienced significantly fewer flashbacks over the week than the no-task control condition (
Measure | No-task (n = 26) | Tetris (n = 26) | Pub Quiz (n = 26) | ANOVA | |||
mean | sem | mean | sem | mean | sem | ||
Diary compliance | 8.53 | 0.24 | 8.46 | 0.23 | 8.53 | 0.19 | |
Recognition memory score | 20.48 | 0.50 | 19.25 | 0.64 | 20.03 | 0.52 |
Cohen's d is a standardized measure of effect size in which values between 0.2 to 0.3 are taken to indicate “small” effect, values around 0.5 indicate a “medium” effect and values >0.8 indicate a “large” effect.
We have proposed that engaging in a visuospatial task such as Tetris may offer a ‘cognitive vaccine’
First, the current experiments provide the first evidence that different computer games have differential effects on the development of flashbacks post-trauma. Playing Tetris after viewing traumatic material reduced later flashbacks compared to no-task control (Experiment 1 and 2), whereas the computer game Pub Quiz did not. In fact, Pub Quiz even worsened these post-trauma symptoms in Experiment 1. This was the case despite the games being rated as equally as enjoyable and as of similar difficulty. There is increasing interest in using computer games as novel psychological interventions in healthcare
Second, our data are the first to demonstrate that the beneficial effects of Tetris on later intrusive memories are retained even when played at 4 hours post-trauma. We account for this based on current models of memory consolidation indicating that certain types of memory may be malleable for up to 6 hours
The insights arising from these two studies support the possibility that tasks such as Tetris (a simple visuospatial task) may be developed as a post-trauma intervention to reduce the flashback symptoms of PTSD, and administered up to 4 hours post-trauma. Such an approach would provide a novel alternative to drug treatment or counseling. Current clinical treatment reviews suggest that an existing counseling early intervention for trauma - “critical incident stress debriefing” - may worsen rather than improve PTSD symptoms
The current data also suggest that some computer games, such as the verbal game Pub Quiz, are unlikely to have a beneficial effect post-trauma, and may even have harmful effects on flashbacks when played soon after. A further implication of our findings is that the seemingly simple, innocuous activities people may choose to engage in after trauma - such as playing different types of computer games - may differentially affect their subsequent mental health. The increase in flashbacks when Pub Quiz is played at 30 min suggests is consistent with our account of nature of trauma memory whereby verbal/conceptual interference may worsen flashbacks in the consolidation phase, as outlined in the
In conclusion, our approach follows directly from hypothesis-driven theory from cognitive science of trauma memory and the visual imagery nature of flashbacks. Data show a differential impact of two types of cognitive task on the development of flashbacks. Thus by directly comparing two computer games and extending the intervention time window, these findings present critical steps required for the development of a computerized ‘cognitive vaccine’ against the development of flashbacks for trauma.
Approval was obtained from the University of Oxford Central University Research Ethics Committee. Healthy volunteers were recruited using advertisements placed online and in the local community. Similar to previous studies
A composite mood score was calculated by summing participants' ratings on three visual analog scales for ‘sadness’ ‘hopelessness’ and ‘depressed’. Visual analog scales for mood were anchored from 0 ‘
After the break, all participants were shown a brief film reminder task in which one neutral but recognizable static image from each of the 15 film clips was presented (via slides in PowerPoint) in order to “reactivate” memories for the trauma film
During the 10-min experimental task manipulation (above) participants in all conditions recorded the frequency of initial intrusions of the trauma film. Afterwards participants completed mood assessments and rated to what extent they had followed instructions (task compliance) on a visual analog scale anchored from 0 ‘
Participants then kept a daily diary for 1 week, in which they recorded the occurrence (frequency) of their flashbacks and then briefly described the content of each of their flashbacks separately (for verification) e.g. a human knee with blood. Flashbacks backs were described as spontaneously occurring image-based intrusive memories of scenes from the trauma film
On return to the laboratory 1 week later, the experimenter went through the diary to verify that the content of each of the flashbacks came from a scene in the trauma film watched 1 week earlier (if not, they were discounted). Participants rated the extent to which they had been able to accurately record their flashbacks in the diary (diary compliance) from 1 ‘
Participants completed a recognition memory task as a measure of voluntary memory for the trauma film. The recognition memory task consisted of 30 written statements describing the film (e.g. A policeman stands watching the wreckage whilst making notes on a clipboard) to which they responded ‘true’ or ‘false’. Participants were then debriefed and reimbursed for taking part.
The method for Experiment 2 was largely identical to that of Experiment 1 with the main difference being the extension of the time frame from 30 min to 4 hours between watching the film and conducting the experimental task manipulation (Tetris, Pub Quiz or no-task control). In addition a 12-min trauma film
78 participants (aged 18–57; mean age = 22.9; 42 females) took part in Experiment 2. After post film mood assessments participants left the laboratory, resumed their daily business, and returned 4 hours later. Upon their return participants were shown a brief reminder task in which one recognizable static image from each of the 11 film clips was presented, via slides in PowerPoint, in order to “reactivate” memory for the film. They were then allocated to one of three 10-min conditions where they either played Tetris, Pub Quiz or sat quietly in the no-task condition. As in Experiment 1 participants recorded their initial intrusions during the 10-min experimental task manipulation. Participants then kept a structured diary for 1 week to record flashbacks. Measures of diary compliance and a recognition memory task were taken 1 week later.
We would like to thank Alan Slater for the illustrative material in this publication. We would also like to thank Katie Grimwood for assistance with data collection. We are grateful to the providers of the film clips used in these studies, including LyleBailie International, Kino International and the British Film Institute National Archives and David Large at the Royal college of Surgeons in Edinburgh.