Following traumatic experiences, some people have flashbacks, which are also called “intrusive memories” and are characterized by involuntary images of aspects of the traumatic event. Although people often try to simply forget traumatic memories, this approach is not very effective. Instead, previous research suggests that a better approach may be to try to change the memory during reconsolidation, which occurs following the reactivation of a previously formed memory. Importantly, research also shows that during reconsolidation, traumatic memories can be altered and weakened to the point that they are no longer intrusive.
Because intrusive memories of trauma are often visual in nature, James and colleagues (2015) sought to explore whether completing a visuospatial task (e.g., tetris) while a memory is reconsolidating would interfere with the storage of that memory, and thereby reduce the frequency of subsequent intrusions. They hypothesized that only participants who complete a visuospatial task after reactivation of the traumatic memories would experience a reduction in intrusive memories. In comparison, simply completing a visuospatial task (without reactivation to trigger reconsolidation) or reactivation (without a visuospatial task), would not reduce the occurrence intrusive memories.
To test their hypothesis, the authors conducted an experiment (N = 72, n = 18 per condition). The procedure is summarized as follows:
Trauma Film: All participants viewed a series of video clips of graphic violence (e.g., a person getting hit by a van while using his phone as he crosses the road) as a way to create memories that should become intrusive memories. Participants then went home and recorded the number of intrusive memories they experienced over the next 24 hours. Because this is before the experimental manipulations, all groups were predicted to have an equal occurrence of intrusive memories during the first 24-hours (called Day 0).
Experimental Task: After this 24-hour period, the participants returned to the lab and completed the experimental task. The experimenters randomly assigned participants to ONE of the following conditions:
Intrusive Memories: All participants were asked to record the number of intrusive memories that they experienced over the next seven days (Days 1 to 7).
Intrusion-Provocation Task: After the seven days had passed, participants completed an Intrusion-Provocation Task, in which they were shown blurred images from the trauma film and asked to indicate whether the blurred image triggered an intrusive memory.
The data are stored in the file
James et al 2015 Experiment 2.csv file available on the
course webpage. Remember to store the file in your working
directory, which is typically the same folder where you have
your lab report markdown document. We load a csv file using the command
read_csv.
The independent variable we are interested in is named
Condition, and it has four levels. The levels are
1, 2, 3, and 4 in
the data-frame, so we rename the levels of the Condition column with
words so we know what they refer to. These will correspond to the four
treatment levels in shown in figure:
-No-task (control) -Reactivation Plus Tetris -Tetris only -Reactivation only
tetris <- read.csv("James et al 2015 Experiment 2.csv")
tetris$Condition <- as.factor(tetris$Condition)
levels(tetris$Condition) <- c("Control",
"Reactivation+Tetris",
"Tetris_only",
"Reactivation_only")Condition, Day_Zero_Number_of_Intrusions,
Days_One_to_Seven_Number_of_Intrusions and
Number_of_Provocation_Task_Intrusions. Create a new
dataframe called tetris.sub that has only these three columns.Remember that before we do any kind of inference or modeling, we should explore the data!
Compute numerical summary statistics for the dependent variables
Day_Zero_Number_of_Intrusions,
Days_One_to_Seven_Number_of_Intrusions and
Number_of_Provocation_Task_Intrusions across each of the
four groups. What do you notice?
Plot the data in an informative way. Describe your observations and place the plot below them in your markdown file.
The methods of ANOVA have a number of assumptions. Let’s consider
the assumption of normality. Plot histograms of each group (each
treatment) versus the dependent variables, and comment on the assumption
of normality. You may use the facet_wrap function in
ggplot to make this very simple.
We first want to show that all of the conditions (i.e. groups) have
the same number of intrusive memories during the 24-hours prior to the
Experimental Task. In other words, we want to see if there was a
difference in the groups before the treatment was applied.
Let’s compare the means of the conditions for the variable
Day_Zero_Number_of_Intrusions using ANOVA.
aov(Day_Zero_Number_of_Intrusions ~ Condition, data=tetris.sub)## Call:
## aov(formula = Day_Zero_Number_of_Intrusions ~ Condition, data = tetris.sub)
##
## Terms:
## Condition Residuals
## Sum of Squares 2.4861 345.1667
## Deg. of Freedom 3 68
##
## Residual standard error: 2.252994
## Estimated effects may be unbalancedThis is hard to interpret! Fortunately, we can wrap this inside the
summary function to get the output in a mway that looks
familiar from class:
summary(aov(Day_Zero_Number_of_Intrusions~Condition, data=tetris.sub))## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 3 2.5 0.829 0.163 0.921
## Residuals 68 345.2 5.076The large \(p\)-value is indicating that there is no significant difference in the mean number of intrusions in the groups on day 0, before the treatment. So we fail to reject the null hypothesis that, on average, each group had the same number of intrusions on Day 0.
Now let’s consider the main research question:
Day_One_to_Seven_Number_of_Intrusions, and report your
conclusion.Note: this next part goes a bit beyond what we have done in this class, but is an important part of an ANOVA workflow.
In the context of one-way ANOVA, a planned comparison is when you decide in advance to focus in on a few scientifically sensible comparisons rather than every possible comparison between groups. Remember that when we are doing multiple comparisons, we need to control for the total Type I error, called the family-wise error rate (i.e., the probability of making a Type I error in any test). There are many ways to do this, but we are going to use Tukey’s Honestly Significant Difference test.
Take a look at the original paper, and compare the results of your individual comparisons with the “Results” section from Experiment 2. Why are your conslusions somewhat different? (Hint: they used two-sample t-tests for each comparison)
Finally, you want to test whether the conditions differed on the
Intrusion-Provocation Task (if you have forgotten what this means, see
description in introduction). Use the variable called
Number_of_Provocation_Task_Intrusions and report your
conclusions.
If there is a significant ANOVA test in the previous question, use Tukey’s HSD to determine if the “reactivation + Tetris” condition experienced fewer intrusions than the other three groups for the Intrusion-Provocation Task.
Original Article: James, E. L., Bonsall, M. B., Hoppitt, L., Tunbridge, E. M., Geddes, J. R., Milton, A. L., & Holmes, E. A. (2015). Computer game play reduces intrusive memories of experimental trauma via reconsolidation-update mechanisms. Psychological Science, 26, 1201-1215.
Original lab activity created by Kevin P. McIntyre, © 2016, and further adapted by your professor.