The Role of Eye Movement and Memory Retrieval in Memory
Enhancement
Sarah Vogels January 19, 2017 University of Amsterdam Student number: 10582045 dhr. dr. R.H. Phaf; C.B. Mulder Number of words: 6,575Table of Contents
Abstract ... 3 Introduction ... 4 Method ... 13 Participants ... 13 Design ... 13 Material ... 13 Categorization Task ... 14 Rehearsal-Preventing-Task ... 14 Manipulation Task ... 14 Recall Task ... 15 Exit-Interview ... 15 Procedure ... 16 Data analysis ... 17 Results ... 18Complementary & Explorative Analyses ... 20
Discussion ... 24
Abstract
EMDR (eye movement desensitization and reprocessing), just like SIRE (saccade-induced retrieval enhancement), has been generating a lot of controversy due to the fact that the relevance of eye movement to memory could not yet be explained. Insight into SIRE could therefore imply progress for EMDR. This study aims to investigate whether the
simultaneous retrieval of negative words and execution of eye movements enhance memory recall. The relevance of simultaneity of retrieval and eye movements for SIRE was manipulated by an incidental learning task. The results showed no SIRE effect (M = - 0.002, Cohen's d = - 0.045, 95% CI [-0.040, 0.035]), both independent and dependent of reimagination. Yet, performance seemed to differ between conditions depending on word categories (disaster/emotion). SIRE therefore needs more follow-up research to define whether it is dependent on a specific nature of information in order to occur.
Introduction
In 2014, the current refugee crisis reached its peak, recording the highest number of refugees since World War Two. As of January 2017, more than 19 million people live in a state of constant uncertainty, have to adapt to new cultural demands and standards, but above all, have to deal with the experiences of their past. It is estimated that about 1 in 5 refugees suffers from a psychological disorder such as post-traumatic stress disorder (PTSD) as a result of trauma.
An efficient and cultural-independent therapy for PTSD is eye movement desensitization and reprocessing therapy (EMDR) (Acarturk et al., 2015). Francine Shapiro first introduced the term after accidentally discovering how her disturbing thoughts seemed to weaken extensively when she followed the falling leaves of trees with her eyes while focusing on her thoughts during a walk outside. She later replicated the situation and found again that moving the eyes horizontally while concentrating on disturbing thoughts made her thoughts less distressful. After conducting this test of her hypothesis, she coined the term EMDR (Shapiro, & Solomon, 1995). From that point onwards, the effectiveness of EMDR has been confirmed multiple times (Lee, & Cuijpers, 2013). Furthermore, EMDR has even been proven superior to other therapies in terms of treating PTSD (Van den Hout, Bartelski, & Engelhard, 2013). However, EMDR still struggles with much controversy due to the fact that its underlying mechanism is still largely unexplained. An essential part of EMDR consists of the patient moving his/ her eyes horizontally in a saccadical manner while focusing their attention on a traumatic incidence. As of today, it cannot be sufficiently explained why the execution of eye movements simultaneous to the retrieval of traumatic memories facilitates coping with a disturbing incident.
Two theories have gained acknowledgment and support for explaining the
The distraction theory argues that producing eye movements and recalling a traumatic event simultaneously leads to an overload of the visuospatial working memory. Subsequently, in order to deal with both tasks at the same time, the working memory would have less capacity left for the traumatic event. By constraining the mental power for dealing with a traumatic event, its emotional content is weakened and the belonging associations and interference for an individual are reduced (Van den Hout et al., 2011). According to this hypothesis, this weakening of emotional content should not necessarily require a memory of a negative, traumatic event in order to occur, but should hold true for positive memories as well. Several studies support this view by arguing that eye
movements serve the same purpose as any other kind of cognitive distracting task (Van den Hout et al., 2011). The distraction theory therefore states that any kind of cognitive distracting task performed simultaneously to the recall of information works equally well to generate the desired changes in intensity of emotional material.
The regulation theory, however, proposes that by performing saccadic eye
movements, a so called „orienting response“ is enabled which searches the environment for potential threats and causes relaxation if none can be found (Phaf, submitted). The orienting response consists of two different responses: the alerting response and the investigatory response. The alerting response is activated when the subject is being confronted with a sudden, negative, environmental stimulus and elicits distress.
Subsequently, an investigatory response is needed to analyze the nature and meaning of the stimuli (Phaf, submitted). If no threat can be determined, relaxation is induced. During an EMDR session, this process could be provoked by asking the participants to imagine the traumatic event and thereby releasing an alerting response. The investigatory
response would then be acted out by performing simultaneous eye movements to realize the safety of the situation and environment. The simultaneous eye movements and
induced relaxation and could therefore be crucial for psychic rehabilitation. Stickgold (2002) supported this theoretical possibility by comparing the saccadic eye movements from EMDR with rapid eye movement (REM) sleeping states. It gained further relevance by Wagner et al. (2001), who showed that the (successful) processing of emotional information is enhanced during these sleep states, which involve high amounts of rapid eye movement and are therefore comparable to saccadic eye movements (Wagner, Gais, & Born, 2001). However, these aforementioned theories contrast each other strongly: while the distraction theory aims for a passive effect, the regulation theory strives to actively motivate the patient to relativize the negative experience by creating an association of relaxation with it. It is thus necessary to acquire further insight into the mechanism of EMDR, to finally exclude and/ or reinforce certain aspects of the already existing therapy for the best possible treatment and recovery (Van den Hout, & Engelhard, 2012).
Saccadic induced retrieval enhancement (SIRE) describes a finding which might offer a new perspective of the functionality of eye movements in not only in EMDR, but also in more fundamental experimental psychological research. As the term “SIRE”
indicates, saccadic horizontal eye movements tended to show an enhancing effect on the retrieval of memories. Yet, despite sharing the execution of saccadic eye movements as their theoretical basis, almost no research exists which links the fields of SIRE and EMDR (Phaf, submitted). Since SIRE has so far produced the strongest findings for negative elements, and EMDR has been applied for negative memories, the relatedness of the two fields cannot be ignored (Phaf, submitted).
Due to the essential element of eye movements in both SIRE and EMDR, an understanding of SIRE could lead to an understanding of the functionality of EMDR. Although being a scientific construct, SIRE could thereby, besides its empirical relevance, have a clinical and social impact on our society and economy. In order to consider insight into SIRE profitable for EMDR research, it is relevant to understand how SIRE research
explains the impact and functionality of eye movements on information processing and retrieval. To relate SIRE and EMDR to each other, it is necessary to consider theories of SIRE and identify their differences and similarities with the earlier revised theoretical
explanations of EMDR. The two most frequently referred to theories of explaining the SIRE effect are the inter-hemispheric interaction theory and the top-down attentional control theory.
The interhemispheric interaction theory was introduced by the researchers who labelled the phenomena of SIRE for the first time: Christman and Propper (2001). They argue that better memory retrieval is caused by an equalization of activation levels of the hemispheres. According to the authors, the alternating saccadic activation therefore leads to an increased inter-hemispheric interaction. The findings that only consistent-handers show a greater effect of SIRE support this view. In this study, only people with consistent-handedness were therefore admitted to participate. Consistent-handers usually have lower levels of hemispheric interaction and thereby have lower memory performance, making them especially prone to the SIRE effect (Lyle, Logan, & Roediger, 2008). However, the theory soon had to be dismissed because vertical eye movements appeared to be as effective as horizontal ones and inter-hemispheric coherence on the EEG did not seem to be increased by the performance of eye-movements (Samara et al., 2011). These findings contradict the inter-hemispheric interaction theory and shifted the attention to the
distraction theory (Gunter & Bodner, 2008).
The above discussed distraction theory at a first glance seems to contradict the SIRE effect: Performing eye movements and retrieving emotional memories
simultaneously leads to an overload of the working memory and thereby weakens the memory instead of enhancing it (as SIRE would predict). Then again, combining this theory with the regulation theory could constitute a new theory, which does account for SIRE: the top-down attentional control theory. According to this theory, overloading the
working memory, as the distraction theory predicts, could increase attentional control and thereby suppress possibly interfering emotional content by performing an orienting
response, as the regulation theory holds. Top-down processes would therefore overcome bottom-up influences. According to this theory, performing eye movements leads to an increase of top-down attentional control, resulting in better memory performance (Lyle, & Edlin, 2015). If the working memory at first gets overloaded, particular emotional memories are weakened and will thereby be more prone to top-down attentional control. Attentional control can therefore suppress possibly interfering emotional content (important for EMDR) and at the same time enhance the retrieval of memories (SIRE). “Overloading the working memory by the execution of eye movements and having enhanced attentional control afterwards can thus be complementary mechanisms that cooperate to produce SIRE” (Phaf, submitted). A second account of this theory gives a similar explanation of the relationship between eye movements and SIRE. The second version of the top-down attentional control theory states that performing eye movements increases attentional control and makes words which are more difficult to retrieve more prone to a SIRE effect. However, in this version, the working memory part is not necessary. It is however
necessary that the words which become more prone to a SIRE effect, because of the execution of eye movements, are elements that are difficult to retrieve. When comparing types of words, negative words tend to be more difficult to retrieve and would therefore need more attentional control to access (Strien & Bourna, 1988). Furthermore, applying negative words would allow an overlap in design of SIRE and EMDR. Given the effect was only found for negative words by Phaf (submitted), we chose to focus on negative words exclusively in this study.
In addition to the lack of confirmed theoretical explanation, SIRE has dealt with a lot of controversy. Matzke et al., for instance, failed to replicate the SIRE effect found by Lyle et al (2008) and Samara, Elzinga, Slagter and Nieuwenhuis (2011) (Matzke, Nieuwenhuis,
Van Rijn, Slagter, Van der Molen & Wagenmakers, 2015). However, these results may rather highlight a drawback of the design instead of a theoretical limitation (Phaf,
submitted). For instance, Phaf argues that much information and progress in psychology gets lost, amongst other by the total reliance on statistical significance instead of valuing the effect size. The non-replication of SIRE by Matzke et al.(2015) is only another example of the failure of design to him. However, in contrast to Lyle et al. (2008), they applied a recency buffer to control for a short-term memory effect. Naturally, this could be an explanation of the SIRE effect found by Lyle et al. (2008). The SIRE effect Lyle et al. (2008) have discovered might therefore, due to this drawback in their design, only represent a short-term memory effect. Additionally, Matzke et al. (2015) did not examine what people engaged in during the execution of eye movement. Whether people retrieved words during the eye movements or not was not controlled for. This lack of experimental control could indeed facilitate invalid results. Furthermore, both studies from Lyle et al. (2008) and Matzke et al. (2015) investigated the possibility of a SIRE effect only for neutral words. However, If the second version of the top-down attentional control theory holds true, memory should only be enhanced for weak elements like negative words. This could be another explanation of the contradicting results research of SIRE has undergone so far.
Since EMDR has been proven efficient, research of SIRE should, due to their similarities, be as closely designed to the EMDR set up as possible to guarantee a mutual profit in terms of theoretical insight and progress, as Phaf suggests. As mentioned above, the second version of the top down theory demands weak- like negative information as a necessity to find a SIRE effect. Likewise, the regulation theory may only explain EMDR based on a negative stimulus eliciting the above-illustrated “orienting response”.
Considering these similarities and that EMDR deals with traumatic, and therefore negative memories, SIRE approaches should focus on negative words. Research of Phaf
explaining SIRE. He compared memory retrieval for negative, neutral and positive
information and could only detect SIRE for negative words. These results may emphasize the significance of the negative valence of the words for enhanced retrieval.
Another crucial element, which has surprisingly come to be dismissed in most of the SIRE research, is the explicit controlling for simultaneity of retrieval and eye movements. This forms a fundamental aspect in EMDR research. Except for the experiments of Phaf (submitted) and Van den Hout (2013), retrieval of the material was performed either before or after eye movements, the participant did not receive instruction when to retrieve the material or the participant was prior to the whole experiment informed of having to
remember it (Christman, Garvey, Propper & Phaneuf, 2003; Lyle, Logan & Roediger, 2008; Matzke et al., 2015). Research in which retrieval did in fact occur at the same time as eye movements and showed an effect, gives further indication that this might indeed be an important aspect to address to the debate and might lead us to deeper insight in the working mechanism of EMDR by equalizing the procedure of SIRE to EMDR (Brunyé, Mahoney, Augustyn & Taylor, 2009). Not controlling for this simultaneity could be another, fundamental explanation for the non-replication of SIRE by Matzke et al. (2015).
Regarding all these differences concerning the (not-) controlling for time simultaneity and the varying results, the question arises if it is necessary to retrieve memories during the execution of eye movements in order to find a SIRE effect.
When looking closely at the aforementioned theories to explain the EMDR and the SIRE phenomena, it is inevitably noticeable that all of those demand a specific time of retrieval.Therefore, time of retrieval might be a crucial aspect and indicate a major role in explaining the phenomena of SIRE.
In this study we therefore address the question: Is simultaneity of retrieval and eye movements necessary or even crucial for memory enhancement of (negative) words? Our main hypothesis will be contrasting the results of Matzke et al. (2015) by stating: memory
retrieval during eye movements will result in an overall better performance of memory in comparison to no retrieval during eye movements (hypothesis 1). In an incidental learning task, participants of this study were therefore asked to categorize a list of negative words into two categories (“disasters” and “emotions”). Then, they performed a “rehearsal-preventing-task”. This intended to strain the working memory and to serve as a form of recency buffer. Finally, participants were instructed to recall as many words as possible of a specific category while either performing saccadic horizontal eye movements or fixating on a point (control condition). If retrieval of the words during eye movements is necessary, a SIRE effect will only be found for words which have been retrieved during the execution of eye movements compared to words which have not been retrieved during eye
movements (see graph1).
If simultaneity of retrieval and eye movements is indeed crucial for finding a SIRE effect (hypothesis 1), the regulation theory will gain relevance. In order to confirm our first hypothesis, we expect a difference between the conditions: Participants in the „eye-movement condition“ should show a significantly stronger SIRE effect than participants in the „fixation condition“. Furthermore, we expect to find a difference within the
„eye-movement condition“ considering the words of the recalled-category or not-recalled category. We expect the words recalled during the eye-movements to be more frequently recalled than words of the other category. Within the „fixation condition“ we do not expect any differences in results in terms of recalled or non-recalled category.
The top down attentional control hypothesis holds that SIRE emerges through the execution of eye-movements, although regardless of simultaneity of retrieval (hypothesis 2). In order to support this view, we expect that a SIRE effect will be found only in the eye-movement-condition for words of both categories (reimagined and non-reimagined). In this scenario, the distraction theory will come to be dismissed as a probable explanation for SIRE (graph 3). The top-down theory does not necessarily have to be dismissed, if arguing
from the perspective of its second version, which requires negative words but no simultaneity to enable a SIRE effect.
The working memory hypothesis (hypothesis 3) will contradict the regulation
hypothesis by stating: Eye-Movements do not enhance retrieval, but weaken it (no SIRE). For this, memory performance is expected to be weakened after the simultaneous
execution of eye movements and word retrieval. This hypothesis is in line with EMDR and would support the working memory approach and thereby question the construct of SIRE. Moreover, the first version of the top-down attentional control theory, will, due to its working theory component, be supported.
Finally, our study will be focused on our main hypothesis: hypothesis 1. We expect to confirm the theories of EMDR (distraction and regulation theory) if time simultaneity is necessary to find a SIRE effect. Confirming theories which explain the EMDR effect in a SIRE study would indicate a close relation of both constructs and provide clearer insight in their functionality, as well as new opportunities for structural improvements of the EMDR therapy.
Graph 1. Hypothetical Results for each hypothesis and corresponding theory
Regulation (Hypothesis 1) Top-Down (Hypothesis 2) WM (Hypothesis 3)
Expected Results according to Theories
EM-Retrieval EM- No Retrieval Fix-Retrieval Fix- No Retrieval
Method
Participants
Forty-three students (18-30 years old) participated in the study. Inclusion-criteria were (1) Dutch as first language; (2) Consistent right or left-handedness (see Lyle et al. 2008) and (3) No history of EMDR or SIRE research experience. Participants were
rewarded with either academic credit or money or participated voluntarily. Motivation was defined by the quantity of incorrectly categorized words in the categorization task.
Participants, who assigned more than 10 percent of the words to the wrong category were excluded from the analyses.
Design
Our experiment was composed as a 2x2 mixed factorial design. Between subject variable was eye movement: participants were distributed randomly in either the
manipulation condition („eye movement condition“) or into the control condition („fixation Condition“). Within subject variable was simultaneity of retrieval (reimagined vs. non-reimagined word category) during the execution of their eye movements/ fixation. The outcome variable was quantity of recalled words.
Material
Instructions and stimuli were presented on a white background of an 23' Asus VG246HE monitor with a 1920x1080 resolution and 100Hz refresh rate. Participants were seated in a height-adjustable chair. As a regulation of the light, an adjustable lamp was set to dimly light before each session.
Categorization task: The participants were asked to classify words into two categories: disaster or emotions. The word list to be categorized and implicitly learned consisted of 48 words: 24 words belonged to the category „disasters“ (Dutch: „Rampen“) and 24 words to the category „emotions“ (Dutch: „Emoties“). Word categories were matched for word length and word frequency in Dutch language use (Stienen, 2004). Words were presented in random sequence, independent of category, in the center of the screen. Word Category labels were positioned on the right and left lower side of the screen. Category name location (right or left lower side) was randomly distributed for participants. After the participant typed the words in a field above the correct category label, both the typed in and the centrally located word disappeared and the next word appeared which had to be categorized. Words appeared in white color ( Arial font, size 32) on a black screen.
Rehearsal-preventing task: Participants were instructed to count back in 2s from 500 aloud. After one minute, the instructor stopped the task.
Manipulation task: Horizontal saccadic eye movements were induced by alternating a fixation cross (+, font size 32, white on black screen) every half second between left and right sides in a distance of 30 cm (53° visual angle). On each side, the fixation cross
appeared 30 times. In the fixation condition a cross (+, font size 32, white on black screen) alternated every half second with a small “x“-letter (x, font size 32, white on black screen) in the center of the screen. Each sign appeared 30 times. Category selection was
randomly distributed. For this task, participants were asked to place their head on a chin rest, which was positioned in a 30 cm distance from the screen.
Recall task: Participants were instructed to type as many words as they could remember from the entire categorization task in. After typing one word, the participant pressed “enter”. The entered word would then disappear and the participant could continue. The experiment stopped when the participant freely announced that he or she was not able to recall any more words.
Exit interview: Finally, the instructor engaged in a short „exit interview“ to check for any possible individual differences in insight, strategy, performance or state of well-being, which might have influenced the experiment in a biasing way.
Procedure
Before starting the experiment, the participant was told that in this study, his/ her attention would be tested and that she/ he will be confronted with negative words. The participant was given an informative brochure of the experiment to read and subsequently asked to sign an “informed consent”. Then the participant was instructed to fill in a
handedness inventory upon which was decided whether he/ she may participate in the study. For participants scoring at least 9 out of 10 (signifying strong righthandedness) or -9 out of -10 ( strong left-handedness), points were admitted to the study. All participants then completed the categorization, rehearsal-prevent, manipulation and recall task and finally took part in the exit interview. The manipulation task took five minutes in total. In both conditions, all participants were asked to try to retrieve as many words from a specific category given by the instructor, during the whole (manipulation) task. In the last task, the recall task, participants were first informed that they had actually been participating in a memory and not an attention experiment. The instructor was present during the whole experiment, and could thereby check whether participants executed the eye-movements / fixation properly.
Data analysis
The main question of this study, whether simultaneity of retrieval and eye movement execution is crucial for a SIRE effect, compromises the hypothesis that when retrieval is not performed parallel to the execution of saccadic horizontal eye movements in time, memory will not be enhanced (no SIRE effect will be found). Results were compared in terms of their proportion; means and percentages of recalled words were calculated in total and of the retrieved or non-retrieved words and finally those were plotted in graphics. Based on those results, first conclusions related to the in the introduction discussed
hypotheses and outlined graphs were drawn. Subsequently, effect sizes and confidence intervals by applying “Cohen's d” were applied to further analyze our results. This method was chosen as the main analysis instead of a statistical test, because we consider the magnitude of an effect to be of greatest informativeness (Cumming, 2013). Indications of (theoretical) trends for further research can thereby become evident, without necessarily finding a “significant” effect in this particular study. Each study might be prone to flaws in design, but this should not interfere with or even inhibit theoretical progress. Effect sizes may reveal the magnitude of an effect much more precisely than by merely differentiating between a well or non-existence. Statistical significance, may only indicate whether an effect is due to chance or not and can therefore not provide any information of the
dimension of an effect. We therefore focussed on the outcome which could be drawn from means, effect sizes and confidence intervals. We thereby hoped to define whether an overall SIRE effect will be apparent.
Results
All 43 participants met the inclusion-criteria and could therefore participate in the experiment. However, eight participants had to be excluded afterwards. The first two participants were presented the wrong word list. During the third trial, the experiment reported an error and one participant reported that the eye-movements were causing pain to her, so that the experiment had to be stopped. Based on the information given in the exit interview, we were able to define 4 different groups: Successful Imagination of only one category (NGroup 1 = 23); imagination of both categories (NGroup 2 = 12); no successful imagination of only one category (NGroup 3 = 7); no information given about depth and/ or success of imagination (NGroup 4 = 1). Through this, we divided and classified all participants into the different levels of word imagination. We excluded participants of group 3 and 4 from the data analyses, because their results may not add any information about the existence and/or functionality of SIRE. All 35 remaining participants were equally distributed over conditions (NBil = 19 (NbilDisaster = 9; NbilEmotions = 10); NFix = 16 (NfixDisaster = 8; NfixEmotions = 8)) and compared in frequency of recalled reimagined and non-reimagined words.
The results showed that participants did not differ in the total amount of recalled words between conditions (MD = -0.002, Cohen's d = -0.045, 95% CI [-0.040,
0.035]).Participants in the eye movement (EM) condition recalled an equal amount of words as participants in the fixation (Fix) condition. Furthermore, conditions did not differ to a considerable extent between the reimagined and non-reimagined category (Figure 4). Neither an effect for words of the reimagined category (MD = -0.001, Cohen's d = -0.017, 95% CI [0.048, 0.046]) nor for the nonreimagined category (MD = 0.003, Cohen's d = -0.031, 95% CI [-0.061, 0.055]) on condition could be found. This contrasts our expectation that time simultaneity is crucial for SIRE. However, within the EM Condition a small
Cohen's d = -0.241). Although within the Fix Condition a difference between reimagined words and non-reimagined words was detected as well, this effect appeared to be very small (MD = 0.014, Cohen's d = -0.147). Given the however rather limited dimension of this effect difference, it most likely cannot be ascribed to our manipulation. Better recall of a category which has been retrieved recently is naturally expectable. Yet, it is possible that the rather small difference hints at a drawback in design instead of the functionality of the construct itself. The finding that within the manipulation condition, a difference between reimagined and non-reimagined words was found, might be an indication for a SIRE effect, but with regard to the aforementioned results, the chances seem rather weak and can thus not be concluded based on our findings (Table 1).
Table 1
Quantity Participants, Mean quantity of total, reimagined and non-reimagined words recalled with standard deviation (in brackets)
Condition N Propreimagined Propnon-reimagined Proptotal Eye-Movement 19 0.210 (0.058) 0.194 (0.074) 0.201 (0.048)
Fixation 16 0.211 (0.079) 0.197 (0.095) 0.203 (0.062)
Figure 4. Mean amount of recalled words for both conditions (EM-1 and Fix-2) for both
EM Fix 0,19 0,19 0,2 0,2 0,21 0,21 0,22 Main Analysis Propreimag Propnonreimag ProptotalRecalled
Complementary & Explorative Analyses
Analysis Imagination Groups
To explore and complement our insight into SIRE, the following analysis aimed at discriminating effects of condition and (non-) reimagination of words between the different groups of imagination (see introduction of results).
The detected means and standard deviations (Table 2) indicated no overall difference between groups of imagination between condition (EM and Fix, Figure 6). However, participants in the EM condition recalled more words of the reimagined category in the imagination group 1 than in group 2 (MD = 0.37, Cohen's d = -0.676)(Figure 5). This suggests the existence of a small SIRE effect between condition, depending on the
participant's imagination success and on the re-imagination of words. However, the high recall reported of participants from the EM condition and in group 1, was revealed to only be as high as the performance of participants in group 1 in the Fix condition for words of the reimagined category (MD = 0.005, Cohen's d = 0.072). The difference within group 1 between conditions thereby loses its strength and informative value in reference to SIRE. Therefore, no difference between groups of imagination could be discovered, which could possibly broaden our insight in SIRE in terms of our detected results.
Table 2
Quantity Participants, Mean quantity of total, reimagined and non-reimagined words recalled with standard deviation (in brackets) for both conditions and both imagination groups
Condition/ Group N Propreimagined Propnonreimagined Proptotal Bil/ Group 1 13 0.221 (0.059) 0.184 (0.070) 0.202 (0.053)
Bil/ Group 2 6 0.184 (0.050) 0.216 (0.086) 0.198 (0.037)
Fix / Group 1 10 0.216 (0.078) 0.193 (0.108) 0.204 (0.047)
Figure 5. Mean proportion scores for both conditions (EM and Fix) for all Imagination-groups and for the reimagined and non-reimagined category
Figure 6. Mean proportion scores for both conditions (EM and Fix) for both Imagination groups and for the reimagined and non-reimagined category
Both categories One category
0,2 0,2 0,2 0,2 0,2 0,21 0,21 0,21
Analysis Imagination Groups
EM-Proptotal Fix-Proptotal
Both categories One category
0 0,05 0,1 0,15 0,2 0,25
Analysis Imagination Groups
EM-Propreimag Fix-Propreimag EM-Propnonreimag Fix-Propnonreimag
Analysis Word Category
Our second additional analysis compromised the division of the variable “word category” within “reimagined” and “non-reimagined” up into “disaster” and “emotion” words. Participants frequently stated during the exit interview that they subjectively experienced “disaster” words as much more concrete and salient in nature and therefore easier to retrieve. Although, due to counterbalancing our conditions, this contrasted our
expectations, it gave rise to the consideration that words of the category “disasters” could have produced unequal results (especially when considering the top-down attentional control hypothesis). If this was indeed the case, not only the valence (negativity) but also concreteness of information could be necessary for SIRE. We therefore revised means (Table 3) and effect sizes to compare possible differences between “disaster” and “emotion” words within the reimagined and non-reimagined word categories for both conditions (EM and Fix). A comparison of the word category within the total recall of both conditions highlighted an unexpected result: Participants in the EM condition appeared to score much higher on “emotions” (MD = 0.20), while in the Fix condition participants scored in line with our prior expectations equally well on both word categories (MD = 0). Both the difference between conditions (EM & Fix) for “emotions” (MD = 0.07; Cohen's d = 0.10) and for “disasters” (MD = 0.013; Cohen's d = -0.28) showed a small, yet
distinguishable effect (Figure 7). Moreover, within the EM condition the effect was up to moderate (MD = 0.020; Cohen's d = -0.416), especially regarding the effect within the Fix Condition (MD = 0; Cohen's d = 0). This might indicate that SIRE does, in fact, depend on the nature of information, being abstract or concrete. However, against our expectations based on theoretical background of EMDR, yet in line with what the results of our study suggest, SIRE might require more abstract information in order to occur.
Table 3
Quantity Participants, Mean quantity of total, reimagined and non-reimagined words recalled with standard deviation (in brackets) for both conditions and both word categories
EM/ Fix (Non-)Reimagined ReimagCat Mean (SD)
EM Propreimag Disaster 0.224 (0.058)
EM Propreimag Emotion 0.196 (0.058)
EM Propnonreimag Disaster 0.155 (0.054)
EM Propnonreimag Emotion 0.229 (0.075)
Fix Propreimag Disaster 0.253 (0.049)
Fix Propreimag Emotion 0.168 (0.083)
Fix Propnonreimag Disaster 0.155 (0.063)
Fix Propnonreimag Emotion 0.238 (0.106)
Figure 7. Mean proportion scores of total recall for both conditions (EM and Fix) for both word categories (disaster and emotions)
Disaster Emotions 0,18 0,19 0,2 0,21 0,22
Analysis Word Category
EM-Proptotal Fix-Proptotal
Discussion
The aim of this study was to investigate the functionality of SIRE, and to define whether simultaneity of retrieval and eye movements is crucial for memory enhancement. Moreover, we aimed to question the results and conclusions by Matzke et al. (2015) by applying certain changes, which should bring SIRE closer in terms of design to that of EMDR. However, we could not find an effect for eye-movements in comparison to fixation. Furthermore, we did not find a difference between recall of reimagined and
non-reimagined words during eye-movements and/ or fixation. Simultaneity of retrieval
therefore may not be crucial for SIRE and/ or explain the results. We also could not find an effect between the groups of participants having imagined word categories differently. However, we did find different results for conditions depending on the type of word recalled.
Our results demonstrated various insights. Our main analysis (analysis 1) indicated no SIRE effect for eye-movements, independent of time of retrieval. Neither for the
reimagined category nor for the non-reimagined category, a difference between eye movements and fixation could be detected. Moreover, when considering recall in its total, participants did not seem to differ between conditions either.
Considering the weak findings contradicting all our hypotheses, we explored our results further. We contrasted the two different groups of participants who either
succeeded in imagining only one category of words during eye-movements or fixation or who imagined both categories (analysis imagination groups). However, no difference between groups based on conditions could be detected. We concluded therefore, that SIRE did not depend on imagination. In our last analysis, however, a difference between the category of words dependent on condition was revealed. Participants of the eye-movement condition recalled “emotion” words much more often than “disaster” words or
participants of the fixation condition for both “emotion” and “disaster” words. With regard to our hypothesis 2 implying the confirmation of the top-down attentional control theory
(second version), the finding that the nature of information might be crucial for recall depending on condition, is of fundamental relevance. As the top-down attentional control theory implies, the nature of elements is crucial to enable the process. Thereby only weak and abstract elements may require higher attentional control and finally lead to SIRE. If however, the elements that are to be retrieved are not weak, this could be another explanation for the non-replication by Matzke et al. (2015). Negativity and abstractness could therefore be necessary characteristics of information to guarantee the weakness of
elements and thus, a possibility for SIRE to occur. Furthermore, the finding of a decreased
memory for disaster words in the EM condition might reflect a memory weakening due to eye movements for concrete scenarios. If follow-up research should support this
possibility, this could have major implications for the EMDR context. Nature of words could thereby explain the functionality of eye movements for trauma therapy. When imagining a traumatic event and performing eye movements, it could be suggested that memory will decrease. If however imagining only the emotion felt at the certain event, this would, due to the abstractness of emotions, induce memory-enhancement.
Although distinguishable, the effect which we found in the last analysis was rather small. However, our experiment contains certain limitations which might have either decreased or increased the effect to its found magnitude.
A first questionable design of our study was the employment of an incidental instead of an intentional learning task (categorization task). When considering EMDR, the
fundamental motivation of human to remember their personally relevant or shaping experiences can be assumed rather high. Although people suffering from PTSD rather desire to forget their experiences, they did not have this motivation prior to the event, because they mostly did not know it would occur. The motivation to remember random
words can however be regarded as rather low, because it has no personal or indirect value (“involvement theory”). Furthermore, not informing the participant that she/ he had to
remember these words afterwards could decrease that motivation even more and also eliminate or inhibit possible social motivations, like for example the motivation to perform very well to impress the researcher. Motivation could therefore be another factor of influence to the SIRE effect and should thus, for example by applying an intentional learning task, be controlled for in following studies. A suggestion for potential intentional learning tasks in follow-up research could be to ask the participants to connect the given words in a way so that he/ she can retrieve them later as easy as possible. This could imply for example writing a short story or text, by using all of the words that had to be remembered.
A second limitation of our study was possible physical impairment. Participants frequently reported to having experienced the eye-movements as cognitively exhausting. As reported earlier, one experiment even had to be stopped because the participant was suffering of increasingly strong headaches caused by the alternating crosses. These experiences might have been caused by the rather bright illumination of the screen, a phenomenon which has been coined “computer vision syndrome”. Furthermore, the
experiment was aborted on a computer with a particular big screen. This however, was not taken into account during the technical design of the experiment. Distance between
alternating crosses has thus presumably been wider than intended. The cognitive influences of this might therefore have interfered in our manipulation and experimental process. Especially for theories like the “working memory theory”, the exclusiveness of performing only two tasks at the same time would be crucial for the type of outcome. However, if working at the computer forms a third tasks or rather challenge, then the working memory would predict no efficient outcome. A suggestion for further research could, based on the participant's reports and the theoretical background of the “Computer
vision syndrome”, be to either adjust the brightness of the screen to a more convenient level, integrate short breaks to rest or close the eyes between eye-movements/ fixation, or even introduce a method of inducing saccadic eye movements without a computer. The last method would then also be closer to the initial design of EMDR, adding another informative source of the relation between EMDR and SIRE.
Based on these conclusions we can reject our first hypothesis, that memory retrieval during eye movements will result in an overall better performance of memory in comparison to no retrieval. Although the drawbacks of our design might still account for that, the detected results revealed no indication of this possibility. Therefore, the regulation theory is based on these results weakened in its reliability. The second hypothesis, that SIRE emerges through the execution of eye-movements, regardless of simultaneity of retrieval, is being supported merely to a limited extent (with regard to the analysis of word category). Further follow-up research is necessary to define if the nature of words is essential for SIRE to support this hypothesis and the connected top-down attentional control theory. The third hypothesis (that eye movements do not enhance retrieval but weaken it) and the thereby connected working memory theory and first version of the attentional control theory, are not supported by the results either, but can with regard to the discussed physical interferences not be fully rejected.
The controversy EMDR and SIRE both deal with can not be dissolved by this study. Although we found a small effect for SIRE, it's extent was extremely small and its
implications therefore should be even more limited. However, our study may highlight an indication for further follow-up research for the SIRE effect. Depending on the nature of information, SIRE would occur after the performance of eye-movements in contrast to fixation. This does not only emphasize the reliability of the top-down attentional control theory, but might further explain how research for SIRE until now has demonstrated that varied and controversial results. If indeed the nature of information is crucial to discover a
SIRE effect, this factor needs to be further investigated and controlled for. An option for follow-up studies could be to take nature of information as an independent variable into the design and compare memory retrieval of abstract and concrete words by an intentional learning task. Given it's necessity in terms of EMDR and the refugee crisis, but also empirically, to solve the controversial debate of the role of eye movements, society but mostly science could benefit from these further insights. Finally, although this study might in regards to the small effect, add only little to the fundamental question in research whether SIRE does exists or not, our results, especially in terms of word distinction for SIRE, emphasize certainly one insight, which Albert Einstein once introduced: “ The important thing is not to stop questioning”.
References
Acarturk, C., Konuk, E., Cetinkaya, M., Senay, I., Sijbrandij, M., Cuijpers, P., & Aker, T. (2015). EMDR for Syrian refugees with posttraumatic stress disorder symptoms: Results of a pilot randomized controlled trial. European Journal of
Psychotraumatology, 6.
Brunyé, T.T., Mahoney, C.R., Augustyn, J.S., & Taylor, H.A. (2009). Horizontal saccadic eye movements enhance retrieval of landmark shape and location information. Brain and Cognition, 70, 279-288.
Christman, S. D., & Propper, R. E. (2001). Superior episodic memory is associated with interhemispheric processing. Neuropsychology, 15, 607.
Christman, S. D., Garvey, K. J., Propper, R. E., & Phaneuf, K. A. (2003). Bilateral eye movements enhance the retrieval of episodic memories. Neuropsychology, 17, 221.
Cumming, G. (2013). Understanding the new statistics: Effect sizes, confidence intervals, and meta-analysis. Routledge.
Gunter, R.W., & Bodner, G.E. (2008). How eye movements affect unpleasant memories: Support for a working-memory account. Behaviour Research and Therapy, 46, 913-931.
Lyle, K.B., Logan, J.M. & Roediger, H.L. (2008). Eye movements enhance memory for individuals who are strongly right-handed and harm it for individuals who are not. Psychonomic Bulletin & Review, 15, 515-520
Lyle, K. B., & Edlin, J. M., 2015, Why does saccade execution increase episodic
memory retrieval? A test of the top-down attentional control hypothesis. Memory, 23, 187-202.
Matzke, D., Nieuwenhuis, S., van Rijn, H., Slagter, H. A., van der Molen, M. W., & Wagenmakers, E. J. (2015). The effect of horizontal eye movements on free
recall: A preregistered adversarial collaboration. Journal of Experimental Psychology: General. 144
Phaf, R.H. (2016) Replication requires psychological rather than statistical hypotheses: An example of eye movements enhancing word recollection. Frontiers in Psychology.
Samara, Z., Elzinga, B.M., Slagter, H.A., & Nieuwenhuis, S. (2011). Do horizontal saccadic eye movements increase interhemispheric coherence? Investigation of a
hypothesized neural mechanism underlying EMDR. Frontiers in Psychiatry, 2, 4 Shapiro, F., & Solomon, R. M. (1995). Eye movement desensitization and
reprocessing. John Wiley & Sons, Inc.
Stickgold, R. (2002). EMDR: A putative neurobiological mechanism of action. Journal of clinical psychology, 58, 61-75.
Stienen, B. (2004). Emotiewoorden.Identificatie met behulp van een perceptuele clarificatietaak.
Van den Hout, M. A., Engelhard, I. M., Beetsma, D., Slofstra, C., Hornsveld, H., Houtveen, J., & Leer, A. (2011). EMDR and mindfulness. Eye movements and attentional breathing tax working memory and reduce vividness and emotionality of aversive ideation. Journal of Behavior Therapy and Experimental Psychiatry, 42, 423-431. Van den Hout, M. A., & Engelhard, I. M. (2012). How does EMDR work?. Journal of
Experimental Psychopathology, 3, 724-738.
Van den Hout, M. A., Bartelski, N., & Engelhard, I. M. (2013). On EMDR: Eye movements during retrieval reduce subjective vividness and objective memory accessibility during future recall. Cognition & emotion, 27, 177-183.
Wagner, U., Gais, S., & Born, J. (2001). Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep. Learning &
