Mechanism of EM in EMDR: Change strength of semantic associations
Description
Research background & objectives: Based on REM-sleep dependent memory reprocessing model, this study examined possible therapeutic mechanisms of eye movements in Eye Movement Desensitization and Reprocessing (EMDR; Shapiro, 1989, 1995, 2001). Stickgold (2002) proposed that sleep induced change in associative memory via activation of weak association during REM state and EMs functioned as REM sleep to integrate the episodic memory of trauma into general semantic memory. Specifically, the effect of EM in relation to order of relatedness of associations on change of strength of semantic associations for negative words after saccadic bilateral eye movements was examined in the present study.
Methods: Sixty-four college students were recruited as participants. Via semantic priming task, a 2 (eye movement condition: horizontal saccadic eye movements vs. no eye movements) × 3 (strength of semantic association: strong priming vs. weak priming vs. unrelated priming) × 2 (block order of presentation: related prime first vs. unrelated prime first) mixed factorial design was performed, with strength of semantic association serving as within Ss factor and the other two variables serving as between Ss factors. The primes consisted of negative word stimuli. Each trials consisted of an 8 seconds saccadic horizontal eye movement manipulation (200 ms per movement in EM condition), followed by 500-msec fixation point and then prime displayed for 32msec, which was followed immediately by the target. The participant was instructed to read the target as soon as possible and the reaction time was recorded by the compuer. While absolute primeability index for strong (or weak) association was calculated as (RT related – RT unrelated) for strong primes or weak primes, respectively, relative primeability index was calculated as (RT related – RT unrelated)/RT unrelated. The dependent variables were correct (%), error (%), and primeability index for each of strong prime and weak prime. Change of primeability of weak associations for negative stimuli as opposed to that of strong ones under different EM conditions and different block order was compared.
Results: For absolute primeability index, the 2 (EM condition) × 2 (order) × 2 (strength of priming) ANOVA showed that only the interaction effect of EM × strength of semantic association was significant ( F(1, 56) = 5.61, p < .02; partial ω2 = .067), indicating that weak priming significantly exceeded strong priming after EM ( F(1, 56) = 18.01, p < .001; partial ω2 = .210), while the opposite was true after non-EM ( F(1, 56) = 25.86, p < .0001; partial ω2 = .280). Further, for weak prime, the priming effect was stronger after EM than after non-EM ( F(1, 56) = 74.62, p < .0001; partial ω2 = .535). For relative primeability index, it also revealed that only the interaction effect of EM × strength of semantic association was significant ( F(1, 56) = 6.09, p < .02; partial ω2 = .074), and tests of simple main effects showed similar patterns as those of absolute primeability index.
Conclusions: 8s EMs was associated with change of strength of semantic associations. While EM didn’t weaken the primeability for the strong association, EM did enhance the primeability for weak associations. Further, weak priming significantly exceeded strong priming after EM, while the opposite was true after non-EM. The results echoed REM-sleep dependent memory reprocessing model, suggesting that EM in EMDR might reflect a shift in associative memory systems by activating different strength of associations of negative semantic nodes for different semantically related words. Given that order of relatedness didn’t play a role and previous research showed that EM resulted in decreased vividness and emotionality and generated greater amount of associations for negative stimuli as well, the implications of the present results from theoretical and psychotherapy point of views and future research possibilities are discussed.
Methods: Sixty-four college students were recruited as participants. Via semantic priming task, a 2 (eye movement condition: horizontal saccadic eye movements vs. no eye movements) × 3 (strength of semantic association: strong priming vs. weak priming vs. unrelated priming) × 2 (block order of presentation: related prime first vs. unrelated prime first) mixed factorial design was performed, with strength of semantic association serving as within Ss factor and the other two variables serving as between Ss factors. The primes consisted of negative word stimuli. Each trials consisted of an 8 seconds saccadic horizontal eye movement manipulation (200 ms per movement in EM condition), followed by 500-msec fixation point and then prime displayed for 32msec, which was followed immediately by the target. The participant was instructed to read the target as soon as possible and the reaction time was recorded by the compuer. While absolute primeability index for strong (or weak) association was calculated as (RT related – RT unrelated) for strong primes or weak primes, respectively, relative primeability index was calculated as (RT related – RT unrelated)/RT unrelated. The dependent variables were correct (%), error (%), and primeability index for each of strong prime and weak prime. Change of primeability of weak associations for negative stimuli as opposed to that of strong ones under different EM conditions and different block order was compared.
Results: For absolute primeability index, the 2 (EM condition) × 2 (order) × 2 (strength of priming) ANOVA showed that only the interaction effect of EM × strength of semantic association was significant ( F(1, 56) = 5.61, p < .02; partial ω2 = .067), indicating that weak priming significantly exceeded strong priming after EM ( F(1, 56) = 18.01, p < .001; partial ω2 = .210), while the opposite was true after non-EM ( F(1, 56) = 25.86, p < .0001; partial ω2 = .280). Further, for weak prime, the priming effect was stronger after EM than after non-EM ( F(1, 56) = 74.62, p < .0001; partial ω2 = .535). For relative primeability index, it also revealed that only the interaction effect of EM × strength of semantic association was significant ( F(1, 56) = 6.09, p < .02; partial ω2 = .074), and tests of simple main effects showed similar patterns as those of absolute primeability index.
Conclusions: 8s EMs was associated with change of strength of semantic associations. While EM didn’t weaken the primeability for the strong association, EM did enhance the primeability for weak associations. Further, weak priming significantly exceeded strong priming after EM, while the opposite was true after non-EM. The results echoed REM-sleep dependent memory reprocessing model, suggesting that EM in EMDR might reflect a shift in associative memory systems by activating different strength of associations of negative semantic nodes for different semantically related words. Given that order of relatedness didn’t play a role and previous research showed that EM resulted in decreased vividness and emotionality and generated greater amount of associations for negative stimuli as well, the implications of the present results from theoretical and psychotherapy point of views and future research possibilities are discussed.
Format
Conference
Language
English
Original Work Citation
Chang, S.-H. (2005). Mechanism of EM in EMDR: Change strength of semantic associations. Presentation at the 113th American Psychological Association Annual Convention, Washington, DC. doi:10.1037/e525012006-001
Citation
“Mechanism of EM in EMDR: Change strength of semantic associations,” Francine Shapiro Library, accessed May 17, 2024, https://francineshapirolibrary.omeka.net/items/show/18694.