An Efficient System for Eye Movement Desensitization and Reprocessing (EMDR) Therapy: Pilot Study (Preprint)
BACKGROUND EMDR is a psychological therapy designed to treat emotional distress caused by a traumatic event from the past, most frequently in the PTSD treatment. We implemented a system based on video, tactile and audio sensors which includes an artificial intelligence chatbot that can replace the trained clinician in the EMDR procedure, thus making the system capable to act autonomously. OBJECTIVE This pilot study aims to develop an EMDR virtual assistant capable of assisting a user in performing an efficient therapy with a minimal intervention of a trained clinician. METHODS An EMDR virtual assistant was created that include audio, video and tactile sensors. A total of 31 participants were included in the pilot study, 14 males and 17 females, with ages between 19 and 33, (Mage = 26.2, SD = 4.21). Participants were invited to remember a traumatic event from their past that caused them discomfort, and they failed to overcome it. After accessing the traumatic memory, primary outcome measures (IES-R and STAI) and secondary outcome measures (VOC and SUD) instruments were applied in the pre-test condition. The designed intervention through application consisted of four phases of bilateral stimulation (visual, auditive, and sensorial). After the discussion with a specialist, and after the completion of the instruments on the pre-test assessment phase, the participants self-administered the intervention through the application and then completed again the instruments in the post-test phase. RESULTS The results showed that there was a significant difference in scores for the pre-test condition (M = 39.8, SD = 14.07) and post-test (M = 13.7, SD = 5.7) measured with IES-R scale; t(30) = 10.5, p < 0.001, and a large effect size (d = 1.89). Also, there are significant differences in scores for the pre-test condition (M = 56.7, SD = 8.3) and post-test (M = 33.8, SD = 4.6) regarding anxiety scores (STAI); t(30) = 11.7, p < 0.001, and a large effect size (d = 2,11) of intervention in reducing anxiety associated with the traumatic event. Related to the secondary outcome measure there was a significant pre- to post-intervention reduction in scores for both VOC; t(30) = 46.8, p < 0.00, and SUD; t(30) = 75.9, p < 0.001. It was shown that the higher scores decreased significantly at the end of the intervention. CONCLUSIONS Our results proved the efficiency of the virtual assistant system for EMDR in reducing anxiety, distress, and negative cognitions and emotions associated with traumatic memories. These findings are filling a gap in the existing literature regarding the efficiency and use of similar systems and could represent a starting point for developing applications and future controlled trials and protocols that can be used in similar cases.