A Therapy Born on a Park Walk
One spring day in 1987, Francine Shapiro, a psychology doctoral student in California, was walking through a park. A persistent distressing thought wouldn't leave her — until suddenly it did. She noticed she had been unconsciously moving her eyes rapidly side to side. When she deliberately recalled the thought, its emotional charge had weakened.
Shapiro repeated the procedure with friends and students, published the first RCT in Journal of Traumatic Stress in 1989, and codified the 8-phase protocol in Eye Movement Desensitization and Reprocessing (1995, 3rd ed. 2018). EMDR was adopted for Vietnam veterans' PTSD and spread rapidly.
That the technique was discovered on a walk has remained a criticism for 30 years: 'a method born of anecdote, not science.' Yet the RCT evidence accumulated since has outweighed the skepticism.
The 8-Phase Standard Protocol
EMDR is not improvised eye-moving but a manualized sequence (Shapiro 1995/2018):
| Phase | Name | Goal |
|---|---|---|
| 1 | History taking | Map trauma, assess suitability |
| 2 | Preparation | Build alliance, install 'safe place' resource |
| 3 | Assessment | Identify target memory's image, negative cognition, emotion, body sensation |
| 4 | Desensitization | Bilateral stimulation (eye movements) to reduce intensity |
| 5 | Installation | Strengthen positive cognition ('I am safe') |
| 6 | Body scan | Process residual somatic sensations |
| 7 | Closure | Stabilize session, self-soothing techniques |
| 8 | Reevaluation | Check change, choose next target |
Sessions typically last 60–90 minutes; simple PTSD often resolves in 8–12 sessions, complex trauma requires more.
Adaptive Information Processing — Shapiro's Model
Shapiro proposed the Adaptive Information Processing (AIP) model:
- The brain has a natural system to integrate and generalize experience.
- In trauma, this system stalls; memory is stored as a maladaptive network with fragmented sensory, emotional, and bodily components.
- Bilateral stimulation reactivates this frozen network into normal processing flow.
AIP is intuitive and clinically useful but not directly validated neuroscientifically. How a 'maladaptive network' is neurally represented, or how bilateral stimulation drives integration, remains unclear.
The Working Memory Hypothesis — A More Testable Alternative
The currently dominant explanation is the working memory hypothesis (Andrade 1997, van den Hout 2011):
- Working memory's visuospatial capacity is limited.
- Recalling a traumatic memory while performing eye movements (a visuomotor load) dilutes working memory resources, reducing vividness and emotional intensity.
- The diluted memory is re-encoded with lower emotional weight.
This cleanly explains why tapping or alternating tones also work — what matters is working memory load, not bilateral stimulation per se. Some studies show counting backwards or Tetris produce similar effects.
The Weight of Evidence — WHO, APA, Cochrane
The mechanism debate is separate from clinical evidence:
- WHO 2013 PTSD guidelines: EMDR and trauma-focused CBT recommended as first-line psychotherapies for adults, adolescents, and children.
- APA 2017 PTSD guideline: conditional recommendation for EMDR.
- Bisson 2013 Cochrane review (70 RCTs, n=4,761): TF-CBT and EMDR both produce clinically meaningful PTSD reduction with no significant difference between them.
- Equivalence: Bisson 2007, Seidler 2006 meta-analyses confirm EMDR ≈ TF-CBT.
Do Eye Movements Really Add Anything? Dismantling Research
The most heated debate: does EMDR work without eye movements? Critics argued EMDR is essentially exposure therapy plus a 'working ritual.'
Lee & Cuijpers 2013 dismantling meta-analysis pooled 26 studies comparing EMDR with and without eye movements. Conclusion: a small but statistically significant added effect from eye movements, consistent across clinical and laboratory settings.
This cuts both ways:
- The strongest skepticism ('eye movements are useless ritual') is rejected.
- But the 'small' effect implies most of EMDR's benefit comes from shared factors: exposure, image reprocessing, therapeutic alliance.
Shapiro herself clarified: 'It's called eye movement, but it's one form of bilateral stimulation.' Modern practice uses alternating taps and alternating tones interchangeably.
EMDR in Korea
- Korean EMDR Association (KEMDR): founded 2007, manages EMDRIA-certified training.
- Ministry of Health PTSD guideline: EMDR included among recommended psychotherapies.
- Disaster mental health: applied to Sewol ferry (2014) families and survivors, and to MERS/COVID-19 frontline medical workers via the National Center for Mental Health and regional trauma centers.
- Insurance: partial coverage when delivered by psychiatrists; usually out-of-pocket when delivered by clinical/counseling psychologists.
Cost is typically 80,000–200,000 KRW per session; 8–12 sessions for simple PTSD makes affordability a real factor. Public trauma centers offer free/low-cost programs; workplace trauma may qualify for workers' compensation.
Limits and Scope
EMDR is not a panacea:
- Complex PTSD: less robust evidence, longer course, extended preparation (phase 2).
- Ongoing threat: not appropriate while domestic violence, war, or active abuse continues — safety first.
- Dissociative or psychotic disorders: require careful expert assessment.
- 'Self-EMDR' risks: YouTube 'follow the dot' videos lack resource installation, assessment, body scan, and closure. They can trigger exposure without processing — potentially worsening symptoms.
Conclusion — Settled and Unsettled
Settled: EMDR works for PTSD; it is first-line, comparable to TF-CBT; eye movements add a small but real effect.
Unsettled: why it works. AIP is a useful metaphor but lacks neuroscientific verification. Working memory is the most testable alternative but may not explain the whole effect.
So what to recommend? WHO and Korean Ministry of Health are clear: EMDR and TF-CBT delivered by qualified clinicians are first-line PTSD treatments. Waiting for the mechanism to be fully solved is no different from ignoring a tool of proven effect.
