C4 · Lab · Psyche · A series · Interactive, in your browser
EMDR
What EMDR is, what its first study reported, and what the trials and guidelines say about the eye movements at its centre.
This isn’t advice about your own care. It explains a therapy and what the trials found.
What EMDR is
EMDR was developed in the late 1980s by Francine Shapiro. The letters stand for eye movement desensitization and reprocessing [Cuijpers 2020 (meta-analysis)].
NICE describes it as a trauma-focused therapy: a structured protocol, a validated manual and repeated sets of bilateral stimulation aimed at specific memories [NICE 2018 (guideline)].
What a session looks like
In the room, the person holds a memory in mind while the therapist’s hand moves back and forth across their view, or a tone sounds in one ear and then the other, or taps alternate. A set lasts a few seconds; then the person reports what came up and how strong the distress is, and the next set begins [NICE 2018 (guideline)].
NICE sets out what a course should do: a phased approach with psychoeducation about trauma reactions, work on target memories and positive beliefs, and repeated sets continued until the memories are no longer distressing. It says a course is typically 8 to 12 sessions, more after multiple traumas, delivered by trained practitioners with supervision [NICE 2018 (guideline)].
The walk in the park
The story EMDR tells about itself begins in May 1987, on a walk in a park. Shapiro said a disturbing thought came up and then went, and that she noticed her eyes moving quickly from side to side as it did. When she called the thought back, it had lost its charge [Rosen 2023 (paper)].
She tried the effect on volunteers and clients, and in 1989 published the study that launched the treatment [Shapiro 1989 (paper)].
The story is disputed. People cannot see their own saccades, Rosen writes, and Shapiro had already written about eye-movement patterns before the walk; a 2026 paper calls the park account false [Rosen 2023 (paper); Rosen & Pankratz 2026 (paper)].
What the story does not change is the study it led to. The first paper reports 22 people, a single session, and a follow-up at one and three months; the abstract records the measures, not the walk [Shapiro 1989 (paper)].
1987
A walk in a park
Shapiro’s account of the origin
1989
The first paper
22 people, one session
2001
The first meta-analysis
The eye movements add nothing
2013
The re-analysis
They add a little
2018
NICE NG116
EMDR recommended for adults
2020
76 trials pooled
Effective; the studies are weak
The first study
Shapiro’s paper reports 22 people whose symptoms followed traumatic events: war, childhood sexual abuse, sexual or physical assault, or emotional abuse [Shapiro 1989 (paper)].
After a single session, anxiety fell, a rating of a positive self-statement rose, and the presenting complaints eased. The paper says the changes held at one and three months [Shapiro 1989 (paper)].
The abstract does not describe a comparison group. Davidson and Parker, looking back at the early literature, call the paper seminal and say it lacked standardised outcome measures [Davidson & Parker 2001 (meta-analysis)].
What made the procedure distinctive was the eye movements. Shapiro described them as the crucial component of it [Davidson & Parker 2001 (meta-analysis)].
What the trials found
A 2020 meta-analysis pooled 76 trials of EMDR for any mental health problem; 62% of them studied PTSD [Cuijpers 2020 (meta-analysis)].
Against control conditions the effect on PTSD was g = 0.93 (95% CI 0.67 to 1.18), with high heterogeneity (I² = 72%). Four of the 27 studies assessed for risk of bias were at low risk, and the authors found signs of publication bias [Cuijpers 2020 (meta-analysis)].
EMDR did better than other therapies, g = 0.36 (95% CI 0.14 to 0.57), but not in the low-risk trials. The authors conclude that the studies are too weak for definite conclusions, though EMDR may work for PTSD in the short term [Cuijpers 2020 (meta-analysis)].
For other conditions the evidence was thin: a few trials in phobias and test anxiety at high risk of bias, and too few trials to pool for the rest [Cuijpers 2020 (meta-analysis)].
Do the eye movements matter?
Davidson and Parker pooled 34 studies in 2001. They found EMDR better than no treatment and about as effective as exposure, with no significant advantage for the eye movements themselves [Davidson & Parker 2001 (meta-analysis)].
Lee and Cuijpers argued that the earlier analysis had methodological problems that could hide an effect, and pooled again: 15 clinical comparisons of EMDR with and without eye movements, and 11 laboratory comparisons, 849 people in all [Lee & Cuijpers 2013 (meta-analysis)].
The eye movements added d = 0.41 in the clinical studies and d = 0.74 in the laboratory. Treatment fidelity moderated the clinical result [Lee & Cuijpers 2013 (meta-analysis)].
A working-memory account gives a reason the eye movements might matter without being special. In one experiment, people recalled a distressing image while doing subtraction; the image became less vivid and less emotional than with exposure alone, and the size of the effect depended on how hard the task was [Engelhard 2011 (paper)].
On that account, any task that loads working memory would do something similar. NICE allows taps or tones instead of eye movements [NICE 2018 (guideline)].
EMDR against control
PTSD, confidence interval
EMDR against other therapies
confidence interval
Eye movements, clinical
point estimate
Eye movements, laboratory
point estimate
Effect size, g. Bars are intervals; points have none in the abstract.
Where the guidelines put it
NICE offers EMDR to adults with PTSD more than 3 months after a non-combat trauma, and says to consider it at 1 to 3 months if the person prefers it [NICE 2018 (guideline)].
For children and young people aged 7 to 17, NICE recommends trauma-focused CBT first and considers EMDR only if they do not respond to or engage with it [NICE 2018 (guideline)].
WHO’s 2013 guidelines put EMDR among the treatments to consider for adults with PTSD, alongside trauma-focused CBT and stress management [WHO 2013 (guideline)].
Limits
This isn’t advice about your own care either. It explains a therapy and what the trials found.
The meta-analysis names the problems: high heterogeneity, few trials at low risk of bias, and signs of publication bias [Cuijpers 2020 (meta-analysis)].
The mechanism is unsettled. The two meta-analyses of the eye movements reached different answers, and the working-memory experiment is a laboratory study, not a trial of therapy [Davidson & Parker 2001 (meta-analysis); Lee & Cuijpers 2013 (meta-analysis); Engelhard 2011 (paper)].
This chapter stays with the short-term comparisons those papers report. It does not cover complex PTSD, the details of trauma-focused CBT, or what happens after a course ends [Cuijpers 2020 (meta-analysis)].
Sources
Cuijpers 2020 (meta-analysis) Eye movement desensitization and reprocessing for mental health problems: a systematic review and meta-analysis. Cognitive Behaviour Therapy 49(3), pages 165 to 180. PubMed record, abstract only, read 2026-11-20.
The 76 trials, the effect against control conditions, the comparison with other therapies, risk of bias, publication bias and the authors’ conclusion.
- Read: abstract
NICE 2018 (guideline) Post-traumatic stress disorder. NICE guideline NG116, recommendations 1.6.9 to 1.6.12 and 1.6.20 to 1.6.27. Recommendations, full text, read 2026-11-20.
What EMDR should do in a course, and where it sits for adults and for children and young people.
- Read: full text
Rosen 2023 (paper) Revisiting the origins of EMDR. Journal of Contemporary Psychotherapy. Open access, full text, read 2026-11-20.
Shapiro’s park account, the research on why people cannot perceive their own saccades, and the earlier writings on eye-movement patterns.
- Read: full text
Shapiro 1989 (paper) Efficacy of the eye movement desensitization procedure in the treatment of traumatic memories. Journal of Traumatic Stress 2(2), pages 199 to 223. Publisher record, abstract only, read 2026-11-20.
The 22 participants, the single session, the three measures and the follow-up at one and three months; no claim beyond the abstract.
- Read: abstract
Rosen & Pankratz 2026 (paper) A “Decisive Moment” confirms the origins of EMDR and draws attention to other claims. Journal of Contemporary Psychotherapy. Open access, full text, read 2026-11-20.
The 1985 photograph and the authors’ claim that the park story is false.
- Read: full text
Davidson & Parker 2001 (meta-analysis) Eye movement desensitization and reprocessing (EMDR): a meta-analysis. Journal of Consulting and Clinical Psychology 69(2), pages 305 to 316. Journal full text, full text, read 2026-11-20.
The 34 studies, the comparison with no treatment and with exposure, the finding on the eye movements, and the criticism of the early measures.
- Read: full text
Lee & Cuijpers 2013 (meta-analysis) A meta-analysis of the contribution of eye movements in processing emotional memories. Journal of Behavior Therapy and Experimental Psychiatry 44(2), pages 231 to 239. PubMed record, abstract only, read 2026-11-20.
The 15 clinical and 11 laboratory comparisons, the two effect sizes, treatment fidelity as a moderator and the reply to the earlier meta-analysis.
- Read: abstract
Engelhard 2011 (paper) Taxing working memory reduces vividness and emotional intensity of images about the Queen’s Day tragedy. Journal of Behavior Therapy and Experimental Psychiatry 42(1), pages 32 to 37. PubMed record, abstract only, read 2026-11-20.
The four conditions, the subtraction tasks, and the finding that taxing working memory reduced vividness and emotional intensity.
- Read: abstract
WHO 2013 (guideline) Guidelines for the management of conditions specifically related to stress. Geneva: World Health Organization. WHO PDF, full text, read 2026-11-20.
The recommendation for adults with PTSD, where EMDR is one of the treatments to consider.
- Read: full text
Each fact is paraphrased from the page named, read on the date shown. “Abstract” means I read the abstract only and say no more than it does. Where a sentence is my own reading, it is marked as my view.
