Working Memory Theory of EMDR: Why Eye Movements May Help

Working memory theory holds that EMDR’s eye movements compete with a distressing memory for the same limited mental workspace, leaving the memory feeling less vivid and less emotionally charged. It’s the best-supported explanation for why bilateral stimulation works, though researchers still disagree on exactly how much it adds and whether it explains the whole effect.

What is working memory, and why would eye movements tax it?

Working memory is the mental workspace holding a small amount of information active at once: a phone number before you dial it, an image while you’re describing it. It’s genuinely limited, and asking it to do two demanding things at once means both suffer.

That’s the core idea behind working memory theory. Holding a distressing memory in mind is one demanding task. Deliberately tracking a moving target with your eyes is a second one.

Run them together, and they compete for the same limited space. The memory that comes out of that competition tends to feel less vivid and less emotionally loaded than it did going in.

A 2012 review in the Journal of Experimental Psychopathology by van den Hout and Engelhard is the clearest modern statement of this account. It’s the theory most researchers reach for first when asked why EMDR’s eye movements might do anything at all.

Where did this idea actually come from?

Not from EMDR’s own research program, which is worth knowing. Psychologist Francine Shapiro discovered the eye-movement effect by accident in 1987, noticing a disturbing thought lose its charge while her eyes moved rapidly back and forth. She built a therapy around what she’d stumbled onto, but her early work didn’t explain why it worked.

That explanation came a decade later, from cognitive psychologists studying memory and mental imagery, not trauma. In an influential 1997 study in the British Journal of Clinical Psychology, researcher Jackie Andrade and colleagues tested a specific hypothesis. One of her co-authors was Alan Baddeley, whose own multicomponent model of working memory is a standard framework in the field. Eye movements, they proposed, compete with a mental image for space in the visuospatial sketchpad, the part of working memory that holds and manipulates visual and spatial information.

Their volunteers pictured neutral and unpleasant images while making eye movements, performing an unrelated spatial tapping task, or doing nothing else. People in both distraction conditions rated their images as less vivid and less emotional afterward than people who simply held the image in mind with no distraction. Neither task involved a therapist, a trauma memory, or anything resembling clinical EMDR. It was a lab test of a cognitive mechanism, run on ordinary volunteers, years before that mechanism got widely credited as a reason Shapiro’s technique might work.

Does it matter which part of working memory gets taxed?

This is the part most quick summaries skip, and it explains a detail later research had to work around. Baddeley’s model doesn’t treat working memory as one single pool. It splits into sub-systems: a visuospatial sketchpad for visual and spatial information, a phonological loop for verbal and auditory information, and a central executive that coordinates both and has its own limited capacity.

Andrade and colleagues’ original hypothesis was specific to the visuospatial sketchpad. A visual image, their reasoning went, should be most vulnerable to a competing visual or spatial task, not to just any distraction. That’s a narrower, more testable claim than “distraction helps,” which is part of why the study still gets cited as the account’s foundation rather than a general observation about busy minds.

The theory has broadened since. Van den Hout and Engelhard’s 2012 review frames the effect more generally, as a demand on limited working-memory resources overall. That broader version helps explain why alternating tones and taps, not just eye movements, also seem to work: they don’t obviously load the visuospatial sketchpad the way tracking a moving target does, but they still demand attention and effort.

Does the evidence actually back this up?

Some of the most convincing support treats “how demanding is the task” as a dial, not a switch, and checks whether turning it up changes the result. A 2015 study in Frontiers in Psychiatry tested two paces of eye movement, a faster 1.2 Hz and a slower 0.8 Hz, against no eye movements at all, while people held a disturbing memory in mind. The faster, more effortful pace came out ahead on every measure that counted: a duller image, a lighter emotional charge, and more trouble pulling the memory back up afterward. That’s the exact pattern working memory theory predicts, more taxing, more effect.

Our practical guide to eye movement speed covers what that means for pacing your own practice.

A related 2016 study in the Journal of EMDR Practice and Research found the effect isn’t just a momentary illusion. People who made eye movements while concentrating on a negative memory reported fewer intrusive memories of it in the days that followed, compared to people who held the same memory in mind without eye movements.

A 2008 study in Behaviour Research and Therapy by Gunter and Bodner adds a different kind of support. A popular alternative theory claims eye movements work by forcing the brain’s two hemispheres to communicate, which would mean the direction of the movement should matter. It didn’t: moving the eyes up and down softened a memory’s vividness and emotional charge just about as much as the standard side-to-side sweep. Effort of tracking seemed to be doing the work, not direction, exactly what a working-memory account predicts and a hemisphere-syncing account doesn’t.

Do the bigger reviews of the evidence agree?

Less than you’d hope, and it’s worth saying plainly rather than picking a side. An older, frequently cited 2001 meta-analysis in the Journal of Consulting and Clinical Psychology, by Davidson and Parker, stripped the eye movements out of EMDR and compared what was left against the full procedure. The eye movements, by that comparison, added nothing measurable.

A larger review reached the opposite conclusion. Lee and Cuijpers’ 2013 meta-analysis in the Journal of Behavior Therapy and Experimental Psychiatry combined 15 clinical trials with 11 laboratory studies and found the eye movements did add a real, moderate reduction in distress, an effect that came through even more strongly under controlled lab conditions than in messier clinical trials.

Two careful reviews, opposite headline conclusions, built from much of the same underlying research. A 2018 systematic review in Frontiers in Psychology surveyed working memory theory alongside competing explanations, including an older “orienting response” idea and a proposed link to REM sleep, and concluded that no single theory yet accounts for every finding on its own. Working memory theory is the leading account here, not a settled one. Our deeper look at what happens in the brain during EMDR walks through those competing theories side by side.

If it’s really about tasking working memory, should tapping and tones work too?

That’s the theory’s most testable prediction, and it’s a fair one to hold it to. If the real mechanism is generic demand on limited attention rather than something specific to eye movements, other effortful, alternating tasks should produce a similar effect.

The evidence leans toward yes, with real nuance. A 2013 study in the Journal of Behavior Therapy and Experimental Psychiatry compared eye movements to alternating tones directly and found a trend toward eye movements reducing a memory’s emotional charge somewhat more, not a case where tones did nothing at all. On the body’s side of the response, a 2025 study in BJPsych Open found visual and tactile bilateral stimulation produced similarly calming shifts in brain activity and heart-rate variability in the same people.

Does tapping work as well as eye movements? covers this split in depth. The short version: modality seems to matter at the margins, not at the theory’s foundation.

What does this mean if you practice bilateral stimulation yourself?

Here’s the practical upshot of a mechanism story like this one: nothing about working memory taxation obviously requires a therapist in the room. Holding a mildly stressful thought in mind while your eyes track a moving target, or while alternating tones play through headphones, taxes the same limited resource whether or not anyone else is watching.

That’s a reasonable basis for using bilateral stimulation as a wellness practice for everyday stress: a looping worry, a tense conversation replaying in your head. It isn’t proof that a self-guided version reproduces what a full course of therapist-led EMDR does for PTSD or serious trauma. The studies behind this theory tested cognitive mechanisms in volunteers and lab settings, not self-administered trauma treatment. Our answer on why EMDR uses eye movements covers that same boundary from a different angle.

EmEase, a self-guided EMDR app, builds on exactly this research: a moving on-screen target for visual bilateral stimulation, and alternating audio tones if you’d rather close your eyes, both adjustable at app.emease.com. Use it for everyday stress and looping thoughts, and bring in a licensed professional for anything heavier.

What the research doesn’t yet settle

A few honest limits belong here. Most of the studies behind this account, Andrade’s original work included, tested healthy volunteers imagining mildly unpleasant scenes in a lab, not people with diagnosed PTSD working through a defining trauma. The two big meta-analyses above disagree on how much eye movements add on top of plain recall, and the 2018 review concluded the field hasn’t settled on one mechanism to the exclusion of the others.

None of this research tested a self-guided app directly, either. It tested the cognitive and physiological pieces a self-guided practice would draw on, not the practice itself.

The bottom line

Working memory theory is the best-supported explanation for why EMDR’s eye movements might do something beyond simply revisiting a memory: two demanding tasks compete for the same limited mental space, and the memory comes out less vivid and less charged. It traces back to a 1997 lab study on ordinary volunteers, not a trauma trial, and it has held up reasonably well since, including in dose-response tests where a more demanding pace produced a bigger effect.

What it hasn’t done is settle the field’s oldest argument. Two respected meta-analyses reach opposite conclusions about how much eye movements add, and a working theory, however well-supported, isn’t the same as a proven mechanism. Treat it the way the researchers behind it do: a genuinely useful explanation, still being tested, not a finished answer.