How EMDR Works: The Science Behind Eye Movement Therapy

Multiple studies have proven that “memory reconsolidation” is a common through line in EMDR. 

While current research offers many different theories to how exactly EMDR works, they all argue that memory reconsolidating functions as a core mechanism. 

These studies include attention to EMDR’s possible relation towards slow-wave sleep cycles (SWS), REM sleep and memory processing, working memory taxation, and brain connectivity.

Working memory is our ability to hold the information that is brought into our focus, and to change it. For example, when we have a conversation with someone, we are not just responding to words, but on the entire thread of a conversation we have just followed, and the context of who we are talking to, roles, etc. We depend on our working memory to make immediate decisions on our goals, relationships, and challenges, with the information presented to us. 

Our working memory also has limited capacity and duration. 

Some studies have shown evidence that holding a traumatic image in mind while simultaneously participating in BLS bilateral stimulation, for example with rapid eye movements or tapping, results in our working memory taxation and “overload.” As a result, the image becomes “degraded,” stored as something less vivid and less emotionally charged.

By utilizing dual attention and focusing on both BLS and negative experiences, negative experiences have to compete for space within working memory and lose their charge as a result (Andrade et al., 1997; Kavanagh et al., 2001).. Negative experiences lose their feelings of vitality, immediacy, the feeling of “right-now.” Over time, the memory is now able to stored and reconsolidated, with less ability to feel immediate and more of a feeling of generality (Nader, K., Schafe, G. E., & LeDoux, J. E., 2000).

These studies have also shown that eye-movement sets show a measurable orienting response, the “what is this?” reaction that we often have with novel experiences, paired with a de-arousal in heart rate and stress responses. 

With repeated EMDR, the nervous system’s orienting response to the BLS habituates over time, easing negative thoughts, feelings, and beliefs out of an“absolute,” defensive posture. 

We are also held in present-moment awareness, grounded in the room, aware the memory is being recalled rather than relived.

Episodic memory is the memory we associate with personal events, our deeply “first-person perspective,” lived experiences. 

Semantic memory is the general knowledge we have of things, i.e. the name of a city, the shape of a phone, types of foods. Information that is categorized and generally feels context-less and non-time specific.

In negative experiences, memories become maladaptively tied to our “I” experience of life in episodic memories, and our brain’s fear-response system (amygdala).

Shapiro’s original idea was that bilateral eye movements mimic REM sleep’s role in emotional memory processing. REM (rapid eye movement) sleep) is theorized to be how our brain replays emotional experiences while our stress response is suppressed. Stickgold (2002) gave this more grounding with their research and brain imaging, though REM sleep has actually been theorized to be a more important function post-EMDR, rather than during.

Additional research and brain imaging recorded during EMDR BLS has shown “delta-range synchronization,” slow-wave sleep brain waves, occurring during BLS, rather than brain waves commonly associated REM sleep cycles. 

In general sleep science, slow-wave sleep has been shown to function as a “reset” of our brain’s links between memories and nervous system responses. This is different than REM’s memory consolidation responsibility. 

Researchers concluded that EMDR’s desensitization phase reduces the fear-memory synapse, compressing what slow-wave sleep normally does across 3–5 nightly cycles into roughly 25–30 stimulation sets in a single session (Pagani et al., 2017).

Slow-wave sleep states are theorized to transfer episodic memory toward semantic memory, before that material gets replayed and consolidated during REM sleep. 

In REM sleep, nervous system responses are suppressed, while memories get appropriately transferred to semantic memory regions (neocortex), where they are replayed and integrated into broader associative semantic networks. 

Similarly in EMDR, complex memories that have been maladaptively coupled between the “I” experience of our episodic memories, and our brain’s fear-response system, ultimately get decoupled. This converts raw, emotionally-loaded episodic memories into more integrated, “storied,” and meaning-laden ones. A traumatic memory that starts as fragmented, sensory, and in the present-tense, ie ”it’s happening to me right now,” gradually becoming narrative, past-tense, and integrated, ie ”this happened to me, and it’s part of my history,” which is often the feeling described as people reprocess memories through EMDR reprocessing. With memories desensitized from fear-responses in Slow-Wave like states, REM sleep is able to adequately function to store these memories in the semantic network later in the day.

One study found that as PTSD symptoms remitted following EMDR treatment, the density of rapid eye movements during patients’ natural REM sleep increased. The study also found this increase was positively correlated with the degree of symptom improvement, with greater REM-density increase tracked with greater symptom reduction (Rousseau et al., 2021).

Slow-Wave Sleep disconnects memories from the brain’s synapses attached to the fear response system, and begins the process to move them into semantic memory networks. REM sleep matters because it’s theorized to be where the brain naturally takes the emotional charge off memories, preserving their content, and settling them into semantic memory. 

The underlying premise with REM, in sleep science specifically, is that the brain replays recent emotional experience while cortical noradrenaline is suppressed. Memories get reactivated and, theoretically, reprocessed under emotionally “cooler” neurochemical conditions than when it was originally encoded, which is thought to be part of how ordinary emotional memories naturally lose their sting over time through normal dreaming.

Typically when we think of memories, we consider them within the same frame that we always do. Memories can become well-trodden paths within our brain and in resulting nervous system responses. Memory reconsolidation theory says that briefly recalling a memory makes it become chemically labile and subject to change. Animal studies have shown this window can last just a few minutes to a few hours, before it re-stabilizes. During that window, new or incompatible information can be integrated before the memory locks back in. This is mechanistically distinct from simply forming a new memory: it is reopening and rewriting one that already exists. (Nader; Schiller; Ecker, Ticic & Hulley)

EMDR’s structure maps onto this closely: the target memory is activated (retrieval → destabilization), new elements are introduced during the labile window (bilateral stimulation, cognitive load, present-moment awareness), and the memory is allowed to re-stabilize with that new material woven in. Reconsolidation theory is compatible with both the working-memory findings above (dual-attention as one way of modifying a reactivated memory) and Pagani’s SWS model (slow-wave sleep as one of the brain’s natural reconsolidation windows, which EMDR may induce artificially while awake).

EMDR is an evidence-based treatment for PTSD, recognized by organizations like the WHO and APA as a leading first-line therapy. Its underlying mechanisms in working memory, sleep science, and memory reconsolidation, are still being actively researched and refined. EMDR’s core tenets remain theoretical, but it’s also generated so much of the enthusiasm around EMDR research as a therapy that works clinically, and is generating more ground for why it is so effective.

Key References

• Andrade, J., Kavanagh, D., & Baddeley, A. (1997). Eye-movements and visual imagery: A working memory approach to the treatment of post-traumatic stress disorder.

• Baek, J. et al. (2019). Optogenetic study of the superior colliculus–mediodorsal thalamus–amygdala pathway in fear suppression (mouse model).

• Christman, S.D., Garvey, K.J., Propper, R.E., & Phaneuf, K.A. (2003). Bilateral eye movements enhance the retrieval of episodic memories.

• Davidson, P.R., & Parker, K.C.H. — Meta-analysis of eye-movement-specific contribution to EMDR outcomes.

• Ecker, B., Ticic, R., & Hulley, L. — Unlocking the Emotional Brain (reconsolidation theory and clinical application).

• Kavanagh, D.J. et al. (2001). Further evidence for the interference effect of eye movements on emotional imagery and memory.

• Kragel, P.A. et al. (2021). Colliculus–pulvinar–amygdala pathway for negative emotion (human data).

• Lansing, K. et al. (2005). High-resolution brain SPECT imaging and EMDR in police officers with PTSD.

• Laugharne, J. et al. (2016). Amygdala volumetric change following psychotherapy for PTSD.

• Meshreky, M., & Lewis, P.A. (2025). Do eye movements in REM sleep play a role in overnight emotional processing? Neuropsychologia, 215, 109169.

• Nader, K.; Schiller, D. — Memory reconsolidation and fear-erasure literature.

• Nardo, D. et al. (2010). Gray matter density in limbic/paralimbic cortices and EMDR treatment outcome.

• Pagani, M. et al. (2017). Neurobiological correlates of EMDR: EEG evidence for slow-wave-sleep-like synchronization during bilateral stimulation.

• Rousseau, P. F., Vallat, R., Coste, O., Cadis, H., Nicolas, F., Trousselard, M., Ruby, P., & Khalfa, S. (2021). Sleep parameters improvement in PTSD soldiers after symptoms remission. Scientific Reports, 11, 8873. https://doi.org/10.1038/s41598-021-88337-x

• Schiller, D. et al. — Fear-erasure and reconsolidation-window studies.

• Schubert, S., Lee, C.W., & Drummond, P.D. (2011). The efficacy and psychophysiological correlates of dual-attention tasks in EMDR.

• Shapiro, F. — Adaptive Information Processing (AIP) model; EMDR standard protocols.

• Stickgold, R. (2002). EMDR: A putative neurobiological mechanism of action.