Reprocessing traumatic memories: EMDR, hypnosis and ketamine compared

The Short Version

Every trauma treatment in these conversations aims at the same thing: keep the memory but loosen the fear attached to it. As Huberman puts it, "you never forget the traumatic experience what you do is you remove the emotional load" [2]. The tools get there differently. EMDR uses side-to-side eye movements that Huberman says quiet the amygdala while a person recounts the event [1][2]. Clinical hypnosis, as Stanford's David Spiegel practices it, uses a focused, dissociated-like state to let people revisit the memory with their body kept calm and to see their own strength in it [3]. Ketamine uses chemical dissociation, letting people recount trauma while feeling something different, and may also rewire mood circuits over weeks [1][4]. The experts disagree about what makes each one work. Spiegel doubts EMDR's eye movements add anything, Huberman thinks EMDR often skips the crucial relearning step, and even Huberman's colleagues doubt ketamine is the ultimate trauma treatment [1][3]. Each option has real evidence and real limits, and each should be done with a trained clinician.

Table of contents

What They Said

The shared model: extinguish, then replace

Huberman's episode on fear gives a framework for comparing every tool [1]. He describes the amygdala as part of a generic threat reflex. It speeds the heart, sharpens vigilance and mobilizes energy, and it is deliberately generic, so it can attach to almost anything [1]. Fear attaches fast: "We get one trial learning to negative experiences," he says, whereas positive learning like languages or music takes many repetitions [1]. At the cellular level, intense events activate NMDA receptors and strengthen synapses through long-term potentiation, so later even the thought of something can set off the threat reflex [1].

The brake is the prefrontal cortex, which exerts top-down control by attaching narrative, meaning and purpose to the otherwise generic response [1]. "There's no negotiating what fear feels like. There's only negotiating what it means" [1].

From that biology Huberman draws a two-step recipe [1]:

1. Extinction. Every clinician he consulted agreed that a detailed recounting of the traumatic event is essential. The first retelling can trigger an anxiety response as big as the event itself, but the response shrinks with each retelling until the memory becomes "what is essentially a boring bad story" [1]. Prolonged exposure therapy, cognitive processing therapy and cognitive behavioral therapy all work this way [1].

2. Replacement. "It is not going to work to simply extinguish a fear," he says [1]. The second step is building a new narrative with positive, rewarding associations and linking it back to the original trauma. His example is a child who stops biking, playing sports and seeing friends after a car accident on the way to soccer practice. The child needs to experience enjoying the bike ride to practice again, holding both in mind [1]. This works because prefrontal inputs to the threat circuit are inhibitory, and the fear circuit connects to the dopamine reward system, so new positive associations can be wired in [1]. Order matters: extinction first, then relearning. "Narrative should not be undervalued as a tool for relieving fear and trauma" [1].

He notes that SSRIs, benzodiazepines, antipsychotics and beta blockers can ease symptoms, but none is directly based on the neurobiology of fear [1]. The tools below are judged by how well they do the two steps.

EMDR: eye movements that quiet the threat system

What it is. EMDR (eye movement desensitization and reprocessing) was developed by psychologist Francine Shapiro in the 1980s [1]. As the story goes, she was walking among the trees behind Stanford, recalling a troubling event, and noticed its emotional weight lessen [2]. In the clinic, patients move their eyes side to side for 30 to 60 seconds and then recount the difficult event [2]. Huberman notes it is one of the few behavioral treatments approved by the American Psychological Association for trauma [2].

Huberman's change of mind. As a vision scientist, Huberman first thought EMDR was crazy [2]. He rejects two popular explanations: that it mimics REM sleep's eye movements (it doesn't, since REM eye movements are erratic and go in all directions) and that it synchronizes the left and right brain [1][2]. What changed his view was research. He says at least five high-quality peer-reviewed papers from the last five years, with a sixth on the way, show that lateral eye movements, but not vertical ones, suppress the amygdala [2], reducing sympathetic arousal [1]. His explanation: side-to-side eye movements are the ones we make reflexively when moving forward through space by walking, running or cycling, and forward self-generated movement is generally incompatible with freezing or retreating [1][2].

Where it fits the model. In Huberman's framework, EMDR mainly addresses extinction. It lowers the physiological response during recounting and sidesteps the intense first retellings [1]. But he isn't aware of it actively including the second step, relearning a new narrative with reward attached, which requires some positive arousal, so he suspects it is often incomplete on its own [1]. Colleagues told him EMDR works best for single-event traumas, such as a car accident or an assault, rather than an entire bad marriage or childhood [1][2]. It should be done in a clinical setting with a certified practitioner [2].

Spiegel's skepticism. Spiegel sees EMDR differently. He thinks it contains a lot of hypnosis, noting that a swinging watch was once the classic hypnotic induction and that people once worried windshield wipers would hypnotize drivers [3]. He agrees many people with trauma improve with EMDR, but says studies that remove the eye movements find they add nothing, which makes it essentially an exposure therapy [3]. He prefers that trauma be processed more deeply [3]. When Huberman raises the newer data suggesting eye movements may calm the amygdala, Spiegel connects it to Allan Hobson's observation that emotion in dreams stays surprisingly constant [3].

Clinical hypnosis: control, not helplessness

What it is. Spiegel, associate chair of psychiatry and behavioral sciences at Stanford, rejects the stage-show image of mind control. He describes hypnosis as a state of highly focused attention that gives people more control over their minds and bodies [3]. "They think you're losing control. You're gaining control" [3].

The brain changes. His fMRI work found three shifts [3]. Activity drops in the dorsal anterior cingulate cortex, part of the salience network that detects conflict and danger. The dorsolateral prefrontal cortex couples more tightly with the insula, a mind-body control area. And the prefrontal cortex decouples from the posterior cingulate, part of the self-reflective default mode network. That dissociation lets people act without worrying about what it means for them, which increases cognitive flexibility and willingness to try something new [3]. More hypnotizable people also show more GABA activity in the anterior cingulate, a natural calming effect [3].

The core idea about trauma. Spiegel's central claim is that "the essence of trauma is helplessness it's not fear it's not pain it's helplessness" [3]. Self-hypnosis is something you make happen and can switch on and off, so it restores a sense of control [3]. That puts his emphasis in a different place from Huberman's fear-extinction model. For Spiegel, the problem is less the fear than the powerlessness.

The technique. Spiegel teaches people to imagine their body floating somewhere safe and comfortable, then picture the problem on an imaginary screen, with the rule that whatever they see, the body stays comfortable [3]. For trauma, he splits the screen. He describes a woman who survived an attempted rape that left her with a skull fracture. Under hypnosis, with her body floating safely, she pictured the attacker on one side of the screen and realized something she hadn't let herself remember: he meant to kill her [3]. On the other side, she pictured what she did to protect herself and realized that by fighting so hard she had probably saved her life [3]. The truth was worse than she'd thought, but she could now see her own strength [3]. In Huberman's terms, that is both extinction (facing the memory with a calm body) and a new narrative (seeing her own strength) in one session.

Why a hypnotic state fits trauma. Spiegel draws on the late Stanford psychologist Gordon Bower's work on state-dependent memory: we recall best in the mental state we were in [3]. Many trauma survivors dissociate during the event, so hypnosis can bring the therapy state closer to the one they were in [3]. He also believes changing one's mental state has therapeutic power in itself, which he calls unsystematic desensitization [3].

Evidence. Spiegel cites a randomized trial from Israel showing that adding hypnosis to PTSD treatment improved outcomes [3]. His broader evidence comes from pain, stress and anxiety. In a Lancet trial, self-hypnosis cut procedural pain by 80% with half the opioids [3]. He usually sees patients only once or twice and teaches them to do it on their own [3].

On confronting pain. "I'm not a big fan of trigger warnings," Spiegel says. What matters is feeling in control of how we approach upsetting material [3]. He cites Karen Parker's stress inoculation studies in young monkeys [3].

Limits. Hypnosis depends on hypnotizability. About a third of adults aren't hypnotizable, about 15% are extremely so, and training raises scores only a little [3]. A simple eye-roll test gives a rough estimate. For low scorers, Spiegel uses other approaches such as medication or mindfulness [3]. People with OCD tend to be less hypnotizable because they're so busy evaluating rather than experiencing [3]. He recommends working with a clinician licensed in their primary discipline [3].

Ketamine: dissociation now, rewiring later

What it is. Ketamine is a dissociative anesthetic, broadly similar in action to PCP [4]. Huberman describes the patient experience through a Stanford colleague: "dissociation is where people feel separate from their body. They can still think, but it's as if they are observing themselves" [4]. One patient described it as feeling like they were getting out of the cockpit and watching themselves [1]. It has a double identity: in the right clinical context, Huberman says, it has proven miraculous for some people, though not all, with depression, suicidality and PTSD, but with a high potential for abuse [4]. Early uses for PTSD date to the 1970s, mainly in soldiers returning from Vietnam, but broad clinical use took off only in the last five to ten years [4]. In the US, ketamine-assisted psychotherapy is legal when prescribed by a physician [1].

How it fits the trauma model. In the fear episode, Huberman places ketamine inside his framework. Dissociation, he says, "is really about not feeling what's happening it's about viewing what's happening from a different perspective" [1]. Ketamine "somehow allows the patient the individual to recount their trauma while feeling either none or a very different set of emotional experiences" [1]. "So it's a remapping of new onto old new meaning new feelings onto old feelings while staying in the exact same narrative" [1]. Like EMDR, it suppresses the threat reflex, but it also seems to bring in new emotional experiences to replace the old ones. So, unlike EMDR, it may cover both extinction and relearning [1]. He cites a Stanford study by Karl Deisseroth, Robert Malenka and colleagues showing that ketamine induces a one-to-three hertz rhythm in layer five neurons of the retrosplenial cortex, linked to dissociation [1]. Spiegel mentions the same research and suggests ketamine may help people approach dissociated experiences in a way they can then work with [3].

Blocking the wiring of fear. In the dreams episode, Huberman stresses a different mechanism. Ketamine blocks the NMDA receptor, which normally opens only during intense experiences and triggers long-term potentiation [2]. His example: if a massive explosion went off while he sat down to chicken soup at his kitchen table, the neurons tied to the table and soup would become coupled to the explosion, so that sitting there later would raise his heart rate and make him sweat. Ketamine blocks that coupling [2]. That's why some emergency rooms stock ketamine to give very soon after severe trauma, so the emotion still happens but doesn't get wired to the experience [2]. He flags the ethical questions this raises, since some emotions should be tied to experiences [2]. He also clears up a common confusion: unlike psilocybin or MDMA trials, which are about getting more in touch with emotion, ketamine-assisted therapy is about dissociating from the emotional part [2].

Rewiring over weeks. The ketamine deep dive adds a third layer [4]. The first clinical study gave seven people with major depression intravenous ketamine at half a milligram per kilogram. Relief came almost immediately and lasted at least three days, but it usually fades within about three days to a week [4]. Stacked regimens, such as twice a week for three weeks, can give relief lasting months [4]. The mechanism is more complex than NMDA blockade alone. Ketamine binds NMDA receptors on inhibitory neurons, releasing the brake on excitatory neurons in mood circuits. They then fire in bursts, which triggers BDNF release and lasting plasticity [4]. Ketamine may also mimic BDNF directly [4]. A Stanford study found that blocking opioid receptors with naltrexone left the euphoria and dissociation intact but eliminated the antidepressant effect [4]. And memantine, another NMDA blocker, "has no anti-depressant effects" [4]. Huberman's point: "The acute or immediate effects of ketamine while one is under the influence of ketamine are just part of the story" [4].

Where it fits best. Huberman is measured. Whether ketamine "turns out to be the ultimate treatment for trauma and for fear I isn't clear my colleagues in psychiatry tell me that that's unlikely" [1]. It does seem helpful "especially people that are experiencing trauma or have existing traumas and fear that are coupled with depressive symptoms because the data on ketamine and depression seems to be quite strong" [1]. It has been shown useful for PTSD, OCD, anxiety and some substance addictions as well as depression [4].

The risks. Ketamine can be addictive [4]. Dose depends heavily on route. An injected study dose of about 50 milligrams for a 100-kilogram person corresponds to about 150 milligrams as a pill or 200 as a sublingual troche [4]. A K-hole is a dose that pushes someone past the dissociative range into a pseudo-anesthetized state, and people metabolize ketamine very differently [4]. Combining it with alcohol or barbiturates can be deadly. It can trigger seizures in seizure-prone people and dramatically raise liver enzymes [4]. At-home and telehealth use carries added risk without someone in the room [4]. His colleagues know of no published evidence that microdosing helps depression [4]. And no drug works alone: "I don't actually believe in miracle drugs," he says, closing with a colleague's line, "better living through chemistry still requires better living" [4].

MDMA: extinction and relearning under a feeling of safety

MDMA is not the focus here, but Huberman's fear episode sets it beside ketamine, and it clarifies what the others lack. MDMA was in clinical trials at the time and "still an illegal drug to possess or to sell" outside them [1]. It uniquely raises dopamine and serotonin at the same time and triggers massive oxytocin release. One paper found plasma oxytocin averaging 83.7 pg per milliliter about 90 to 120 minutes into a session, compared with a typical level of 18.6 [1]. Dopamine likely drives euphoria, serotonin a sense of safety and comfort, and oxytocin deep feelings of connection [1]. "What it seems to allow is a very fast re learning or new associations to be tacked on to the previously traumatic experience" [1]. In Huberman's model, the safety lets people lean into parts of their story they'd otherwise avoid, so extinction and relearning can happen quickly without many painful repetitions [1].

REM sleep: the therapy you already get

Huberman adds a natural comparison. During REM sleep, norepinephrine and its body counterpart epinephrine (adrenaline) are essentially absent. "It's probably one of the few times in our life that epinephrine is essentially at zero activity within our system" [2]. That lets the brain replay real events, and distortions of them, without fear attached, adjusting our emotional relationship to hard experiences [2]. REM deprivation makes people irritable, makes little things feel big and leads to catastrophizing [2].

He sums up the parallel neatly: ketamine blocks plasticity chemically, EMDR suppresses the amygdala behaviorally, and REM sleep removes epinephrine naturally [2]. "A certain component of our sleeping life is acting like therapy," he says, and sleep deprivation "is deprivation of self-induced therapy every time we go to sleep" [2].

Self-directed stress and social connection as supports

Two more tools from the fear episode cut across all the treatments. First, social connection: the neuropeptide tachykinin rises in the central amygdala soon after trauma and reinforces fear learning. Isolation raises it further, while trusted connection, such as conversation, a shared meal or appropriate touch, lowers it [1]. People working through trauma, especially alone or in a transactional therapy relationship, should deliberately seek regular trusted connection [1].

Second, brief self-directed stress. In a mouse study, five minutes a day of intense stress reversed the effects of chronic stress, while 15 minutes made things worse [1]. Huberman links this to unpublished work with Spiegel on five-minute breathing protocols, cyclic sighing and cyclic hyperventilation, both of which showed impressive stress reduction [1]. Spiegel's insight ties it to trauma: "it's not just the state that you are in or that you go into, it's how you got there and whether or not you had anything to do with it" [1]. That fits Spiegel's view that helplessness is the core of trauma [3]. Huberman strongly cautions people with panic or anxiety disorders against cyclic hyperventilation [1].

Where They Agree, and Where They Don't

They agree on the goal. Every source describes separating emotion from memory, not erasing memory. Huberman: you remove the emotional load [2]. His extinction step turns trauma into a "boring bad story" [1]. Spiegel's screen technique keeps the body calm while the mind revisits the event [3]. Ketamine lets people recount trauma while feeling something different [1].

They agree that dissociation and focused states can be therapeutic. Spiegel links hypnosis to state-dependent memory in people who dissociated during trauma, and sees a parallel with ketamine [3]. Huberman describes ketamine's dissociation as viewing the experience "from a different perspective" [1]. Both mention the same Deisseroth research on the dissociation rhythm [1][3]. There's an irony here that Huberman notes: dissociation is also a primary symptom of PTSD [4].

They disagree about EMDR's active ingredient. This is the sharpest split. Huberman, persuaded by at least five studies, says lateral eye movements suppress the amygdala [1][2]. Spiegel says studies that remove the eye movements find they add nothing, so EMDR is essentially exposure therapy, perhaps with some hypnosis mixed in [3]. Both agree many people improve. They disagree on why, and Spiegel would rather trauma be processed more deeply [3].

They put the core of trauma in different places. For Huberman, trauma is fear embedded in the nervous system that shows up at the wrong times, and treatment means extinguishing it and wiring in a new narrative [1]. For Spiegel, "the essence of trauma is helplessness" [3], so treatment means restoring control. These overlap. Huberman's self-directed stress work leans on Spiegel's point about having something to do with how you got there [1]. But they lead to different emphases: repetition and reward for Huberman, agency for Spiegel.

They rate the tools' completeness differently. By Huberman's two-step test, EMDR is often incomplete because it handles extinction without clear relearning [1]. Ketamine and MDMA seem to carry both steps: ketamine through remapping new feelings onto the old story, MDMA through fast relearning in a state of safety [1]. Hypnosis, as Spiegel describes it, also does both in the split-screen method [3], though Huberman doesn't score it in his framework.

Ketamine is described in apparently conflicting ways. In the dreams episode, Huberman stresses that ketamine blocks plasticity, preventing emotion from being wired to experience [2]. In the ketamine episode, he stresses that it drives plasticity, releasing inhibition, triggering burst firing and BDNF and rewiring mood circuits, and says NMDA blockade alone isn't the explanation, since memantine doesn't work [4]. The two views can fit together. Blocking the NMDA receptors that would lock in a fresh trauma is one effect. Releasing the brake on inhibitory neurons so mood circuits can rewire over weeks is another. But readers should know the sources describe several mechanisms at different time scales, not one simple story [4].

They agree nobody has a miracle cure. Huberman's colleagues doubt ketamine is the ultimate trauma treatment [1], he says he doesn't believe in miracle drugs [4], EMDR suits single-event traumas best [1][2], and hypnosis doesn't work for about a third of adults [3].

Why It Works

Fear is a learned link that can be weakened. Trauma forms through one-trial learning and long-term potentiation at NMDA receptors [1]. The same machinery runs in reverse through long-term depression, the weakening of connections that happens when a fear is extinguished [1]. Every tool here aims to weaken the link between memory and alarm.

Calm while recalling. Extinction needs the memory to be recalled without the full threat response. EMDR's lateral eye movements suppress the amygdala [1][2]. Hypnosis quiets the salience network and keeps the body floating while the mind watches the screen [3]. Ketamine's dissociation lets people observe rather than feel [1][4]. REM sleep replays events with epinephrine at zero [2].

A new story needs reward. The fear circuit's link to the dopamine system allows positive associations to be wired onto old events [1]. MDMA's dopamine, serotonin and oxytocin supply reward, safety and connection all at once [1]. Ketamine appears to make the reward pathway more available by weakening the habenula's inhibition and strengthening reward-frontal connections [4].

Matching states helps recall. State-dependent memory means a dissociated-like therapy state may reach memories encoded during dissociation [3].

Agency is itself therapeutic. If helplessness is the essence of trauma [3], then a self-directed state, whether self-hypnosis or chosen stress, directly counters it [1][3].

Timing and dose matter. Ketamine given soon after trauma may keep emotion from wiring to the experience [2]. Stacked ketamine sessions extend relief from days to months [4]. Five minutes of self-directed stress helps while 15 minutes harms in mice [1]. And calming compounds taken before an extinction session may short-circuit extinction by damping the response [1].

Where It Applies, and Where It Doesn't

EMDR: Best for single-event traumas that can be recalled in detail, such as a car crash or an assault. Less suited to diffuse traumas like an entire childhood or marriage [1][2]. It needs a certified practitioner [2]. It may need a separate relearning step on top [1].

Clinical hypnosis: Suits hypnotizable people, roughly two-thirds of adults, with a third not responding and training barely helping [3]. It is less suited to people with OCD [3]. It is brief, with one or two sessions and then self-practice, and has an Israeli trial supporting it as an addition to PTSD treatment [3]. It should be done with a licensed clinician [3].

Ketamine: Legal when prescribed in the US [1]. Especially useful when trauma comes with depression [1]. Relief is short unless sessions are stacked [4]. It carries serious risks: abuse, K-holes, deadly combinations with alcohol or barbiturates, seizures and liver stress. There is no evidence for microdosing, and it depends on behavior outside the session [4].

MDMA: Legal only in clinical trials at the time Huberman described it [1].

REM sleep: Available to everyone and free, but it is support, not stand-alone treatment for PTSD [2]. Protecting it protects the brain's nightly emotional processing [2].

For everyone: None of these is a do-it-yourself fix for serious trauma. Huberman repeatedly urges clinical support [1]. Detailed journaling can help people without access to therapy, but with caution [1]. If you're struggling with trauma, work with a qualified mental health professional to choose among these options.

What To Do With It

1. Understand the two steps. Look for care that both lowers the emotional charge of the memory (extinction) and builds a new, rewarding meaning around it (replacement) [1]. Ask any therapist how their approach does both.

2. Match the tool to the trauma. For a single, well-defined event, EMDR with a certified practitioner is a reasonable option [1][2]. If you use EMDR, ask how the relearning step will be handled, since Huberman suspects it's often missing [1].

3. Check your hypnotizability. Tilt your head back, look up and slowly close your eyes. If mostly the whites show, hypnosis may suit you [3]. If you're hypnotizable, ask about hypnosis-based trauma work, and find clinicians through the Society for Clinical and Experimental Hypnosis or the American Society of Clinical Hypnosis [3].

4. Consider ketamine for trauma with depression, under medical supervision. It seems most helpful when trauma is coupled with depressive symptoms [1]. Use it only through a prescribing physician, understand how dose varies by route, never combine it with alcohol or barbiturates, and don't use it alone at home [4].

5. Keep up antidepressive habits between sessions. Morning sunlight, regular quality sleep, good nutrition and social engagement reinforce the circuit changes that ketamine and therapy open up [4]. "Better living through chemistry still requires better living" [4].

6. Protect your sleep, especially REM. REM sleep does natural emotional processing without adrenaline [2]. Keep a consistent sleep amount, and avoid alcohol, THC and serotonin-boosting supplements before bed, which disrupt the depth and sequence of sleep [2].

7. Don't go it alone. Seek regular, trusted social connection, such as conversation, a shared meal or appropriate touch, which lowers fear-reinforcing tachykinin [1].

8. Reclaim control in small ways. Brief, self-chosen challenges counter the helplessness at the heart of trauma [1][3]. Cyclic sighing for five minutes is a calming option. Avoid cyclic hyperventilation if you have panic or anxiety disorders [1].

9. Time calming aids carefully. If you use anti-anxiety supplements or medications, talk with your clinician about taking them outside sessions meant to extinguish a fear, since damping the response can short-circuit extinction [1].

Sources

1. Erasing Fears & Traumas Based on the Modern Neuroscience of Fear · Huberman Lab
2. Understand and Use Dreams to Learn and Forget · Huberman Lab
3. Dr. David Spiegel: Using Hypnosis to Enhance Health & Performance · Huberman Lab
4. Ketamine Benefits and Risks for Depression, PTSD & Neuroplasticity | Huberman Lab Podcast · Huberman Lab