Why Integration Fails

This is the final essay in a ten-part series on our Signal Loss Model. It assumes familiarity with the model developed across SLM 1–SLM 9.

Psychedelic therapy has a durability problem, and the trial literature has been slow to say so. Taken one at a time, the results look encouraging. Taken together, they converge on something more specific: the acute, short-window effect is real and replicated, while durable remission is far less settled.

Start with what is solid. Across molecules, sites, and designs, psilocybin produces a strong acute antidepressant response, often between 50% and 70% at three to six weeks. Two randomized trials against active placebo reproduced the rapid effect as recently as 2023 and 2026 (Raison et al. 2023; Yngwe et al. 2026). The short-window result is not seriously in dispute.

Durability is where it comes apart. The COMPASS Pathways Phase IIb trial, the largest randomized study at the time, reported strong response at three weeks and a much thinner signal by twelve, with sustained response in the 25 mg arm falling to roughly 20% (Goodwin et al. 2022). The 2026 Karolinska trial, the first placebo-controlled study to follow patients a full year, found psilocybin superior to niacin at six weeks but not at twelve months on clinician-rated depression; the self-reported advantage reached about three months, and shortened to roughly ten weeks once participants who resumed antidepressants were censored (Yngwe et al. 2026). The EPISODE trial in treatment-resistant depression did not meet its primary response endpoint at six weeks at all (Mertens et al. 2026).

The most-cited sustained result does not close the question either. The Johns Hopkins twelve-month follow-up reports remission holding at a high level across the year (58% at one week, 54% at three months, 71% at six, 58% at twelve), but it is a small open-label sample with clear expectancy confounds and no blinded control (Gukasyan et al. 2022). And in the one head-to-head against a standard antidepressant, psilocybin showed no significant separation from escitalopram on the primary endpoint at six weeks (Carhart-Harris et al. 2021). The durable picture is mixed where it is not simply thin.

The throughline is consistent. Rapid short-window relief is well supported. Durable remission is not, and the gap between the two widens as the trials improve: longer follow-up, more active placebos, closer tracking of antidepressant use and unblinding. The acute window is real. What happens after it is the open question.

The standard explanation is "insufficient integration." That explanation is correct and almost entirely useless. It's the equivalent of diagnosing a collapsed building as having had insufficient structural support. The interesting questions are mechanical: what specific biological processes govern the window of opportunity, what closes it, and why do so many integration programs ignore the mechanism of closure?

The Window, Briefly

Earlier essays in this series have covered the relevant neurobiology in detail. SLM 8 established that psychedelics reliably open a temporary plasticity window by resetting autonomic balance, upregulating BDNF, reducing default mode network rigidity, and promoting dendritic spine growth and synaptogenesis. SLM 9 argued that this window creates the biological conditions for constraint installation: the system is temporarily plastic enough to accept new organizational inputs.

What the series has not yet addressed is what governs the closing of that window. Closure is usually treated as a passive process, a fading, a return to baseline, as though plasticity simply dissipates like heat from a cooling surface. That framing is wrong in ways that matter for everything that follows.

The window doesn't fade. It is actively closed.

Why the Window Closes

The neuroscience of plasticity regulation has been studied most rigorously in the context of developmental critical periods. The visual cortex is the best-characterized example. In early development, the visual system is extraordinarily plastic: it rewires in response to experience, establishing the neural architecture for binocular vision, edge detection, and spatial processing. Then, over a defined period, that plasticity is shut down. The system consolidates what it has learned and locks the configuration in place.

Takao Hensch's landmark work, published in Nature Reviews Neuroscience in 2005, established the mechanisms. Critical period closure is driven by the maturation of specific inhibitory circuits, particularly parvalbumin-expressing (PV+) interneurons. As these fast-spiking inhibitory neurons mature, they impose increasingly tight temporal control over excitatory signaling, narrowing the system's capacity for reorganization. Simultaneously, the extracellular matrix surrounding these neurons condenses into structures called perineuronal nets (PNNs), physical lattices of proteoglycans that stabilize existing synaptic connections and prevent the formation of new ones (Hensch, 2005).

The closure is not incidental. It is engineered. The brain has evolved dedicated molecular machinery whose function is to end plasticity once a learning window has served its purpose.

This matters for the psychedelic context because the same closure machinery appears to operate in adult brains. Pizzorusso and colleagues demonstrated in 2002 that enzymatically degrading perineuronal nets in the adult visual cortex reopened critical-period-like plasticity: the system could reorganize again as though the developmental window had never closed (Pizzorusso et al., 2002). The lock on adult plasticity is physical. It can be picked. And it reasserts.

Emerging evidence suggests that psychedelics may work, at least in part, through related plasticity mechanisms, including critical-period-like reopening and downstream extracellular-matrix remodeling. There is growing interest in whether serotonergic psychedelics may alter the plasticity-regulating machinery that governs critical period closure, including PV+ interneuron dynamics and extracellular-matrix structures such as perineuronal nets, in ways that parallel developmental critical period reopening (Calder and Hasler 2023). More recent work has strengthened the case: Zhang and colleagues' 2026 review in Biological Psychiatry explicitly maps psychedelic mechanisms onto extracellular-matrix remodeling and metaplasticity pathways, connecting the developmental critical period literature to the psychedelic plasticity literature in mechanistic terms (Zhang et al., 2026).

Gül Dölen's laboratory has provided the most direct experimental evidence for this framework. In a 2019 study published in Nature, Nardou and colleagues demonstrated that MDMA reopens the critical period for social reward learning in adult mice. Crucially, the reopened window had a defined duration. It did not persist indefinitely. A 2023 follow-up in Natureextended the finding across multiple psychedelic compounds, showing that critical period reopening is a shared property of the drug class and that extracellular-matrix reorganization is a common downstream mechanism. The window opens, learning becomes possible again, and then the biology closes it on a clock (Nardou et al., 2019; Nardou et al., 2023).

Shao and colleagues provided complementary structural evidence. Using two-photon imaging of mouse cortex, they showed that psilocybin induced rapid dendritic spine formation within twenty-four hours. Some of these new spines persisted at one month. Many did not. The spines that survived were those that had been functionally reinforced through use. The rest were pruned (Shao et al., 2021).

This is the critical insight, and it reframes the entire integration question. The plasticity window is not a gift that fades. It is a temporary override of a system that was designed to prevent exactly this kind of reorganization. The brain built perineuronal nets, matured its inhibitory circuits, and locked its synaptic configurations for a reason: stability is metabolically cheaper than plasticity, and from an evolutionary standpoint, an organism that perpetually rewires is an organism that can't act on what it has already learned. Plasticity is expensive, destabilizing, and time-limited by design.

Psychedelics temporarily override those protections. The override is real. It creates genuine opportunity. And the homeostatic machinery that closed the window in the first place will close it again, because that is what it was built to do.

The practical consequence is stark: whatever reorganization occurs during the window must be reinforced strongly enough to survive the pruning that follows. New dendritic spines that are not functionally consolidated will be eliminated. New circuit configurations that are not repeatedly activated will be overwritten. The system is not passively returning to baseline. It is actively restoring prior architecture. Integration is a race against a biological process that is working to undo everything the experience made possible.

There is a further implication, and it is the one that matters most: plasticity is directionless. An open window does not favor therapeutic reorganization over pathological reconsolidation. The system will crystallize around whatever patterns are most active during the plastic period. If a person returns from a psychedelic experience to the same environmental stressors, the same relational dynamics, the same autonomic triggers, and the same absence of embodied constraint, the open window will consolidate those inputs with fresh biological cement. The pruning process does not selectively eliminate old patterns and protect new ones. It eliminates whichever connections are least reinforced. 

A person who spends the window in reflection, without structured behavioral and physiological change, may find that the window's closing deepens the prior state rather than restoring it, because the system has now consolidated the old architecture through a full cycle of plasticity and restabilization.

This is the cruelest feature of the mechanism. The window is not a grace period. It is a competition between old patterns and new ones for biological survival, and the old patterns have every structural advantage: existing synaptic infrastructure, established circuit dynamics, and the full weight of environmental reinforcement. New patterns can win that competition. They cannot win it passively.

Why Integration Misses the Mechanism

Psychedelic therapy, to its credit, has recognized that integration matters. Virtually every serious program includes some form of post-experience support: reflective sessions, journaling prompts, group calls, referrals to talk therapy, mindfulness practices, community connection. These are all reasonable things to offer someone who has just had a profound experience. They are also, considered against the biology just described, largely operating at the wrong level of mechanism.

The Signal Loss Model identifies three interlocking systems whose simultaneous dysfunction produces the treatment-resistant conditions most commonly seen in high-functioning populations. The model's therapeutic logic, developed across SLM 6–SLM 9, holds that durable change requires addressing all three during the window when the biology permits reorganization. Most programs address, at best, one of the three. Usually the easiest one. Usually incompletely.

Untethered Cognition (UC)

The cognitive-architectural problem described in SLM 2 and SLM 3: the human simulation engine requires continuous real-world constraint and feedback to remain calibrated. When constraint is removed or becomes pathological, simulation decouples from reality, producing rumination, identity distortion, and the recursive cognitive patterns characteristic of the Achievement Paradox.

During a plasticity window, the simulation system is temporarily available for recalibration. New constraint inputs can be accepted. The system can, briefly, reorganize around a different relationship to reality.

The standard response to this mechanism is, almost universally, reflective conversation. "What did you see? What does it mean? How does it connect to your life?" This is cognitive processing of simulation content. It is simulation aboutsimulation. It does not provide the system with new constraint. It provides the system with new material to simulate about, which the unconstrained mind will "happily" incorporate into its existing recursive architecture. You can journal about your breakthrough for six weeks and arrive at a beautifully articulated narrative that has changed nothing about how your simulation engine operates. The narrative becomes another loop.

What the UC mechanism requires during the window is embodied, relational, sensorily grounded experience that provides the simulation system with real-world feedback it cannot generate internally. Physical work. Novel environments that demand present-tense engagement. Relational encounters that resist abstraction. The specific content matters less than the structural quality: the input must be the kind that constrains simulation rather than feeding it.

Neuroimmune Dysregulation (ND)

The biological problem described in SLM 4: chronic stress produces persistent low-grade inflammation that suppresses neuroplasticity and physically locks maladaptive patterns in place through inflammatory-mediated changes to synaptic function, glial behavior, and neural circuit maintenance.

The psychedelic experience appears to temporarily disrupt this lock-in. Acute psychedelic administration has been associated with reductions in inflammatory markers, restoration of neurotrophic signaling, and a temporary lifting of the inflammatory suppression of plasticity. This is part of what makes the window a window: the biological brakes on reorganization are briefly released.

For this mechanism there is essentially no standard model at all. Standard post-experience protocols do not include any systematic approach to autonomic regulation, vagal tone maintenance, or inflammatory management during the recovery window. The assumption appears to be that the experience itself resolves the inflammatory problem, or that the inflammatory dimension is outside the scope of psychological integration.

The biology suggests otherwise. If vagal tone is not actively maintained during the post-experience period, if the autonomic system is not given structured input that sustains the parasympathetic shift initiated by the psychedelic experience, the inflammatory cascade described in SLM 4 will reassert. The same chronic stress physiology that produced the original lock-in will rebuild it. The process is physiological, and it requires a physiological intervention during the window. Breathwork, cold exposure, sustained aerobic movement, specific somatic practices that drive vagal tone: these are not wellness accessories. They are, in the context of the SLM, mechanistically necessary components of integration that most programs have filed under optional lifestyle recommendations.

Pursuit-Reward Decoupling (PRD)

The motivational problem described in SLM 5: sustained cortisol elevation downregulates dopamine receptor density, collapsing incentive salience and draining the external world of motivational pull. The person knows, cognitively, what they should want. The wanting itself is physiologically unavailable.

Many people report a temporary restoration of motivation, curiosity, and engagement with the world in the days and weeks following a psychedelic experience. Colors are brighter. Music is more affecting. Social connection feels available again. This is the dopaminergic system temporarily freed from the cortisol-mediated suppression described in SLM 5. It is the biological opportunity, and it is routinely mistaken for the solution.

The restoration of incentive salience during the post-experience window is a signal that the reward system is temporarily accessible for retraining. If that window is filled with novel, effortful, real-world reward-seeking activity, the dopaminergic system can begin establishing new reward associations that may survive the window's closure. If the window is spent reflecting on the experience, the temporary restoration of feeling will fade as the cortisol-dopamine relationship reasserts, and the person will be left with a vivid memory of what motivation felt like and no durable access to it.

The qualitative research bears this out. Watts and colleagues, studying psilocybin therapy patients, found that those who maintained improvements reported actively restructuring their daily lives during the post-experience period: new activities, new social patterns, new behavioral commitments. Those who reverted described the experience as meaningful but reported no sustained behavioral change (Watts et al., 2017). The insight was real. The behavioral reinstallation didn't happen. And the window closed.

The Mismatch

Map the three mechanisms against the integration toolkit in common use.

Reflective processing, journaling, and talk therapy primarily address the narrative-cognitive dimension of the experience. They are useful for making meaning of what was revealed. They operate almost entirely within the simulation system and do not provide the embodied constraint that UC requires.

Community support and group integration circles provide social connection, which has real value for autonomic regulation and reward engagement. But unstructured social processing, without specific attention to vagal tone maintenance or dopaminergic retraining, provides attenuated and inconsistent biological input.

Mindfulness practices, depending on the specific technique, can address autonomic regulation. But mindfulness is typically offered as a general recommendation rather than as a targeted intervention calibrated to the post-experience recovery timeline and the specific biological processes it needs to support.

No widely used integration framework currently addresses all three mechanisms during the window with the specificity and timing that the biology demands. The programs in wide use are psychologically thoughtful and biologically incomplete. The problem is structural: a mismatch between the mechanism of the intervention and the mechanism of its delivery, not a failure of intention.

The reason this matters is that signal loss, as the series has argued from SLM 6 onward, is a coupled system. Addressing one mechanism while the others remain active doesn't produce partial improvement. It produces temporary improvement that the untreated mechanisms pull back toward baseline. You can restore vagal tone, but if the simulation system remains unconstrained, rumination will re-drive the autonomic dysregulation that collapses vagal tone. You can temporarily reopen incentive salience, but if inflammation continues suppressing plasticity, the new reward associations won't consolidate. You can provide embodied constraint during the retreat, but if the person returns to the same environmental inputs that produced the original failure, those inputs will reinstall the old patterns through the very plasticity the experience opened. Each mechanism, left unaddressed, actively undermines the gains made in the others. This is why single-modality integration plateaus: the untreated levels aren't neutral. They are countervailing forces.

The relapse curves in the clinical data are the predictable result of this mismatch. The experience reliably opens the window. The integration model doesn't address what closes it. The homeostatic machinery reasserts. The spines are pruned. The inflammation returns. The reward collapse resumes. And the person is left wondering why something that felt so transformative didn't last.

What This Means for the Model

The Signal Loss Model began, in SLM 1, with a simple observation: that depression, anxiety, and treatment resistance in high-functioning populations are better understood as system-level failures than as discrete diseases. Across nine essays, the model identified the three mechanisms that produce and sustain that failure, explained their interactions, and described the biological conditions under which change becomes possible.

This final essay adds one claim: that the most promising intervention currently available for these conditions is systematically failing at durability, and the failure is predictable from the biology.

The plasticity window opened by psychedelic experience is real, measurable, and temporary. It is temporary because the brain has evolved dedicated machinery to close it. The closure is active, not passive. The integration model required to produce durable change during that window must address all three SLM mechanisms with appropriate specificity and timing, or the system will restore its prior configuration.

SLM = UC + ND + PRD. That equation describes both the collapse and the requirements for its reversal. Research and practice have focused, understandably, on opening the window. The question that remains is whether anyone will build what happens during it with sufficient biological precision to survive its closing.

The series ends here. The building does not.

References

Calder, Abigail E., and Gregor Hasler. 2023. “Towards an Understanding of Psychedelic-Induced Neuroplasticity.” Neuropsychopharmacology: Official Publication of the American College of Neuropsychopharmacology 48 (1): 104–12. 

Carhart-Harris, Robin, Bruna Giribaldi, Rosalind Watts, et al. 2021. “Trial of Psilocybin versus Escitalopram for Depression.” The New England Journal of Medicine 384 (15): 1402–11. 

Davis, Alan K., Frederick S. Barrett, Darrick G. May, et al. 2021. “Effects of Psilocybin-Assisted Therapy on Major Depressive Disorder: A Randomized Clinical Trial.” JAMA Psychiatry 78 (5): 481–89.

Goodwin, Guy M., Scott T. Aaronson, Oscar Alvarez, et al. 2022. “Single-Dose Psilocybin for a Treatment-Resistant Episode of Major Depression.” New England Journal of Medicine 387 (18): 1637–48.

Gukasyan, Natalie, Alan K. Davis, Frederick S. Barrett, et al. 2022. “Efficacy and Safety of Psilocybin-Assisted Treatment for Major Depressive Disorder: Prospective 12-Month Follow-Up.” Journal of Psychopharmacology 36 (2): 151–58.

Hensch, Takao K. 2005. “Critical Period Plasticity in Local Cortical Circuits.” Nature Reviews. Neuroscience 6 (11): 877–88.

Mertens, Lea J., Michael Koslowski, Felix Betzler, et al. 2026. “Efficacy and Safety of Psilocybin in Treatment-Resistant Major Depression: The EPISODE Randomized Clinical Trial.” JAMA Psychiatry 83 (5): 448.

Nardou, Romain, Edward Sawyer, Young Jun Song, et al. 2023. “Psychedelics Reopen the Social Reward Learning Critical Period.” Nature 618 (7966): 790–98.

Nardou, Romain, Eastman M. Lewis, Rebecca Rothhaas, et al. 2019. “Oxytocin-Dependent Reopening of a Social Reward Learning Critical Period with MDMA.” Nature 569 (7754): 116–20.

Pizzorusso, Tommaso, Paolo Medini, Nicoletta Berardi, Sabrina Chierzi, James W. Fawcett, and Lamberto Maffei. 2002. “Reactivation of Ocular Dominance Plasticity in the Adult Visual Cortex.” Science 298 (5596): 1248–51.

Raison, Charles L., Gerard Sanacora, Joshua Woolley, et al. 2023. “Single-Dose Psilocybin Treatment for Major Depressive Disorder: A Randomized Clinical Trial.” JAMA 330 (9): 843.

Shao, Ling-Xiao, Clara Liao, Ian Gregg, et al. 2021. “Psilocybin Induces Rapid and Persistent Growth of Dendritic Spines in Frontal Cortex in Vivo.” Neuron 109 (16): 2535-2544.e4.

Watts, Rosalind, Camilla Day, Jacob Krzanowski, David Nutt, and Robin Carhart-Harris. 2017. “Patients’ Accounts of Increased ‘Connectedness’ and ‘Acceptance’ After Psilocybin for Treatment-Resistant Depression.” Journal of Humanistic Psychology 57 (5): 520–64.

Yngwe, Hampus, Pontus Plavén-Sigray, Carl Johan Ekman, et al. 2026. “Short-Term and Late-Term Effects of Psilocybin on Symptoms in Major Depression: A Randomized Clinical Trial.” JAMA Network Open 9 (5): e2612589.

Zhang, Jin, Cong Lin, Xinyou Lv, Huiying Zhao, and Xiaohui Wang. 2026. “Psychedelics and the Extracellular Matrix: Rewiring Neuroplasticity and Metaplasticity for Next-Generation Psychiatric Therapies.” Biological Psychiatry.

Essays on treatment resistance, altered states, and the conditions under which change becomes possible.

Subscribe