Dissociatives

The founding observation is nearly forty years old and still the cleanest statement of the mechanism. In 1983, ketamine and phencyclidine were shown to selectively reduce the excitation of central neurons by N-methyl-aspartate while leaving responses to other excitatory amino acids largely intact ^[3]. One receptor, selectively silenced.

The way they silence it matters. These drugs are open-channel blockers: they bind a site inside the pore of the NMDA receptor, which means the channel has to open before the drug can get in. Blockade therefore builds up with activity, a property called use dependence ^[4]. The natural comparison is magnesium, which the body already uses to plug the same channel at rest and which is expelled when the cell depolarises. A dissociative is a magnesium block that does not leave.

Two consequences follow. The busiest synapses are blocked hardest, so the drug acts where glutamate signalling is most active rather than uniformly. And because the drug is trapped inside a closed channel when activity falls, how long the effect persists depends on how readily the compound escapes, which is the axis the next section is about.

The paradox is that blocking the brain's main excitatory receptor increases cortical excitation. Subanaesthetic ketamine raises extracellular glutamate in the prefrontal cortex, and the cognitive disruption it causes is prevented by blocking AMPA receptors, so the downstream signal is carried by AMPA rather than by NMDA ^[7]. The resolution came later: NMDA hypofunction quiets fast-spiking inhibitory interneurons more readily than it quiets the pyramidal cells they restrain, so removing the inhibition lets the output cells fire more ^[8]. Cortical disinhibition, not sedation.

That surge is the likely bridge between the acute state and everything else these drugs do. It explains the dose curve, in which low exposure is stimulating and disorganising while high exposure is anaesthetic; it explains the reliable psychotomimetic effects catalogued when subanaesthetic ketamine was first given to healthy volunteers under controlled conditions, including perceptual distortion, thought disorder and blunted affect ^[6]; and, as the antidepressant section describes, it is the leading candidate for why a single dose can change mood for a week ^[2].

Chemically the class is not one family but four, joined only by what they do to the channel. The arylcyclohexylamines are the largest group: phencyclidine and its analogues, ketamine, and the long tail of research chemicals built on the same scaffold, which is where methoxetamine and its successors came from ^[1]. Morphinans contribute dextromethorphan, which is structurally an opioid skeleton with none of the opioid activity. Adamantanes contribute memantine, a licensed Alzheimer's drug. And nitrous oxide is a gas with no organic scaffold at all, shown to be an NMDA antagonist only in 1998, more than a century after it entered anaesthesia ^[9].

Almost none of them is clean. Dextromethorphan is simultaneously a sigma-1 agonist and a serotonin and norepinephrine reuptake inhibitor, which is why it carries a genuine serotonin syndrome risk that no other dissociative does, and why its metabolism through CYP2D6 matters so much: poor metabolisers and anyone taking a CYP2D6 inhibitor reach far higher concentrations from the same amount ^[10]. Phencyclidine analogues hit the dopamine transporter and sigma receptors as well as the NMDA channel, which is where their stimulant and unpredictable character comes from ^[1]. See sigma receptors and pharmacokinetics.

The research-chemical branch deserves its own warning. When methoxetamine appeared it was marketed as a bladder-safe ketamine substitute, a claim that had no evidence behind it and was contradicted by the case reports that followed ^[1]. Its successors arrived faster than the toxicology could, and most of what is known about any of them is a handful of emergency-department presentations. See research chemicals.

Memantine is taken every day by frail elderly patients. Dizocilpine, an experimental blocker of the same channel, produces neuronal injury in rodents and was never given to people as a therapy. Both are NMDA antagonists. The difference is entirely kinetic, and it is the single most useful idea in this whole area ^[5].

Three properties separate them. Affinity determines how much of the channel population is blocked at a tolerable concentration. Voltage dependence determines whether the drug leaves when the neuron is at rest, which is what allows normal physiological signalling to continue between bursts. And trapping determines whether the drug is sealed inside the channel when it closes, or escapes as the gate shuts. Memantine is low affinity, strongly voltage dependent, and escapes readily, so it blunts pathological continuous activation while leaving ordinary transmission largely intact. Dizocilpine is high affinity, slow to leave and fully trapped, so it shuts the receptor down and keeps it down ^[5].

Ketamine sits between the two, which is exactly why it produces a vivid altered state at a dose that does not abolish consciousness. Nothing about that position was designed; it is an accident of chemistry that turned out to be clinically useful twice, first as an anaesthetic and then as an antidepressant.

The kinetic story also produced the field's most instructive failure. If NMDA blockade were the antidepressant mechanism, then a blocker with a gentler kinetic profile should work with fewer dissociative effects. Lanicemine was that drug. It looked promising in early studies and then failed to separate from placebo in a larger, adequately powered adjunctive trial ^[17]. A cleaner NMDA antagonist did not produce a cleaner antidepressant, which is a strong hint that NMDA blockade is not the whole of what ketamine is doing.

The practical version of all this: when a source describes a new compound as an NMDA antagonist, that sentence tells you the binding site and almost nothing about the dose range, the duration, the tolerability or whether anyone has taken it. Ask for those separately.

One of the genuinely important findings in modern psychiatry is that a single subanaesthetic infusion of ketamine can lift severe depression within hours. The first controlled report was a crossover study in seven patients in 2000 ^[11], and the finding held up in a randomised, placebo-controlled crossover trial in treatment-resistant depression in 2006, where the response appeared within a day and persisted for about a week ^[12]. Two decades of replication later, the acute effect is not in doubt. The mechanism still is.

The leading account runs through the glutamate surge described above. Blockade of NMDA receptors on inhibitory interneurons disinhibits cortical pyramidal cells, glutamate is released, AMPA receptors carry the signal, and the result is a burst of protein synthesis and new spine formation in prefrontal cortex; in rodents this is mTOR dependent and is abolished by blocking AMPA receptors ^[13]. A parallel line found that blockade at rest deactivates eukaryotic elongation factor 2 kinase and rapidly increases BDNF translation, which is the plasticity link (see BDNF and neuroplasticity) ^[14].

That account is well supported and it is not unchallenged. The table below sets out the four live explanations and what each rests on.

What follows from the uncertainty is practical. Esketamine, the S enantiomer delivered as a nasal spray, is approved for treatment-resistant depression on the strength of trials in which it beat placebo when added to a newly started oral antidepressant, by a real but modest margin ^[18]. Approval rests on that comparison, not on a settled mechanism. Nitrous oxide has a proof-of-concept crossover trial showing improvement in treatment-resistant depression after a single hour of inhalation, which is intriguing and remains a small study ^[19]. And the most important unknown is durability: nothing in this class has been shown to hold a remission without repeated dosing, and the long-term consequences of repeated dosing are the subject of the next section.

Beyond the obvious impairment of coordination and judgement while under, this class carries several signature harms that do not resemble those of other drugs.

Ketamine damages the urinary tract. This was described as a new clinical entity in 2007 from a small case series ^[20] and confirmed the following year in a much larger cohort presenting with frequency, urgency, pain and, in the worst cases, a contracted bladder and damage extending to the ureters and kidneys ^[21]. A large survey of recreational users found urinary symptoms to be common and clearly related to how much and how often a person used ^[22]. The damage is dose related, can begin within months of heavy use, and is not always reversible on stopping. There is no equivalent syndrome for the other dissociatives, and the arylcyclohexylamine research chemicals have not been shown to be exempt.

Frequent use impairs memory and raises delusional thinking. A one-year longitudinal study of frequent ketamine users found impairment on working and episodic memory alongside increased delusional ideation, and the deficits tracked the amount used ^[23]. Given that the acute drug blocks the receptor that memory formation depends on, this is the least surprising harm in the class and the most often ignored.

Compulsive redosing is a feature of the pharmacology, not a character flaw. The effect is short, the return to baseline is abrupt, and tolerance to the subjective effect builds quickly, so the natural pattern of use is a session that extends itself. Dependence is well documented and is psychological rather than a classical physical withdrawal syndrome ^[24]. See tolerance and dependence.

Nitrous oxide has a harm all of its own, and it is a vitamin problem. Nitrous irreversibly oxidises the cobalt atom in vitamin B12, inactivating it. Repeated use produces functional B12 deficiency and, through it, subacute combined degeneration of the spinal cord: numbness, weakness, unsteady walking and, if it goes far enough, permanent damage. The case literature is substantial, and the presentation is easy to miss because blood B12 can look normal while the active form is depleted ^[25]. This risk scales with the total amount inhaled and has nothing to do with the acute experience.

Dextromethorphan carries a serotonin syndrome risk that the rest of the class does not, because it inhibits serotonin reuptake as well as blocking the channel ^[10]. Its clearance depends on CYP2D6, so the same amount produces very different concentrations in different people, and a CYP2D6 inhibitor taken alongside it, including several common antidepressants, raises exposure sharply.

Stacking depressants is the way people die. Every drug here depresses the central nervous system, and combining one with alcohol, a benzodiazepine or an opioid multiplies the risk of airway compromise and respiratory depression. Ketamine preserves airway reflexes better than most anaesthetics, which is a genuine advantage that people over-generalise into a false sense of safety; the reflexes are less protective once another depressant is on board. The interactions and stacks tool flags these combinations. This is educational information, not medical advice.

Three honest gaps are worth stating plainly, because silence about them reads as confidence.

Whether the dissociative experience is necessary for the antidepressant effect. It is the field's central question and it is unresolved. Lanicemine's failure argues that blocking NMDA receptors without producing dissociation does not work ^[17]; the metabolite hypothesis argues the opposite, that the therapeutic action can be separated from the NMDA block entirely ^[15]. Both cannot be right, and neither has been tested in a way that settles it.

What repeated therapeutic dosing does over years. The bladder and cognitive harms are documented in frequent recreational use, at exposures far above clinical practice. Whether a maintenance schedule of infusions or nasal spray carries a fraction of that risk, or none of it, is not known, because the trials that supported approval were not long enough to find out ^[18].

Almost everything about the research-chemical branch. For most arylcyclohexylamines sold online there is no potency data, no metabolism data, no toxicology, and no way to know whether the specific harms established for ketamine apply to them, are absent, or are worse ^[1]. Absence of reported harm in a compound nobody has studied is not evidence of safety.