What the Drugs Did to the Brain

The previous posts in this series described what MKUltra did to people. This post describes what MKUltra did to brains — at the molecular, circuit, and systems level. Every claim is sourced from peer-reviewed neuroscience. The pharmacology is real. The doses are documented. The intersection of the two is what the CIA paid for.

LSD-25: the molecule that wouldn’t let go

LSD is active at microgram doses — a millionth of a gram. Its primary target is the serotonin 5-HT2A receptor, where it acts as a partial agonist. But what makes LSD unusual is not its affinity. It is its residence time.

In 2017, Wacker et al. published the crystal structure of LSD bound to the human serotonin 5-HT2B receptor at 2.9 angstrom resolution (Cell). The key finding: a segment of extracellular loop 2 — specifically residue L209 — forms a “lid” that closes over LSD in the binding pocket. The lid makes extensive hydrophobic contacts with both the drug molecule and the surrounding transmembrane helices, physically trapping LSD inside the receptor.

The consequence: LSD’s residence time at the 5-HT2A receptor is 221 minutes at body temperature — nearly four hours of continuous receptor activation from a single binding event. When the researchers mutated the lid residue (L209A), residence time dropped tenfold, from 44 to 4.3 minutes. The mutation also selectively reduced beta-arrestin2 recruitment without affecting the initial G-protein signal — meaning the lid doesn’t just slow dissociation; it changes what the receptor does while LSD is inside it.

This is why a dose measured in micrograms produces effects lasting 8-12 hours. The molecule enters the receptor. The lid closes. The receptor signals continuously until the lid finally opens.

Binding affinities (Liechti, 2017, Neuropsychopharmacology):

Receptor	Ki (nM)	Role
5-HT2A	4.0	Primary psychedelic target
5-HT1A	3.0	Anxiolytic/modulatory
5-HT2C	15.0	Mood, appetite
D2	25.0	Dopaminergic — psychotomimetic
D1	310	Weaker dopaminergic

LSD binds potently to all 13 human serotonergic GPCRs. Its D2 affinity is comparable to its serotonin affinity — this dopaminergic component contributes to psychotic symptoms at high doses, operating through D2-5-HT2A heteroreceptor complexes that enhance dopamine signaling.

Preller et al. (2018, eLife) proved the 5-HT2A receptor is necessary and sufficient: the antagonist ketanserin completely blocked all subjective and neural effects of LSD. Every hallucinatory, ego-dissolving, and perceptual effect routes through this single receptor.

What LSD does to neural networks

Carhart-Harris et al. (2016, PNAS) administered 75 micrograms of LSD intravenously to 20 healthy volunteers and performed multimodal neuroimaging (fMRI, arterial spin labeling, MEG). The findings:

Default Mode Network disintegration. The DMN — the network active during self-referential thought, daydreaming, and the sense of being a continuous self — lost internal coherence. Connectivity within the DMN decreased. The degree of DMN disintegration correlated with subjective ego dissolution (r = 0.49). A specific disconnection between the parahippocampal cortex and the retrosplenial cortex — two regions that anchor the self in space and memory — correlated with ego dissolution at r = 0.73.

Network desegregation. The brain normally operates through specialized, segregated networks. Under LSD, boundaries between these networks dissolved. Connectivity increased between networks that normally don’t talk to each other — visual and parietal, auditory and parietal, DMN and salience. The brain shifted from compartmentalized processing to a more integrated, less differentiated state.

Visual cortex activation. Cerebral blood flow increased in V1 (primary visual cortex). V1 connectivity expanded to numerous cortical regions. Blood flow increases in V1 correlated with complex visual imagery ratings (r = 0.64). This is the mechanism of hallucination: the visual cortex, normally constrained by thalamic gating, generates its own activity when LSD opens the gates.

Thalamic gating disruption. The thalamus normally filters sensory information before it reaches cortex — it decides what you consciously process. LSD, acting through 5-HT2A receptors, reduces thalamic filtering via the cortico-striato-thalamo-cortical (CSTC) circuit. The gates open. Cortex is flooded with unfiltered sensory and associative information.

Carhart-Harris called this the “entropic brain” — a state of elevated neural entropy where the brain operates near criticality, the transition zone between order and disorder. More possible states. More sensitivity to initial conditions. Less predictable.

Dose-response: from wonder to terror

Holze et al. (2021, Neuropsychopharmacology) conducted the first controlled dose-response study in 16 healthy subjects at 25, 50, 100, and 200 micrograms:

Measure	25 μg	50 μg	100 μg	200 μg
”Good drug effects” >50%	37%	91%	96%	91%
“Bad drug effects” >50%	0%	9%	27%	31%
Ego dissolution (mean)	3.4%	13%	20%	22%
Duration (hours)	6.7	—	—	11.0

The critical finding: good effects ceiling at 100 μg, but ego dissolution and anxiety continue to increase at 200 μg. Higher doses don’t produce more pleasure. They produce more disintegration of the self while bad effects climb.

Physiological effects (Holze et al., 2022, Psychopharmacology, pooled analysis of 83 subjects):

At 200 μg: systolic BP rose to 142 mmHg (from 129 baseline), heart rate to 90 bpm, body temperature to 37.8°C. LSD independently elevated cortisol, corticosterone, prolactin, oxytocin, and epinephrine — it activates the stress axis even in a calm, supportive clinical setting.

What happened at Lexington: 77 days

The Isbell experiments at the Addiction Research Center in Lexington, Kentucky, administered LSD daily to seven incarcerated Black men for 77 consecutive days at 1.55 μg/kg (approximately 140 μg for a 90 kg person), escalating to quadruple doses (~600 μg).

Tolerance. Isbell’s own data shows 47% decrease in responses by day 2, near-maximal tolerance by day 4. Subjects “simply read and watched TV normally.” Even quadruple doses had minimal effect after tolerance developed.

The mechanism: Daily LSD produces 5-HT2A receptor downregulation in frontal cortex (Buckholtz et al., 1990). The receptors are internalized and degraded — the brain literally removes the binding sites. Tolerance reverses after 3 drug-free days. No physical withdrawal occurs.

But the first days were the window of maximum harm. Before tolerance developed, subjects experienced escalating doses against a naive serotonergic system — maximum DMN disruption, maximum ego dissolution, maximum thalamic gating failure — in a coercive institutional environment with no psychological preparation, no autonomy, and the knowledge that their payment was the heroin their bodies craved.

This matters because of set and setting.

Set and setting is neuroscience, not philosophy

Mueller et al. (2017, Translational Psychiatry) showed that LSD (100 μg) reduced amygdala reactivity to fearful faces in a supportive clinical environment — the greater the drug effect, the greater the fear dampening. In a safe room, LSD makes you less afraid.

But the mechanism is bidirectional. LSD doesn’t suppress fear. It removes the brain’s filtering and compartmentalization systems. The thalamic gates open. The DMN boundaries dissolve. The self becomes permeable to the environment.

In a calm room, this produces wonder. In a coercive room, it produces maximal vulnerability.

The compounding works through four identified pathways:

Thalamic gating disruption floods cortex with unfiltered sensory input. In an interrogation room, every threat signal arrives at full intensity.
DMN dissolution eliminates the psychological boundary between self and environment. Ego dissolution in a supportive context is mystical experience. In a coercive context, it is identity disintegration — you cannot psychologically distance yourself from what is being done to you.
HPA axis potentiation. LSD independently elevates cortisol. Environmental stress also elevates cortisol. The two compound. Cortisol potentiates amygdala reactivity while LSD simultaneously removes the prefrontal regulatory capacity to manage it.
Entropic sensitivity. The brain at elevated entropy is maximally sensitive to initial conditions. Small environmental inputs produce disproportionate psychological effects. A kind voice becomes profoundly reassuring. A threatening voice becomes profoundly destabilizing.

The MKUltra protocols at Lexington weaponized every one of these mechanisms: coercive environment, incarcerated subjects with no autonomy, opiate-dependent subjects paid in their drug of addiction, escalating doses before tolerance could develop, and 77 consecutive days ensuring chronic neuroendocrine stress even after pharmacological tolerance to LSD itself.

The other drugs

MKUltra used far more than LSD. Each compound attacked a different system.

Scopolamine — one of the original “truth serums.” Anticholinergic: blocks muscarinic M1 receptors in the hippocampus, the receptors required for new memory formation. Produces a twilight state with anterograde amnesia. The subject can talk but cannot remember talking. The problem for interrogation: it also impairs coherent thought and promotes confabulation. The subject will say things, but the things may not be true.

Barbiturate-amphetamine combination — simultaneous IV barbiturate into one arm and amphetamine into the other. Barbiturates are GABA-A positive allosteric modulators: they enhance inhibitory chloride conductance, suppressing cortical function. Amphetamines flood synapses with catecholamines (dopamine, norepinephrine), activating arousal circuits. The simultaneous depression and stimulation of the CNS produces a state where the subject is cognitively impaired (barbiturate) but physically activated (amphetamine) — “babbling incoherently,” per the documents. Sometimes useful answers could be extracted. Often not.

Sodium pentothal (thiopental) — the most famous “truth serum.” Ultra-short-acting barbiturate, GABA-A agonist. Produces disinhibition and drowsiness. The belief was that lowered inhibitions would produce truthful disclosure. The science says otherwise: under thiopental, subjects become more suggestible and confabulate more freely. They do not become more truthful. A 2004 review in the Indian Journal of Anaesthesia concluded it is “not a foolproof method of interrogation” because subjects can still “willfully lie.”

BZ (3-quinuclidinyl benzilate) — the Army’s “super-hallucinogen” tested at Edgewood Arsenal. An anticholinergic: blocks muscarinic acetylcholine receptors throughout the brain. Unlike LSD (which works through serotonin and produces perceptual distortion while consciousness remains), BZ produces true delirium — complete loss of contact with reality lasting 3-4 days. Subjects experience vivid hallucinations indistinguishable from reality, hyperthermia, tachycardia, urinary retention, and total disorientation. The Army tested BZ on thousands of soldiers at Edgewood Arsenal from the mid-1950s through the 1970s.

Chlorpromazine (Thorazine) — used by Cameron for prolonged drug-induced comas. A first-generation antipsychotic: D2 receptor antagonist with antihistaminic and anticholinergic properties. At sedating doses maintained for weeks to months, it produces sustained dopamine blockade → post-withdrawal dopaminergic supersensitivity → risk of tardive dyskinesia (involuntary movements from irreversible D2 receptor changes) and psychotic rebound. Extended use also causes metabolic disruption, weight gain, and cardiac conduction changes.

Why “truth serums” don’t work

Amy Arnsten’s research at Yale (2009, Nature Reviews Neuroscience) provides the molecular explanation for why coercive interrogation with pharmacological agents produces unreliable information.

Under uncontrollable stress — which includes being drugged without consent in an interrogation setting — high levels of norepinephrine and dopamine rapidly weaken prefrontal cortex function. The specific mechanism: dopamine D1 receptor overstimulation activates cAMP signaling, opening HCN channels that shunt excitation away from prefrontal neurons. Norepinephrine alpha-1 stimulation activates PLC cascades causing further excitation loss through K+ channels.

The result: working memory failure, loss of executive function, inability to organize and retrieve complex information. The subject becomes less capable of accurate recall, not more compliant with truthful disclosure. Even mild acute uncontrollable stress causes “rapid and dramatic loss of prefrontal cognitive abilities.” More prolonged stress causes architectural changes in prefrontal dendrites.

Every MKUltra pharmacological approach — LSD dissolving cognitive boundaries, barbiturates suppressing cortical function, anticholinergics blocking memory formation, stress hormones impairing prefrontal retrieval — attacked the very neural systems required to produce the accurate information the CIA wanted. The drugs made the brain less able to tell the truth, not more.

The program’s stated objective was defeated by the pharmacology of its own methods.

— Cael