A Model for the Application of Target-Controlled Intravenous Infusion for a Prolonged Immersive DMT Psychedelic Experience

N,N-dimethyltryptamine (DMT) produces rapid, profound and highly reproducible alterations of consciousness, commonly described as the replacement of ordinary awareness by an ‘‘alternate universe’’ populated by complex visual objects and apparent sentient ‘‘beings.’’ The compound is endogenously produced in humans, crosses the blood–brain barrier in animals, and is metabolised rapidly by monoamine oxidase (MAO) enzymes; these pharmacological features, together with the striking phenomenology, make DMT of considerable interest for psychology and neuroscience. Despite this, detailed phenomenological characterisation has been limited in part because the effects of inhaled or intravenous bolus DMT are very brief, typically resolving within 20–30 min, which hampers prolonged observation and functional neuroimaging during the state. Gallimore and Strassman set out to determine whether target-controlled intravenous infusion—a methodology from anaesthesia used to achieve and maintain a stable brain concentration of drug—could be applied to DMT. Using time-series plasma DMT data from prior human dosing sessions, the investigators developed and simulated pharmacokinetic (PK) and effect-site models to test whether a controlled infusion protocol could produce a sustained, stable DMT effect-site concentration, and to derive candidate infusion parameters for research use.

The investigators adapted the principles of target-controlled infusion: constructing a compartmental pharmacokinetic model that describes central (plasma), peripheral tissue, and effect-site (brain) compartments, and using those parameters to calculate an initial bolus and time-varying infusion rates that compensate for elimination and inter-compartment transfer. They explain the general two-compartment formalism (central and peripheral compartments with rate constants k12, k21 and elimination k10) and incorporate a first-order transfer to the effect site (k1e) and its elimination (ke0). Because enzymatic metabolism by MAO-A dominates DMT clearance, the model includes Michaelis–Menten (non-linear) elimination. Empirical plasma data were taken from a previously published human study in which ‘‘fully psychedelic’’ bolus doses of DMT fumarate (0.2 or 0.4 mg/kg) were administered intravenously over 30 s with a 15 s saline flush, and venous blood samples were collected before dosing and at 2, 5, 10, 15, 30 and 60 min after the end of infusion. Plasma DMT concentrations were measured by gas chromatography–mass spectrometry (GC–MS). The extracted text states that a total of nine subjects were used in the analysis and refers to ‘‘9 sets each of 0.4 and 0.2 mg/kg time series data;’’ the extraction does not clearly report whether each of the nine subjects contributed both dose series or how the datasets were paired. For model fitting the authors used Matlab’s Simbiology toolkit. They fitted one-, two- and three-compartment models to the 0.2 and 0.4 mg/kg time series separately, and report that both dose datasets were best fitted by a two-compartment model with enzymatic (Michaelis–Menten) clearance. Parameter estimates were obtained from naïve averaged concentration–time curves (mean plasma concentration at each time point across subjects); the authors note that inter-subject variability was substantial and that non-linear mixed effects methods can give more robust population estimates and identify covariates. Having obtained PK parameters, the researchers extended the model to include an effect-site compartment, tuning k1e and ke0 so that the simulated peak effect-site concentration matched clinical observations (peak subjective effects at ≈3 min). They then used the bolus–elimination–transfer (B.E.T.) approach commonly applied in anaesthesia to derive an initial bolus and a minute-by-minute, time-decreasing infusion rate that compensates for early rapid peripheral transfer and approaches a steady-state maintenance infusion once peripheral equilibration has decayed.

Fitting the plasma time-series data yielded a two-compartment pharmacokinetic model with enzymatic (Michaelis–Menten) clearance, consistent with rapid MAO-A metabolism of DMT. The mean plasma concentration–time profiles varied considerably between subjects; the extracted text reports mean coefficients of variation in plasma DMT concentration between 2 and 30 min of 83% for the 0.2 mg/kg group and 58% for the 0.4 mg/kg group. Despite this plasma variability, variability in the subjective response (measured with the Hallucinogen Rating Scale) was lower, with coefficients of variation of 46% and 32% for the 0.2 and 0.4 mg/kg groups respectively. Simulations of the bolus 0.4 mg/kg infusion reproduced clinical observations: peak effect-site concentrations were reached at approximately 3 min after infusion start, consistent with autonomic and behavioural peak responses recorded between 2 and 5 min. The authors interpret an effect-site concentration of roughly 60 ng/ml as the ‘‘breakthrough’’ threshold for entry into the full DMT state, with peak subjective intensity occurring at concentrations just over 100 ng/ml; subjects typically remained in the immersed state for 6–7 min following breakthrough in the bolus paradigm. Using the parameter set from the 0.4 mg/kg data the team calculated infusion rates required to reach and maintain a target effect-site concentration of ≈100 ng/ml for a 75 kg simulated subject. Key quantitative findings from these simulations include: a steady-state maintenance infusion rate (Rss) of approximately 0.93 mg/min; an early peripheral transfer peak of about 3.3 mg/min at ~2.3 min that must be compensated for by a higher initial infusion rate; and the need to taper the infusion to avoid overshoot as peripheral compartments equilibrate. The investigators contrasted their model with an earlier infusion protocol (0.3 mg/kg bolus followed by 0.02 mg/kg/min infusion) reported in the literature; simulations of that regimen for a 75 kg subject showed a continued rise of effect-site concentration to ~150 ng/ml by session end, an outcome the authors suggest would produce excessively intense and potentially distressing effects. As a practical candidate protocol, the authors present simulations in a 75 kg subject using an initial bolus of 25 mg over 30 s, an infusion starting at 2 min at 4.2 mg/min with minute-by-minute updates decreasing according to peripheral transfer decay, and transition to a constant maintenance infusion of 0.93 mg/min once steady state is reached after ~20 min. In this simulation plasma concentration briefly spikes above 200 ng/ml while the effect-site concentration approaches and holds near 100 ng/ml with minimal overshoot. The extracted text also notes behavioural observations from the original dataset (e.g. deep rhythmic breathing, REM-like eye movements, impaired memory for measurements at peak vs later time points) that align with the simulated time course. Ethical approvals and witnessed written informed consent for the original blood-sampling study are reported in the extracted material.

Gallimore and colleagues interpret their modelling work as a proof-of-principle that DMT’s pharmacokinetic and pharmacodynamic properties—rapid onset, short duration, and lack of acute tolerance to subjective effects—make it amenable to target-controlled intravenous infusion. They argue that a PK-informed infusion protocol could reliably induce, stabilise, and titrate the depth of the DMT state, enabling prolonged phenomenological observation, safer control of intensity (including down-titration to allow communication), and compatibility with functional neuroimaging protocols that require extended, steady-state drug effects. The authors position this approach as complementary to, and distinct from, the prolonged DMT-like state produced by ayahuasca: while ayahuasca yields multi-hour effects, its oral activation requires MAO inhibition and includes additional psychoactive constituents, so it cannot provide a precisely controlled pure-DMT effect-site concentration. They also suggest potential clinical uses, proposing that a controllable infusion could be ‘‘titrated’’ in psychotherapy to adjust duration and intensity for processes such as trauma re-exposure, and note that brief inter-dose lucid periods in prior repeated-bolus DMT studies facilitated therapeutic dialogue and progressive thematic development across doses. Key limitations the authors acknowledge include the need for more extensive pharmacokinetic sampling and more sophisticated population-level modelling to define definitive parameters, quantify inter- and intra-subject variability, and identify covariates (for example weight, age, gender, and liver function) that influence dose–concentration relationships. They note the large plasma concentration variability observed in their dataset (mean coefficients of variation 83% and 58% for the two dose groups) but emphasise that such variability is comparable to that seen for anaesthetic agents and that subjective-response variability was substantially lower, suggesting a tolerable target window. The authors also emphasise that infusion protocols typically undergo iterative refinement as models and empirical data accumulate. In conclusion, the investigators present target-controlled infusion of DMT as a feasible research tool that could advance the phenomenological characterisation of the DMT state, support neuroimaging studies, and potentially enable novel, titratable psychotherapeutic applications; they recommend further empirical work to refine PK/PD parameters and to validate safety and tolerability in controlled human studies.