Correlation between the potency of hallucinogens in the mouse head-twitch response assay and their behavioral and subjective effects in other species

Halberstadt and colleagues outline the pharmacological diversity of classical serotonergic hallucinogens, noting major structural families (ergolines, phenylalkylamines, tryptamines) and differences in receptor selectivity. Earlier research established that 5-HT2A receptor agonism is central to hallucinogen effects in humans and that animal models have been useful for probing neurochemical mechanisms. However, the behavioural complexity of human psychedelic effects has made it challenging to develop animal assays with both construct and predictive validity for potency. Against this background, the investigators evaluated whether the mouse head-twitch response (HTR) — a rapid, high-frequency head rotation elicited by 5-HT2A activation — can provide a quantitative measure of hallucinogen potency that translates across species. Specifically, they aimed to test whether ED50 values derived from the mouse HTR correlate with reported behavioural and subjective potency measures in rats (drug-discrimination studies) and in humans, thereby assessing the HTR assay's utility for structure–activity relationship (SAR) studies and cross-species prediction of hallucinogenic potency.

Male C57BL/6J mice (6–8 weeks old) were used for all experiments. Animals were group-housed in an AAALAC‑approved vivarium on a reversed light cycle, with testing conducted during the light phase. All procedures complied with NIH guidelines and institutional animal care approvals. A wide panel of hallucinogens from phenylalkylamine and tryptamine classes (including DOM, DOET, DOBU, DON, DOC, R-(-)-DOI, MDA, R-(-)-MDA, α-Et-2C-D, 25D-NBOMe, 25I-NBOH, DMT, DET, DPT, DIPT, 5-MeO-AMT and others) were obtained from established suppliers and authenticated by mass spectrometry and NMR. Test substances were dissolved in isotonic saline and administered intraperitoneally at 5 mL/kg. HTR measurement used an electronic detection method: a small neodymium magnet was surgically affixed to each mouse’s skull and, after a two‑week recovery, animals were placed in a coil-equipped glass cylinder immediately after drug injection. Magnetometer signals were recorded for 30 minutes, digitally filtered and inspected by trained personnel using predefined waveform, frequency (≥ 40 Hz), amplitude and duration (< 0.15 s) criteria to count head twitches. Experiments employed a between-subjects design with pseudorandom group assignment and at least 7 days between sessions to minimise carryover; mice were tested repeatedly over months, and prior observations indicate no tolerance with weekly dosing. Statistical analysis comprised one-way ANOVA for head-twitch counts with Dunnett’s post hoc tests when applicable. ED50 values and 95% confidence intervals were estimated by nonlinear regression. Correlations between HTR-derived ED50 values and potency measures in rats and humans were assessed using linear regression and Pearson correlation, with analyses performed in Prism 7.00.

Across the expanded dataset (41 compounds tested in HTR experiments), hallucinogens from phenylalkylamine and tryptamine classes reliably induced head twitches in C57BL/6J mice. Dose‑response curves were often non‑monotonic, showing an ascending phase and a marked decrement at high doses. To assess translational validity, the study compared HTR-derived ED50 values (moles/kg) with reported human hallucinogenic potencies (total dose in moles) and with ED50 values from rat drug-discrimination studies. Human potency estimates were compiled from clinical studies, published accounts and archival sources; where reported as ranges the mean of the range was converted to molar dose. For 36 hallucinogens, HTR ED50 values correlated strongly with reported human potency: Pearson r = 0.9448 (95% CI: 0.8937–0.9717), p < 0.0001. HTR potencies were also highly correlated with drug-discrimination ED50 values from rats trained to discriminate DOM (n = 21): r = 0.9564 (95% CI: 0.8937–0.9825), p < 0.0001. Several drugs showed nearly identical potency estimates across assays; for example, LSD ED50 in the mouse HTR was 52.9 μg/kg (132.8 nmol/kg) versus 52 μg/kg (130.5 nmol/kg) in DOM-trained rat substitution, and DOM ED50 values in HTR and drug-discrimination were 1.75 μmol/kg and 1.79 μmol/kg, respectively. The authors quantified similarity between HTR and drug-discrimination potencies using a normalized difference score. For 21 compounds in DOM-trained rats the mean normalized difference was 31.6 ± 5.7% (range 1.8%–78.6%), whereas for 16 compounds in LSD-trained rats it was 330.2 ± 94.4% (range 22.7%–1,215.4%). Restricting analysis to nine compounds present in both drug-discrimination datasets yielded lower difference scores for DOM-trained rats (27.3 ± 9.4%) than for LSD-trained rats (99.8 ± 25.7%; t16 = 2.608, p = 0.019). Correlations between HTR and drug-discrimination ED50s for the nine-compound subset were r = 0.9544 (95% CI: 0.7926–0.9906) with LSD training and r = 0.975 (95% CI: 0.8819–0.9949) with DOM training; LSD and DOM drug-discrimination ED50s were themselves highly correlated (r = 0.9722). The dataset included first-time mouse HTR characterisation for several compounds (DOET, DOBU, DON, DOC, α-Et-2C-D, MDA, R-(-)-MDA, 25D-NBOMe, 25I-NBOH, and DET). Some agents were excluded from the human potency correlation because effective dose ranges in humans were uncertain (for example DIPT, DON and α-Et-2C-D). The authors note that the HTR ED50 versus human potency regression produced a high coefficient of determination (reported in the Discussion as r2 = 0.8927 for the full set and r2 = 0.9602 for a 15‑compound subset), indicating a tight linear relationship despite heterogeneity in chemical classes and human dosage data.

Halberstadt and colleagues interpret their findings as evidence that the mouse HTR assay, while traditionally used qualitatively, can yield robust quantitative potency estimates that translate across species. After expanding their HTR dataset to 41 compounds chosen for structural diversity and availability of comparative data, they found strong positive correlations between HTR ED50 values, human subjective potency estimates and rat drug-discrimination ED50s. Because HTR, drug-discrimination cues and human psychedelic effects are all linked to 5-HT2A receptor activation, the authors argue that these results support substantial construct validity for the paradigms and predictive validity for the HTR assay. The investigators acknowledge several limitations and sources of uncertainty. Human potency estimates are imperfect: many are reported as dose ranges from heterogeneous sources, routes of administration differ (oral versus sublingual or parenteral), and subjective measures are inherently variable. The HTR dose–response relationship is often non‑monotonic and baseline spontaneous head-twitch rate varies, which complicates curve fitting and potency estimation. Moreover, the HTR can be susceptible to false‑positive induction by non‑5-HT2A mechanisms (examples cited include rolipram, enkephalins and agents that increase serotonin release), although the compounds profiled here largely have known 5-HT2A agonist activity and appropriate receptor affinities. Pharmacological complexity is further discussed: interactions with 5-HT1A and 5-HT2C receptors can modulate HTR expression and potency (for instance, 5-HT1A activity can dampen 5-HT2A-mediated effects), and these influences appear to be similar across mice and humans. Strain differences in mouse sensitivity to 5-HT2A agonists are noted, so the generalisability of results obtained in C57BL/6J mice to other strains remains to be tested. The authors also highlight that equivalence between HTR and drug-discrimination potencies may be partly driven by coincidence between HTR peak dose ranges and the training doses commonly used in discrimination studies. Looking ahead, the paper suggests follow-up work to establish whether HTR-derived regression models can prospectively predict human potency for novel 5-HT2A agonists, and whether class-specific models (tryptamines, ergolines, phenylalkylamines) might perform better than a single unified model because of differences in distribution, metabolism and off‑target interactions. The authors emphasise that the HTR assay does not model subjective psychedelic experience per se, but conclude that HTR potency measures are relevant and translatable to other species and useful for SAR and mechanistic studies, while calling for additional investigations to resolve unanswered pharmacological questions (for example, why certain 5-HT2A agonists such as lisuride fail to elicit HTR).