Case file · ketamine · a TNU animation · Joe Christensen

The Ketamine Case

One forty-minute infusion, and within a day most patients with treatment-resistant depression had responded. How can one molecule reach a mood within a day, and why does the same molecule lift mood, bend perception and, at a higher dose, switch off consciousness? An investigation from a person down to one trapped molecule, and back.

8:01 · narrated · for Translational Neuropharmacology (MSc) · Translational Neuropsychiatry Unit, Aarhus University

What's real in it

  • Anatomy and cells come from real data: the MNI152 MRI template (an average of 152 adults), the BigBrain histology specimen at 1 µm, human cortex from the H01 electron-microscopy dataset, and the cryo-EM structure of the NMDA receptor with ketamine in its pore (PDB 7EU7).
  • Every number on screen is traced to its paper, figure or table, with species, preparation and dose. Mouse and rat findings are labelled as such.
  • Derived values are flagged: the "% of channels blocked" is a simple binding estimate from plasma levels and IC₅₀, not a measured occupancy; the time bands in the alibi are schematic.
  • The people are AI-generated illustrations, not trial participants. Cell types on H01 neurons are illustrative stand-ins.

The case, in short

  • Clue 1, the dose: an antidepressant dose blocks only a few percent of NMDA channels; anaesthesia needs ~20× the plasma level. Yet low doses raise prefrontal glutamate: a blocker that excites.
  • Clue 2, who is blocked first: fast-firing interneurons, the brakes. Block the brakes, and excitation rises (still debated: SST or PV first).
  • Clue 3, lights out: at high dose every neuron is blocked.
  • Clue 4, the alibi: by day one, plasma ketamine is ~1% of its peak, yet the mood effect remains.
  • The suspects: new synapses, a metabolite (HNK), opioid signalling, the lateral habenula, and a protein-synthesis brake. The case is still open.

Sources

History: Domino et al. 1965, Clin Pharmacol Ther (CI-581 in man). Patients: Zarate et al. 2006, Arch Gen Psychiatry · Murrough et al. 2013, Am J Psychiatry · Luckenbaugh et al. 2014, J Affect Disord · Auerbach et al. 2018, J Abnorm Psychol · Satinsky et al. 2021 · Rotenstein et al. 2016, JAMA. Brain and synapse: Abdallah et al. 2017 · Fischl & Dale 2000, PNAS · Amunts et al. 2013, Science (BigBrain) · Shapson-Coe et al. 2024, Science (H01) · Goncalves et al. 2020, PNAS · Zhang et al. 2021, Nature (7EU7). Dose and circuits: Moaddel et al. 2023 · Kotermanski & Johnson 2009, J Neurosci · Moghaddam et al. 1997, J Neurosci · Homayoun & Moghaddam 2007, J Neurosci · Gentet et al. 2010, Neuron · Garst-Orozco et al. 2020 · Ali et al. 2020, Nat Commun · Cichon et al. 2022, Nat Neurosci · Chen et al. 2009, J Neurosci. Suspects: Li et al. 2010, Science · Moda-Sava et al. 2019, Science · Zanos et al. 2016, Nature · Williams et al. 2018, Am J Psychiatry · Jelen et al. 2025 · Yoon et al. 2025 · Yang et al. 2018, Nature · Autry et al. 2011, Nature · Holmes et al. 2022 (SV2A PET).

Narration: ElevenLabs v4 (voice: Elizabeth). Music and sound: ElevenLabs. Rendered from real data with a custom WebGL pipeline. Data licences: BigBrain CC BY-NC-SA, H01 CC BY 4.0, PDB CC0; figures shown are CC BY with credit, others re-plotted or typeset.