Pl. IX

Part IV · Inner Life, Sentience & Ethics · Chapter 9

Sleep, Two-Stage Sleep, and Possible Dreaming in Octopuses

Sleep in octopuses was first established behaviorally, then—remarkably—shown to have a two-stage architecture rivaling the vertebrate distinction between non-REM and REM sleep. The foundational work (Brown et al., 2006; Meisel et al., 2011) demonstrated that Octopus vulgaris meets the classical behavioral criteria for sleep: a reversible quiescent state with reduced activity, elevated arousal thresholds during quiet periods, and homeostatic rebound when quiescence is prevented. This satisfied the three-part behavioral definition of sleep (raised arousal threshold, rapid reversibility, and rebound after deprivation) without any need for a vertebrate-style cortex.

The two-stage discovery came from Sylvia Medeiros, Sidarta Ribeiro, and colleagues at the Brain Institute (UFRN, Brazil), published as Medeiros/Sampaio et al. (2021, iScience). Video-recording four adult Octopus insularis, they quantified a cyclic alternation between "quiet sleep" (QS)—pale, uniform skin, closed slit-pupils, stillness, median episode ≈415 s (≈6.9 min)—and a briefer "active sleep" (AS) phase (median ≈40.8 s) of dynamic chromatophore-driven skin color and texture changes, rapid eye movements, sucker and mantle twitches, and quickened breathing. Critically, they demonstrated AS is true sleep, not waking: arousal-threshold tests using visual stimuli showed the longest response latencies and highest non-response rates during AS (median ≈32 s vs. ≈6 s in QS and ≈4 s alert; p ≈ 2.2×10⁻¹⁶). AS almost always followed a long bout of QS (≈82% of the time) and recurred with a periodicity clustered around 30–40 minutes. The behavioral resemblance to REM—twitches, eye movements, high arousal threshold—prompted cautious speculation that octopuses might experience something dream-like, while the authors stressed the cephalopod and vertebrate brains are non-homologous, making this convergent rather than inherited.

The decisive neural evidence came from Aditi Pophale, Sam Reiter, and colleagues at the Okinawa Institute of Science and Technology (OIST), published in Nature (Pophale et al., 2023). Using Octopus laqueus and Neuropixels probes in the central (supra-oesophageal) brain, with brain regions localized via CUBIC tissue clearing and light-sheet microscopy registered to a 3D atlas, they found that local field potential (LFP) activity during AS resembles waking. The superior frontal and vertical lobes—regions homologous in function to learning/memory circuits—showed the strongest AS activity, with the superior frontal lobe generating prominent ≈30 Hz oscillations and the vertical lobe producing large (≈700 μV) low-frequency waveforms; waking and AS LFP correlated strongly (Pearson R ≈ 0.74 low-frequency, 0.95 high-frequency). Strikingly, during QS the superior frontal lobe produced 12–18 Hz oscillatory events lasting up to ≈1 s that resemble mammalian sleep spindles in frequency and duration—often with no visible behavioral correlate.

Pophale et al. also refined the timing: AS bouts recurred roughly hourly, each lasting ≈60 s (75 ± 28 s), interleaved with QS bouts (≈50 min); over a night at 22°C animals showed ≈10 active bouts and ≈12 QS bouts, with the interval temperature-dependent (a 1°C rise shortened intervals by ≈5 min). Their sleep-deprivation experiment is the strongest functional evidence: keeping animals awake ≈2 days produced a significant rebound—more frequent active bouts on the two following nights (P = 0.0065 and P = 0.0216, Wilcoxon)—establishing AS as a homeostatically regulated, essential sleep stage, not incidental restlessness.

The most provocative finding concerns possible dreaming. Using 8K video and a VGG-19 neural-network analysis of skin patterning, the team found AS skin dynamics rapidly cycle through the very same patterns octopuses deploy while awake and camouflaging—suggesting a "replay" or offline refinement of skin-pattern motor control, loosely analogous to rodent hippocampal replay or the structured head-direction activity seen during mammalian REM. Whether this constitutes dreaming remains unknowable; as Reiter and others emphasize, "we cannot ask the octopus."

Context from cuttlefish reinforces convergence: Frank et al. (2012) first reported a REM-like state in senescing Sepia officinalis, and Iglesias, Frank et al. (2019, J. Exp. Biol.) documented its cyclic nature (≈2.4 min REM-like bouts alternating with ≈34 min quiescence). A separate biorxiv preprint (Medeiros et al., 2023) reported bizarre "abnormal behavioral episodes" in senescing O. insularis—possible parasomnias or "nightmares"—though this remains anecdotal and unreviewed. The overarching interpretation: octopuses and mammals, whose last common ancestor lived ≈550 million years ago, independently evolved two-stage sleep, implying the biphasic architecture serves something deeply fundamental.

Striking / counterintuitive:

• Active-sleep LFP brain activity in the octopus is nearly indistinguishable from waking activity (Pearson R up to 0.95 in high-frequency bands), yet arousal thresholds are highest during this stage.
• Quiet sleep contains 12–18 Hz oscillatory events lasting ≈1 s that resemble mammalian sleep spindles—strikingly convergent given ≈550 million years of separate evolution and no homologous brain structures.
• During active sleep, octopuses rapidly cycle their skin through the exact camouflage patterns they use while awake, hinting at offline 'replay' or rehearsal of skin-pattern control.
• Sleep-deprived octopuses rebound specifically by increasing active-sleep bouts, proving the REM-like stage is homeostatically defended and essential, not incidental.
• Active-sleep interval timing is temperature-dependent: a 1°C rise shortens the cycle by roughly 5 minutes.
• A preprint reports bizarre 'nightmare-like' abnormal episodes in senescing octopuses, though this is anecdotal.

Open questions:

• Do octopuses subjectively experience anything dream-like during active sleep, or is the skin-pattern 'replay' purely offline motor maintenance with no phenomenology? This is likely unanswerable with current methods.
• Is the skin-pattern cycling during active sleep true memory replay (like hippocampal replay) or stochastic churn through the motor repertoire?
• What is the mechanistic function of the sleep-spindle-like 12–18 Hz waveforms in quiet sleep—memory consolidation, as in mammals?
• Given non-homologous brains, is two-stage sleep a case of deep convergent evolution driven by shared computational constraints, or independent solutions that merely look alike?
• Do young, healthy octopuses show the same 'abnormal'/parasomnia-like episodes, or are those artifacts of senescence and captivity?
• How generalizable are findings across octopus species (O. insularis, O. laqueus, O. vulgaris) and to other coleoid cephalopods?

Key researchers/labs: Sam Reiter (Computational Neuroethology Unit, OIST), Aditi Pophale (OIST, co-first author), Sylvia Medeiros (Brain Institute, UFRN Brazil), Sidarta Ribeiro (Brain Institute, UFRN), Marcos G. Frank (Washington State University; cuttlefish REM-like sleep), Teresa L. Iglesias, Ruth A. Byrne / David Meisel (early octopus sleep behavior), E. R. Brown (Stazione Zoologica Naples).

Key papers #

• Aditi Pophale, Kazumichi Shimizu, Tomoyuki Mano, Leenoy Meshulam, Sam Reiter, et al. (2023). Wake-like skin patterning and neural activity during octopus sleep. Nature — Neuropixels recordings show active-sleep brain activity resembles waking and QS contains sleep-spindle-like waveforms; deprivation triggers rebound, and skin patterns replay waking camouflage—key evidence for a REM-like, possibly dream-like stage.
• Sylvia L. de S. Medeiros, Mizziara M. M. de Paiva, Paulo H. Lopes, Sidarta Ribeiro, et al. (Sampaio group) (2021). Cyclic alternation of quiet and active sleep states in the octopus. iScience — First to define two-stage octopus sleep (quiet vs active) in O. insularis, with arousal-threshold data proving active sleep is genuine sleep and speculation on dreaming.
• Marcos G. Frank, R. J. Waldrop, M. Dumoulin, et al. (2012). A preliminary analysis of sleep-like states in the cuttlefish Sepia officinalis. PLOS ONE — First report of a REM-like state in a cephalopod (senescing cuttlefish), establishing the cross-cephalopod precedent for REM-like sleep.
• Teresa L. Iglesias, Marcos G. Frank, et al. (2019). Cyclic nature of the REM sleep-like state in the cuttlefish Sepia officinalis. Journal of Experimental Biology — Documented cyclic ≈2.4 min REM-like bouts alternating with ≈34 min quiescence, showing periodic two-stage structure in cuttlefish.
• David M. Meisel, Ruth A. Byrne, et al. (2011). Contribution of the visual system of the octopus to determination of sleep-like behavior. Journal of Experimental Marine Biology — Helped establish behavioral sleep criteria (arousal threshold, reversibility) in Octopus vulgaris.
• E. R. Brown, S. Piscopo, R. De Stefano, A. Giuditta (2006). Brain and behavioural evidence for rest-activity cycles in Octopus vulgaris. Behavioural Brain Research — Foundational demonstration of circadian rest-activity cycles and quiescent behavioral sleep in octopus.