The subjects of this experiment were five laboratory-reared adult/senescent Sepia officinalis (15–19 months post-hatching), housed in a round tank (823 L) containing a variety of plastic plants. For the experiment, individual cuttlefish were transferred from group housing in a large (approximately 10–12 foot diameter pool) to a rectangular experimental tank (240 L). A sleeping site was prepared using a submerged clear plexiglass “V,” with the bottom point of the “V” running parallel with the length of the tank. Below the plexiglass “V” and within the narrow base of the “V” was a layer of crushed oyster shells. At each end of the “V,” two large stones were placed to create a single central resting site. Behind the tank, a few plastic plants and a large piece of black plastic served as a backdrop. A high-resolution, black and white camera (Sanyo VCB-3384, 30 frames/sec) was positioned directly above the sleeping site, and focused at a plane that maximally captured chromatophore activity on the dorsal surface of each animal. A second camera was placed facing the side of the tank.

At 8:30 AM on day 1 of the experiment, one cuttlefish was placed into the experimental tank for 2 days of acclimation. A short period of acclimation was used because these animals were very close to the end of their lifespan (most died within a few months after we began our experiments). It was fed 1 live fiddler crab (Uca spp); immediately thereafter, a black plastic tarp was wrapped around the entire experimental area to minimize human interaction. On the morning of day 3, the two cameras were turned on to run continuously for 48 hours. Frame data were captured on a PC with a GV-2004 video card and analyzed using accompanying software. During each minute of recording, the following behaviors were scored: ‘inactive’ (“0” buried in or lying motionless on the gravel bed) or active (“1” hovering or swimming in the water column); eyes: closed (0) or open (1); and fin movement: still (0) or moving (1). However, as the inactive/active scoring was as informative as any other single measure or combination of measures, it was used as our principle measure of rest and activity. When sleep-like states with chromatophore activation (CA; hereafter referred to as ‘sleep-like states+CA’) were observed, changes in individual chromatophore pattern components as defined by Hanlon et al [20] were tabulated second-by-second as either present or not present.

All adult cuttlefish exhibited periods of quiescence where they rested against, or buried themselves in substrate on the bottom of the aquarium (Figure 1, Table 1). If provided enough substrate, the animals would completely bury themselves, leaving only their eyes above the gravel. This state was rapidly reversible as the animals quickly moved from the bottom and began swimming when disturbed. There was considerable variability in the amounts of rest across animals and on average there were no day-night differences in activity (Table 1). For three of the five cuttlefish, a state of complete quiescence was followed by a second, sleep-like state that contained phasic motor activity, as reported previously by Duntley et al., [13], [14], [15]. The eyes appeared to rapidly move beneath closed lids, chromatophore activity (CA) suddenly intensified, and the tips of the arms curled and twitched in a way not observed when the cuttlefish was awake (Figure 2 and Video S1,S2). This sleep-like state+CA lasted on average 135 seconds (st.dev. ±23 seconds) and only occurred once (in the night, when there was minimal illumination and disturbance) for each animal during the recording period.

The overall body patterning could best be described as mottled [20], which camouflage the animals well against mixed gravel substrates; during the sleep-like state+CA, particular components of the mottled body pattern turned on and off in an irregular sequence. These body pattern changes did not appear to be random firings of uncontrolled and uncoordinated neurons. We used two strategies to quantitatively evaluate this possibility.

First, we used a dendrogram analyses to determine if the chromatophore patterns observed during this state were randomly generated. Dendrograms were created using correlation coefficient distance (Ward Linkage) showing the relatedness of chromatophore pattern components (as described in [20]) during the sleep-like state+CA(Figure 3). For example, the appearances of the white square, white head bar, and white arm triangle components were positively correlated in all three cuttlefish. These components also clustered in a previous analysis of the structure of cuttlefish body patterning [21]. Similarly, white neck spots, anterior head bar, anterior paired mantle spots, and posterior paired mantle spots usually occurred concurrently with one another in all three cuttlefish. In two of the three cuttlefish (C-Check and Tiny, Fig. 3a,b), the dark arms component occurred independently and was not correlated with any other component. With the third cuttlefish (KY, Fig. 3c), the dark arms component was associated with the presence of anterior paired mantle spots.

Second, we used an autoregressive model to determine if body pattern components, taken as a whole, turned on and off at random. For each second during the sleep-like state+CA period, we determined the total number of components that were turned on as a measure of total activity. This total activity time series data was analyzed using a Box-Jenkins ARIMA model. Each of cuttlefish had a separate time series of total activity. If the body patterning components were randomly turning on and off there would be no significant autoregressive (AR) or moving average (MA) effects. Two of the three cuttlefish showed strong AR(1) time series (C-Check and KY: p<.01 for each series; p<0.10 for the third cuttlefish, Tiny) while all three showed strong MA(1) series (p<0.05 all). The moving average (MA 1) model for all three cuttlefish had a P value of less than 0.05 (C-Check: 0.00; KY: 0.001; Tiny: 0.022). Taken together, these two analyses indicate that body patterning changes were probably not manifestations of random firings of uncontrolled and uncoordinated neurons.

The subjects of this experiment were four fifth generation, laboratory-reared sub-adult cuttlefish (approximately 9 months post-hatching). An all-glass experimental tank (47 L) was divided into ad lib sleep and sleep-deprivation chambers. One vertical camera was positioned directly over each chamber. Both cameras collected continuous frame data throughout the experiment. A third, horizontal camera facing the sleep chamber recorded one frame every ten seconds (time-lapse). The sleep chamber consisted of a 2.5 cm layer of crushed oyster shells, one brick, and several small stones placed by the in-flow valve to prevent the cuttlefish from resting underneath the valve. A few plastic plants were placed on the outside of the tank. The sleep-deprivation chamber was free from oyster shells and small stones. A vertically-facing LCD computer monitor was positioned beneath the bottom of the tank. Hans Zimmer's “King Arthur” film score (2004) was played (Microsoft's Window Media Player) and the visualizer option was used to project random shapes on the screen. Angled mirrors placed around the sides of the tank reflected the screen image all around the cuttlefish to ensure constant, gently moving visual stimulation. Pilot studies showed that cuttlefish remained in constant motion and off the bottom of the aquarium while this stimulus was displayed.

Cuttlefish were acclimated to the sleep chamber for one week (a longer period was used as these animals were not as close to the end of their lifespan). The cameras were then turned on and baseline video data collected continuously for two days as described above. On days 3 & 4, the cuttlefish were housed in the sleep-deprivation side of the tank; during this time, the computer monitor was turned on and the CD was played. On day 5, the cuttlefish were returned to the sleep-side of the tank and video behavior was recorded for an additional 3 three days. Rest and activity were scored as in Experiment 1.

In contrast to the senescent adults (Experiment 1), all juvenile cuttlefish showed clear diurnal organization in rest and activity (Table 1). Activity was highest at night, consistent with nocturnal activity patterns of wild-caught cuttlefish [22]. In addition, juvenile cuttlefish showed significantly more rest during the day than senescent cuttlefish. 48 hours of rest deprivation resulted in compensatory increases in rest time during the first 24 hours of recovery, indicative of homeostatic regulation. (ANOVA F = 6.3, p<0.013, Figure 4). None of the juvenile cuttlefish displayed a sleep-like state+CA (see Experiment 1) either before or after the 48 hours of rest deprivation.

An important strategy in the search for sleep function is the study of sleep in diverse animal species. Through this phylogenetic approach, scientists have discovered sleep-like states behaviorally similar to human sleep in the fruit fly [7], [33] and even nematodes [8]. These states are not only behaviorally similar to human sleep, but they may be controlled by similar molecular mechanisms [34]. Thus, the study of sleep in diverse organisms may eventually provide insights into how and why sleep evolved in the animal kingdom. This may be particularly true for REM sleep, which to date has only been consistently observed in two vertebrate orders. More studies are needed before firm conclusions can be drawn regarding the presence of sleep and REM sleep in Sepia. For example, we did not measure arousal thresholds in these preliminary experiments, and thus cannot confirm earlier, unpublished reports of their presence in Sepia [13], [14], [15]. Nevertheless, our findings provide suggestive evidence that sleep states exist in Sepia and that REM sleep may have evolved separately in this invertebrate animal. The ultimate cause of the evolution of such a state in Sepia are unknown, but it is interesting to speculate that it may be related to the relatively complex nervous system of this animal; a situation that might also explain the presence of REM sleep in birds and mammals.