Frigate birds keep their eyes wide open during flight – most of the time!

Sleep patterns of flying birds

Snap-Shot of the study

Birds fly enormous distances during migratory flight. It seemed reasonable therefore to think that birds sleep while flying, especially since birds can go to sleep -one brain-half at a time.

But then, how can they function and be attentive to the demands of flying, foraging, avoiding predators, finding their route, often over water, over such long periods of time? Imagine the frigate birds with the wind on their wings as they begin their flight across the ocean, flying continuously and majestically for upto 2 months! In an elegant and well written piece of research, the authors find that frigate birds sleep very little during their long flight (1). By recording brain activity in flying birds, the authors show for the first time that these birds take short and light naps but mostly forgo sleep during migratory flight.

Background to this study

A single late night is enough to send most adult humans into a downward spiral the next day, fruit-flies that don’t sleep have their lives cut short (2-4). Birds however have been shown i) to go on long flights without resting (5) ii) to sleep uni-hemispherically – single eye open – keeping one part of the brain alert to the requirements of the environment while the other part rests (6). iii) are able to function more or less normally even when deprived of sleep (7)

So what happens during flight? Given the high energy/metabolic demands of flying the need for rest and sleep must be high, on the other hand the need for attention is also high during flight, can birds afford to sleep on the wing? One way to find out is to record the patterns of brain activity in migratory birds during flight and look for patterns of sleep and wakefulness.

What did they do?

In this study (1), sleep patterns of 15 female great frigate birds flying over the Pacific Ocean and after returning to their nest on Genovesa Island (Galápagos) were recorded using implanted devices that measured brain activity (EEG – electroencephalogram), movement of the head as well as acceleration. No behavioural differences were observed in the birds with implants during and after removal of the devices. Using these devices, the researchers measured movement of the bird, acceleration, deceleration, flapping of the wings and brain activity near the primary visual area and also collected data on weather conditions. Measuring movement of the head allowed them to separate patterns formed by head movement from actual changes in brain activity.

What did they find?

The overall EEG patterns were similar in the birds on land and during flight, allowing the researchers to look at duration and intensity of sleep. They describe three sleep-awake states roughly – wakefulness, rare episodes of REM sleep (like in humans, this is deep sleep characterized by rapid eye movement) and slow wave sleep – this is the most frequent type of sleep described in birds which can be bi-hemispherical or uni-hemispherical or assymetric (6).

During the day, the birds showed patterns of wakefulness (fast head movements together with high amplitude signals in the EEG), even at night during flapping of wings wakefulness patterns were observed. These were interspersed with slow waves which the authors identify as slow wave sleep. Rarely, between bouts of slow wave sleep, short bursts of deep sleep or REM sleep patterns characterized by dropping of the head, and twitching were also observed. Interestingly, birds ascended in altitude during slow wave sleep and descended during wakefulness. All types of slow wave sleep, including unihemsipherical, asymmetrical (when one hemisphere was more active than the other)  and bihemispherical slow wave sleep were observed during flight. There was an increase in asymmetric sleep in flight than on land but this was not correlated with any one type of movement.

Overall, frigate birds seemed to sleep very little during flight – in shorter bursts and less soundly –  a homeostatic balance was restored when these birds landed. Further, preliminary evidence from this study suggests that these bursts of sleep are enough to sustain the birds during flight.

Why is this interesting?

The very fact that these birds are able to accomplish Himalayan tasks such as follow migration routes, feed themselves, with such low levels of sleep suggests that, at least for frigate birds, sleep may be dispensable during flight. Are they postponing this need? What sort of adaptations allow them to postpone sleep or perform sleeplessly? This study is a step towards understanding adaptations to lack of sleep and perhaps a way to understand the very nature of sleep itself.

1. Evidence that birds sleep in mid-flight. Rattenborg NC, Voirin B, Cruz SM, Tisdale R, Dell’Omo G, Lipp HP, Wikelski M, Vyssotski AL. Nat Commun. 2016 Aug 3;7:12468. doi: 10.1038/ncomms12468. PMID: 27485308

2.Reduced sleep in Drosophila Shaker mutants. Nature. 2005 Apr 28;434(7037):1087-92  Cirelli C, Bushey D, Hill S, Huber R, Kreber R, Ganetzky B, Tononi G.

3. Genetics of sleep and sleep disorders. Cell. 2011 Jul 22;146(2):194-207. doi: 10.1016/j.cell.2011.07.004. Sehgal A, Mignot E.


5. Frigate birds track atmospheric conditions over months-long transoceanic flights. Science. 2016 Jul 1;353(6294):74-8. doi: 10.1126/science.aaf4374. Weimerskirch H, Bishop C, Jeanniard-du-Dot T, Prudor A, Sachs G.

6. Half-awake to the risk of predation Nature 397, 397-398 (4 February 1999) | doi:10.1038/17037 Niels C. Rattenborg, Steven L. Lima & Charles J. Amlaner

7. Adaptive sleep loss in polygynous pectoral sandpipers. Science. 2012 Sep 28;337(6102):1654-8. Epub 2012 Aug 9 Lesku JA, Rattenborg NC, Valcu M, Vyssotski AL, Kuhn S, Kuemmeth F, Heidrich W, Kempenaers B.


An interview with Alexei Vyssotski

Q. Can you tell us a little more on how you determined that the birds were sleeping? How did you identify the pattern corresponding to sleep when you got all your recordings?

Sleep was identified by visual inspection of 4-sec episodes of raw EEG records. Slow-wave sleep is characterized by large amplitude low-frequency oscillations in EEG. These episodes are easily-detectable. Visual scoring was used because automated methods of sleep staging can’t separate properly large-amplitude EEG events from movement artifacts in all cases. Locomotor artifacts can have similar amplitude to slow-waves during in sleep.

Q. Do you think these birds sleep a lot during an annual cycle? Do migratory birds tend to sleep longer than non-migratory birds on average – there must surely be some compensatory mechanisms?

We have found that frigatebirds sleep on average 9.3 hours per day when on land and only 0.69 hours per day when flying. We did measurements only during breeding period. While the frigatebirds live in equatorial area with relatively small weather seasonal changes, one can suppose that the duration of sleep is linked stronger with the pattern of animal activity than with the time of the year per se. It is known that migratory species can stay on the wing for a long time. Extrapolating our findings to these species one can suppose that they should sleep in the flight smaller amount of time than on land, and might compensate migratory sleep loss on land later like our frigatebirds did. However, the compensatory increase in sleep duration and intensity upon landing after the trip is relatively small comparatively to missed amount of sleep on the wing. One can speak only about partial compensation of sleep loss. The birds have, probably evolutionary formed, an ability to stay without sleep significant amount of time without physiological dysfunctions. Unlike most mammals, the birds do not have so called sleep-deprivation syndrome. If a rat will be forced to stay without a sleep for significant time, it will die, but a homing pigeon can stay month-long awake and still behave properly. Migratory birds definitely reduce amount of sleep during migration, but whether they sleep longer than non-migratory birds in other situations is difficult to say.

Q. Do you think that migratory birds produce neurochemicals that resist the urge to sleep or have special brain structures?

The neurochemistry of avian sleep is investigated in much less extent than mammalian sleep. To the best of my knowledge, no special anatomical structures that are responsible for sleepless in birds have been discovered. It is assumed that the avian sleep control brain system is similar to mammalian. However, additional studies are needed to check how strong this similarity is indeed and what are the differences.

Q. Does the slow sleep wave recordings refer to the activity of only certain regions of the brain?

No. It is assumed that like in mammals, the vigilance states in birds are controlled by deep brain structures that modulate activity of superficial brain regions in a synchronous manner. However, contrary to most of mammals that have only bi-hemispheric sleep, the birds have so- called unilateral sleep, when one hemisphere is sleeping and another is awake. Thus, the avian brain hemispheres are more independent from each other than mammalian. One should note that the phenomenon of “local sleep” has been also discovered in birds. This means that if a particular region of the brain has been used intensively in a wake state, the slow waves of increased amplitude will be observed during the following sleep in this brain area.

Q. Did you have independent recordings of some of the parameters using a high speed camera to set the baseline for each bird?

Do you mean Ca2+ brain cells imaging? No, we did not do this, but of course, if would be nice to monitor how activity of different cells ensembles changes in sleep in birds. The recently developed head-attached microscopes can help to film brain in freely behaving animals.

Q. It is astonishing that you got so much information that you could put together, did you expect that when you captured the 15 birds? What were the unexpected challenges in the study?

That is true, this time we collected more information than in our previous studies. We alreadyhad experience with EEG and GPS logging. This time we compensated luck of GPS precision by the acceleration data to reveal the flight mode of the animal. 3-D acceleration data practically doubled the dataflow, but this was not a problem to log in 1 GB onboard flash memory. We did not predict the particular way of the data analysis in advance, but observing three-modal distribution of sway acceleration leaded us to separate analysis of EEG in three different flight modes (straight flight, circling to the left and to the right). The real challenge was to master the surgery and handling in an animal-friendly way to have the birds back with the equipment. Indeed, the rate of return 14 from 15 birds exceeded our expectations. To be honest, I expected larger losses and was very happy when returns exceeded 50%.

Q. Have you been to the Galapagos? What is it like to do research in that setting? Do you think that nearly 180 years after Darwin’s voyage, the biodiversity of Galapagos still holds new discoveries?

Yes, I have a real luck to spend a week at the Darvin station in Santa Cruz and then a week on Genovesa island working in the bird colony. My colleagues Bryson Voirin and Ryan Tisdale spent two weeks more waiting for birds return. This is the best place for animal study that I have ever seen. Wild birds behave almost like tame animals there and do not run away from humans. Thus, it is easy to handle them. This is, of course, one of the features that attracts biologists. The biodiversity of these islands is definitely not studied completely and will attract scientists for a long time.