Sensory processing during sleep in Drosophila melanogaster

When asleep, most animals enter a state of altered consciousness that detaches them from the surrounding world allowing for prolonged episodes of undisturbed rest. At the same time, however, they must also maintain some level of awareness necessary to wake up when challenged by relevant stimuli. Stimuli that are mostly quantitative (e.g. the loudness of a bang, the brightness of a light) can be filtered out by lowering the response threshold at the neuronal level, but it is unclear how a sleeping brain could discriminate stimuli qualitatively (e.g. the meaning of a word, an angry tone of voice). It was first shown in the 1960s that sleeping humans respond to the sound of their own names being called and conceptually similar observations were later made in rats, cats, and primates. To explore the neuronal underpinnings of this phenomenon in a genetically amenable model, we turned to the vinegar fly, Drosophila melanogaster. Sleep in flies recapitulates most of the behavioural characteristics of sleep in mammals, including a stereotypical modulation of arousal threshold. It has been widely established that the reaction of a sleeping fly to a mechanical stimulus varies with the intensity of the stimulus and the internal state of the animal, but can sleeping flies, like humans, discriminate stimuli qualitatively too? To address this question we built a robotic machine able to selectively probe single flies with air puffs of identical mechanical intensity but different odour saliency and using this new experimental paradigm we were able to prove that flies have the ability to recognise the qualitative saliency of a stimulus during sleep, similarly to what humans do.

The research we are proposing builds on these preliminary findings and promises to extend them further thus pioneering a little new research field: the study of sensory processing of information during sleep in a genetically amenable animal model. In particular, we want to better understand what are the neurons that regulate this phenomenon. What are the neurons that act as sentinels in the brain and maintain the ability to perceive and decode external sensory information? How can they afford to do so while the rest of the other neurons must instead "sleep" unresponsive?

This research is important for the larger neuroscientific community because it provides information on how a sensory input is decoded within the brain during different states of conscious awareness: wake and sleep. It is also important for the sleep community because it addresses one of the most puzzling aspects of sleep: its cost. We still do not know what is the function of sleep but whatever that is, it requires animals to disconnect from the external world at a potentially very high cost in terms of danger. This research can shed light on how neurons achieve this disconnect and how and why some of them actually manage to escape it.

In 1960, Oswald et al. showed in a seminal study that sleeping humans have the ability to selectively wake up to stimuli that are considered salient, such as the sound of their own name being called (but not the sound of an unfamiliar name). We now know that this processing ability is common among animals, having been described in similar conceptual terms, also in rodents, cats, and other mammals. This well-known phenomenon prompts a question that remains unanswered: how can a sleeping brain recognise the qualitative valence of sensory input during sleep? Can a circuit be concomitantly asleep and yet able to decode complex information, such as the meaning of a sentence or of a tone of voice? This is not just a descriptive question, for it touches on one of the most mysterious properties of sleep itself: the physical necessity to disconnect from the outer world and enter a state of altered consciousness.

We have recently created and presented a new paradigm to study this phenomenon in the vinegar fly, Drosophila melanogaster (French et al., Nature, 2021). This pioneering work was well received by the community because 1) it shows that even flies have the ability to wake up selectively to salient stimuli, adding a new layer of evolutionary complexity to sleep; 2) it provides an initial description of the neuronal circuit responsible for this phenomenon.

We here propose to continue this initial work, expanding the anatomical and functional description of the circuit and investigating in greater detail how the circuit can be modulated. We have three complementary objectives: understand how the circuit regulating sensory processing during sleep can be modulating its activity in response to a change in internal states, such as hunger or previous sleep history; how does sensory processing correlate with different sleep stages, such as light sleep and deep sleep; explore whether the circuit is plastic and whether new saliency can be acquired through learning.