Divergence in sleep architecture associated with interspecific shifts in memory investment

Abstract

Sleep plays a critical role in stabilising memory. However, whether the architecture of sleep co-evolves with investment in learning and memory is unknown. To address this, we exploit the extreme levels of neural and behavioural divergence among Heliconiini butterflies. Compared to other Heliconiini, Heliconius have massively expanded mushroom bodies, the insect centres of learning and memory, which are associated with specific enhancements in visual learning and memory. Using a purpose-built video system, machine learning-facilitated behavioural tracking, and in-depth analysis of micro-behaviours during nighttime quiescence, we describe multi-state sleep in Heliconius melpomene and an outgroup Heliconiini, Dryas iulia. In both species, sleep is predominantly composed of quiet sleep, however, Heliconius and Dryas deviate in specific features. In Dryas iulia, quiet sleep is characterised by highly stereotyped, rhythmic antennal oscillations. Conversely, the antennae of Heliconius melpomene remain largely motionless but occasionally perform large movements which interrupt quiet sleep. These antennal deflections are associated with bouts of activity that include increased overall antennal displacement, isolated head twitching, proboscis movements and less frequently, wing movements. While these episodes were also seen in Dryas iulia, they were considerably less frequent. We further demonstrate that the behavioural profile of these active sleep bouts is constituently different from bouts of induced wakefulness, providing experimental support for multi-state sleep in Lepidoptera. This work shows that Heliconius melpomene, a species distinguished by its expanded mushroom bodies, exhibits more frequent bouts of active sleep than the outgroup species Dryas iulia. These findings raise the possibility that increased investment in learning and memory, as well as the structures that support it, may be associated with changes in sleep architecture.