1.
Introduction
Animals face the challenge of discriminating between the sensory consequences of their
own motor actions, as opposed to sensory stimuli resulting from events taking place
in the external world [1]. Indeed, this ability is key to nearly all locomotor behaviours
such as territorial chases, prey-tracking or positional control, and serves as an
inspiration for the development of artificial autonomous vehicles. Insects are especially
fascinating from this perspective because they accomplish these tasks in flight, at
extraordinary speeds, yet with incredibly stringent size, weight and power requirements
(e.g. [2]). This suggests that they are able to extract, filter and process their
sensory feedback in ways that we do not fully comprehend.
Traditionally, researchers have assumed that insect behaviour is determined by external
stimuli which trigger a chain of activities ranging from sensory signal transduction
to the generation of motor commands [3]. This includes the parallel acquisition and
processing of information within a given sensory system followed by cross-modal signal
integration to generate robust multimodal feedback signals in pre-motor and motor
circuits, ultimately causing muscle contraction and movement. In sharp contrast to
this notion, animal nervous systems can internally generate motor activity, in the
absence of external stimuli (e.g. [4]).
This broad theme is explored by five papers in this special mini-issue centred on
the topic of processing of multisensory information in insects. These studies, which
focus on diverse insect systems, offer unique insights into how insect nervous systems
acquire, process, integrate and weigh information from disparate sensory modalities.
In the first paper, Ruiz & Theobald [5] explore how multisensory integration enables
corrections to disturbances in the flight trajectory of flies. They point out that,
whereas rapid flight trajectory corrections in response to perturbations are strongly
mediated by mechanosensory feedback (e.g. halteres in flies), the role of visual inputs
is pre-eminent over longer timescales. Specifically, they explore which features of
the visual world are most critical for in-flight path correction. Visual and mechanosensory
modalities are widely recognized as two complementary and critical sensory modalities
for locomotor control, yet the nature of the critical features of the visual world,
particularly in natural environments, represents an open problem. For example, the
image statistics of natural environments differ considerably from the simple visual
sensory stimuli that are typically used in studying visual motion sensing. In natural
environments, the image statistics of scenes above the fly (the dorsal field of view)
differ considerably from that beneath the flying animal—sky views present simpler
moving images than ground views. The authors presented laterally moving dot fields
of different densities (above and below the tethered fly) and found that, in the fly
Drosophila melanogaster, the amplitude of stabilizing responses depends strongly on
the density of dots. The dorsal field of view has little effect, whereas the ventral
field strongly modulates counter-steering. This paper thus shows that both vision
and mechanosensory systems are important, but local specialization of the eye in the
lower region also plays a critical role in translational flight stabilization.
The second paper, by Deora et al. [6], probes the role of both mechanosensory and
visual sensory systems and their influence on the feeding behaviour of the tobacco
hornworm hawkmoth Manduca sexta. These moths feed while hovering above a flower and
probe the surface of the flower with their long flexible proboscis to localize a tiny
nectary opening in which they insert their proboscis to feed. As this task is strongly
tactile, mechanosensory input is critical [7]. However, the precise role of visual
input in this behaviour, although essential, is somewhat unclear. To address this
question, the authors quantified hovering flight patterns and feeding performance
of freely behaving hawkmoths feeding from three-dimensional printed artificial flowers
under two different light conditions: a low light level that simulates the light conditions
of these crepuscular moths, and a high light level at which visual feedback should
be more prominent. In both light levels, moths improved their ability to locate the
nectary and feed from the flower. However, at the higher light levels, they took longer
and were less effective in the feeding process. Why increased light should impair
feeding performance remains an open question. These results also help illuminate how
animal behaviour is impacted by increased light pollution in habitats near human populations.
In addition to the dual roles of mechanosensory and visual feedback, the role of odours
in modulating visually guided behaviours was the focus of the third paper by Cheng
& Frye [8]. This paper shows that responses to small objects in the visual field of
flies elicit aversive behaviours; flies tend to steer away from small objects in the
margins of their field of view. Cheng & Frye used a flight arena in which tethered
flies, again D. melanogaster, freely rotated and performed steering manoeuvres, which
were detected by both body rotation and wing-driven turns (via a wingbeat analyser).
These experiments revealed that flies do indeed tend to steer and rotate away from
small objects. This behaviour would of course be maladaptive if the fly needed to
aim towards a food source in their visual field. Because relevant food sources have
odours associated with them, the authors posited that the addition of a plume of odour
(e.g. apple cider vinegar) would elicit frontal fixation of an object, which was indeed
the case. Thus, flies showed a greater probability of frontal fixation (flying towards
the object) and a lower probability of spinning and aversive steering. Taken together
with Ruiz & Theobald's experiments [5], these data suggest interesting additional
experiments that might reveal an odour-gated visual behaviour, depending on whether
objects are presented in the dorsal versus ventral fields of the visual world.
Frontal fixation of visual images is part of gaze (head) stabilization, a key flight
control behaviour that can reduce motion blur, particularly in fast aerial manoeuvres.
This was the focus of the fourth paper, by Chatterjee et al. [9], which explored the
interaction between mechanosensory information mediated by the antennal Johnston's
organ and visual feedback in the oleander hawkmoth, Daphnis nerii. Their study highlights
the critical aspects of sensory processing time—with fast mechanosensing occurring
alongside slower visual image processing. These diverse sensory modalities, operating
as negative feedback loops, have considerably different timescales [10], which likely
results in emergent oscillatory dynamics. In the case of gaze stabilization, the authors
showed that the head of the hawkmoth D. nerii undergoes lateral wobble (small-amplitude
oscillations) [9]. A series of experimental manipulations of the mechanosensory antenna
(e.g. clipping mass of the flagellum, gluing the Johnston's organ) revealed that the
amplitude of head wobble increases when the mechanosensory information is reduced.
Interestingly, light level changes (as also explored by Deora et al. [6]) may also
influence the emergent dynamics, because lower light levels mean longer processing
times. These data also suggest that the frequency of head wobble is lower under darker
conditions. Thus, the enhanced head wobble may be an outcome of broken or delayed
mechanosensory feedback acting within a larger reflex loop.
Visual feedback occurs across many wavelengths, and in a number of studies, polarized
light plays an especially important role (e.g. [11]). The fifth paper, by Uemura et
al. [12], explored how polarized light influences the behaviour of larval insects
as they navigate terrestrial environments. They focussed on an intriguing emergent
behaviour in which social caterpillars of two moth species, the European Thaumetopoea
pityocampa and the Australian Ochrogaster lunifer (Lepidoptera: Notodontidae), walk
single file, crawling head-to-tail as they navigate to feeding sites or pupation sites,
laying down a silk trail along the way. Whereas chemosensory (e.g. pheromone) and
mechanosensory input (tactile information between individuals) likely support this
behaviour, the role of visual feedback is essential in reinforcing this group behaviour.
As it turns out, the polarization of light detected by a novel visual organ identified
by the authors plays a key role in the multisensory behaviour. Thus, orientation is
maintained via feedback from multiple sensory inputs including odour and tactile inputs,
as well as external environmental visual input.