Imaging the neural mechanisms of distance estimation in insects

Stereopsis is the ability to perceive distance using small differences between the images seen by the two eyes. Long thought to be a complex ability confined to "higher" animals such as primates, it has now been identified in species as diverse as mice, owls, toads and cuttlefish. It has so far been demonstrated in only one type of insect, the praying mantis. This raises fascinating questions about how tiny insect brains achieve stereopsis. Notably, we do not know how they solve the fundamental problem of "stereo correspondence": the matching of an image of an object in one eye, with its corresponding image in the other eye. This is challenging in complex scenes when multiple similar objects are present, yet mantids hunt successfully in such environments
We have recently developed a "virtual reality" laboratory for mantids (1,2), which allows us to study this remarkable behaviour under experimental conditions. Our behavioural work has found that insects and humans achieve stereo correspondence in very different ways. Our neurophysiology work to date has involved microelectrode recordings of visually responsive neurons within the mantid brain, but this technique is limited to just one neuron at a time. This project will instead utilise the latest microscopy technology - 2-photon imaging - to visualise the mantid brain in action, using calcium biosensors. Ca2+ provides a read-out of neuronal activity, because every time a neuron fires, Ca2+ enters the cell, allowing us to follow the activity of many neurons simultaneously and with single-cell resolution.
The supervisory team includes experts in mantid neurobiology, insect genetic engineering (to introduce the biosensor) and 2-photon Ca2+ imaging. In Year 1, we will establish imaging of the mantid brain, using Ca2+ -sensitive dyes that don't require genetic manipulations. During Years 1 and 2, we will also attempt to create transgenic mantids in which specific neuronal populations express Ca2+ sensors intrinsically. This will allow us to target the sensors to subpopulations of neurons, helping us to dissect out the neuronal interplay during stereopsis perception.
By Year 3, the student should be fully trained and competent in the imaging techniques, and we should be routinely able to image neural activity. We will now use the "insect 3D cinema" used in our previous neurophysiology experiments, where the brain is exposed via the back of the head with the insect lying face down, viewing a computer monitor via a mirror and coloured filters which enables different images to be presented to each eye (mantids lack colour vision so this does not create artefacts). We have previously found that activity in the relevant neurons can be evoked by briefly flashed stimuli, so we will be able to interleave stimulus presentation and fluorescence recording, thus avoiding contamination of the fluorescence signal with the visual stimulus.
We will first replicate previous brief experiments, demonstrating that different classes of neurons in mantis brain are tuned to the location of a coarse bar stimulus in 3D space. The longer experiments possible with imaging will then enable us to probe the properties of these neurons in far more detail, including how these neurons respond when multiple objects are present and thus how/where insect brains solve stereo correspondence. We envisage that data collection will occupy most of Year 3, while in Year 4 the student will construct mathematical models summarising the behaviour of these neurons.
The results will be of interest well beyond insect neuroscientists, since they will reveal to what extent different taxa (e.g. arthropoda, mammalia, aves) have convergently evolved similar solutions, versus discovered novel solutions. For instance, our results will inspire robotic approaches for visual perception machines, as well as shedding light on our own perception.