European Commission logo
English English
CORDIS - EU research results

Development of the circuits in the locust brain for the early detection and avoidance of looming objects

Final Report Summary - NEURAL DEVELOPMENT (Development of the circuits in the locust brain for the early detection and avoidance of looming objects)

Dr Rind's laboratory has pioneered the field of biologically inspired robotics using circuits inspired by insects for collision detection in vehicles. This was based on the research done over the last 30 years on the circuitry involved in looming detection (the lobula giant movement detectors, LGMD1 and LGMD2, which respond selectively to an object approaching towards its eye) and evasive behaviours in adult locusts. It was focussed on last moment reactions displayed just before collision.

Our project searched for a better understanding of the looming detection system by analysing the ultrastructure together with the responses of looming sensitive neurons and evoked behaviours in each phase of the locust's development. Based on the excellent understanding of the LGMDs ultrastructure and synaptic connections in the adult, it would be very informative to study the circuitry during development. This might lead to a breakthrough if, for example, some parts of the circuitry develop later than others. This, together with studying the development and the response changes of the LGMD1 (through extracellular recordings performed in the postsynaptic DCMD neuron) to looming stimuli, may lead to the possibility of improving the models already in use.

We focused our research on escape behaviours on the ground, particularly the hiding response (a positioning behaviour in which locusts sitting on branches tend to hide from an approaching stimulus by moving behind the object), that require more time to prepare and thus involve detection of a predators at a longer range. Unveiling the circuitry beneath this type of behaviour may enable the development of technologies in cars / vehicles for hazard detection over a longer range.

During the course of this project we made the following advances:
- We discovered that locusts were able to perform the hiding response in all developmental stages (even in 1-day-old first instars) implying that the neural pathway involved matures early in development.
- For evaluating the possible role of the LGMD neurons in such behaviour we used a multiple-techniques approach.

Our principal findings were:
- A description, for the first time in juvenile instars, of the 3D structure of the LGMD1 and LGMD2, two key neurons for the detection of approaching stimuli. We are now analysing the changes in their structure found during development to establish the order of maturation of the different subparts of the neurons.
- We found that both LGMD1 and LGMD2 have mature synapses at the earliest stage of development and already show the same functional organisation as the adult. We did find a consistent increase in the density of synapses during development and are analysing possible asymmetries in the maturation of the LGMD1 compared with the LGMD2.
- We proved that even the LGMDs of the first instars can serve as functional detectors of approaching stimuli. We recorded from DCMDs in the different instars and found a consistent activation of the neuron towards the end of the approach of the stimulus. We did find differences in the exact timing of the responses (onset, following of the progression of the loom and offset) when comparing the responses from young animals and adults. Dr Simmons is now further analysing the fine properties of these responses (using different sets of stimuli) to understand the development of the different subfields and their function in the complex computations known to take place in the adult neurons.
- Through different results, we established the LGMD2 as the candidate neuron involved in triggering the hiding response. Particularly important is the fact that we could trigger the hiding response by using computer-generated stimuli that were programed following the response preferences of the LGMD2 but not those of the LGMD1.

In summary, we made significant advances towards the planned objectives and are expecting to be able to incorporate our finding into biologically inspired circuits.