The effect of feedback connections on information processing at the first visual synapse of Drosophila and on the animal behaviour

The brain selects where, what and how we gaze. From mammals to insects we know that there are important feedbacks from the brain centres to the sense cells, and within the sense organs themselves. However, we know little how these contribute to the processing of sensory information. Whereas humans can focus their gaze on objects of interest, insects cannot do so, as the lenses of their compound eyes have fixed focal ranges. Instead, it is likely that their attention control may take place in a form of coordinated signalling at the level of the first visual synapses, the sites where the photoreceptor neurons innervate the brain. In many animals the first visual synapses utilise continuous electrical messages instead of electrical impulses to transmit information as this provides richer communication. This is also so in the photoreceptor-LMC (large monopolar cell; the first visual interneurone) synapse in fly, where six identical photoreceptors that look at the same small visual field connect to one LMC to improve the quality of their joint transmission. Yet, this is an overly simplified view of the information processing taking place at the first visual synapse. As in the vertebrate eye, the communication from photoreceptors to the brain occurs not just one way, but instead the cells that are synaptically linked to photoreceptors feed also information back to photoreceptors. The significance of this feedback communication is, however, poorly understood, in part because monitoring in electrical messages between small visual neurones is very difficult. Here, we propose to study how the communication between (i) photoreceptors and peripheral neurones and between (ii) photoreceptors and more central visual neurones of the fly enables it to extract and filter information from the visual scene in space and in time, and relay this information to higher visual centres and to the brain. This strategy is facilitated by (1) the possibility of genetically engineering fruit flies (Drosophila melanogaster) whose neurones display reduced or increased activity, (2) the favourable location and architecture of their visual systems, (3) our ability to perform high quality electrical recordings from visual neurones and (4) the behavioural tests, where we compare the learning performance of wild-type flies and mutant variants that have malfunctioning communication channels feeding back to photoreceptors.

Recent behavioural studies have shown that fruit fly, Drosophila, has attentive mechanisms that, according to the proposed model (Tang et al., 2004), should also affect early visual processing. This idea is further supported by the highly organised feedback connections within the fly eye that also link central visual brain centres to the first optic neuropile, lamina, in the periphery (Meinertzhagen and Sorra, 1991). Moreover, our recent electrophysiological data from the first visual interneurones, LMCs (large monopolar cells), have shown that signal transmission from photoreceptors to LMCs, and further towards the brain, is regulated by this feedback network. Here we propose to investigate the role of individual feedback connections in regulating the transfer of visual information from photoreceptors towards the brain in in vivo Drosophila. We will genetically dissect synaptic pathways between peripheral and more central visual neurones within modular visual processing units, called neuro-ommatidia. Comparing neural responses of output neurones in wild-type flies to those in flies that have altered connectivity will enable us to elucidate the rules, patterns and limitations of synaptic signal transfer from photoreceptors to LMCs. We will focus on two questions: First, How does the distribution of input signals over the receptive field of a LMC change with time and intensity? and second, How does the feedback from the brain and from the neighbouring interneurones affect these signals to generate particular patterns of network activity and whole animal behaviour? We will study these questions using intracellular voltage responses of LMCs to light patterns of a novel feedback-controlled LED matrix. Since this stimulator uses a running statistical measure of an LMC's responses, to regulate the generation of temporal and spatial light patterns, we will learn how adaptation and attention dynamics within the neural network of visual neurones control the transmission of information towards the fly brain. We will also conduct behavioural studies with Prof. Tang (Beijing, China) using a flight simulator in order to link the molecular, cellular and network physiology to whole animal behaviour. Finally, the results will be quantified with modelling and statistical analysis.