Fly Attention
5 Apr, 2007 04:53 pm
In this article I show how one can study selective attention in flies by measuring their brain responses to competing visual stimuli. Fly mutants with defective short-term memory were found to also display defective attention-like processes.
Much of our understanding of memory mechanisms results from work done in model organisms such as the fruit fly, Drosophila melanogaster. But measures of selective attention have somehow been lacking in the fly, even though one might suspect that some aspects of its learning and memory might require attention-like processes (e.g., stimulus selection, suppression, and alternation). One reason that attention and memory have not been studied in parallel in the fly is because the most effective learning paradigms have been olfactory, not visual. Visual paradigms are in comparison much better suited for studying attention, because fine spatio-temporal control of competing visuals allows for precise measures of response selection and suppression.
Ingenious behavioral experiments by Martin Heisenberg and Reinhard Wolf testing visual responsiveness (in tethered flight) to competing objects uncovered the first evidence of attention-like processes in Drosophila. The recent discovery that Local Field Potential (LFP) oscillations in the fly brain correlated with behavioral responses was further evidence. LFPs resemble electro-encephalograms (EEGs), the electric fields or "brain waves" which are often associated with behavioral states in humans. LFPs in the fly brain were found to increase in amplitude when flies responded to relevant visual objects, and decrease in amplitude during sleep, anesthesia, or simply when an object was not relevant anymore because some time had elapsed.
I investigated the temporal dynamics of fly brain responses to visual novelty. Fruit flies are novelty-seekers. When flies were presented with a recurrent novel object every 100 seconds, their brain LFP activity locked onto it while suppressing responses to a competing control object which never changed. These selective brain responses lasted on average 9 seconds before extinguishing or alternating to the competing object. When the recurrently novel object re-appeared more often (every 25 seconds) it did not elicit a selective brain response anymore, suggesting it was not novel any more if the fly had last seen it more recently. Flies attention thus appears to display stereotypical temporal dynamics, and these are active processes distinct from simple habituation or fatigue.
What is the neural basis of visual selective attention in Drosophila? It seems likely that attention mechanisms overlap to some extent with learning and memory systems. Fly memory mutants, for example, may be unable to remember well because they are not paying attention appropriately. Drosophila short-term memory mutants such as dunce and rutabaga, which have defective memory in both visual and olfactory paradigms, are thus intriguing candidates for detailed investigations into selective attention. Predictably, these mutants did not attend to visual novelty in the same way as normal flies, by either failing to respond or by not sustaining their attention long enough. The effect of the dunce mutation on attention could be corrected by adding a normal version of the gene, but only if the gene was expressed during the growth and development of the fly brain. This suggests that connections made during brain development define the tempo of attention-like processes in the adult fly. Perhaps similarly, our brain development as children constrains to some extent our attention idiosyncrasies as adults.
To effectively dissect mechanisms of attention in Drosophila will require screening through many more mutants and genetically engineered flies. It was therefore important to find an efficient behavioral correlate for fly attention. Surprisingly, mutants with defective visual attention were found to respond more strongly to moving gratings in a simple behavioral paradigm. Although perhaps counter-intuitive (flies with defective attention respond more to moving objects?), this behavioral observation actually makes some sense. Optomotor responses (orienting to motion) are reflexes, which should be suppressible by shifting attention to other stimuli. Dunce and rutabaga mutants may be unable to suppress optomotor responses for the same reason that they have trouble detecting visual novelty: defective attention processes. By using both behavioral and electrophysiological approaches in the Drosophila model, future work should begin to uncover insights into general principles underlying selective attention in this small brain.
This is a great paper!