I was so busy at my two posters that I had almost missed the talk by my good colleague and friend Bruno van Swinderen! Many thanks to Kit Longden for coming and picking me up at my poster.
Bruno started with a great general introduction into what attention might be and how you can define it to study it in animals. He showed how experiments with bistable images in humans can be mimicked by choice paradigms in animals, even insects.
Bruno studies the fruit fly Drosophila and records local field potentials (LFP) in its brain. There is an increase in 20-30Hz activity in the LFP whenever a visual pattern passes in front of the fly. These increases can be modulated by salience and by other sensory modalities. The 20-30Hz LFP activity is sensitive to novel stimuli such that only novel stimuli elicit a strong LFP response. The novelty effect in the LFP lasts about 12 seconds and then disappears. In 12 seconds the animal gets about three presentations of the novel pattern. Counting how often the LFP response switches between patterns it turns out that these switches are not random, but a feature of the brain's attention system.
Since such a system must involve memory (i.e., "how long have I looked at this pattern?"), Bruno started studying the LFP response in learning mutants. The first mutant studied was dunce. Dunce shows a significant reduction in the LFP response. This deficit can only be rescued by expressing the dunce gene during development.
How does this deficit in an attention-like process manifest itself in the behavior of dunce mutants? To study this, Bruno studies the distractability of flies in an optomotor essay. Dunce flies are better than any other strain Bruno tested in optomotor behavior. This appears to be due to a lack in distractability. Bruno tested this by adding distractors around the optomotor machine. As suspected, dunce is less distracted than control flies during optomotor behavior.