The second part of the session on neural networks underlying learning and memory was started by Randolf Menzel on "Neural signatures of learning and memory retrieval in mushroom body extrinsic neurons". Extrinsic PCT neurons are thought to provide feedback-inhibition from the mushroom-body alpha lobe to the calyces. These neurons do show some correlates during olfactory conditioning, but deriving a clear role is difficult. One identified neuron, the PE1 reduces its activity in response to the conditioned odor. Potentially, this reduction in activity could be due to increased activity from a presynaptic PCT neuron. In summary, PCT neurons could be found which increase and those which decrease their responses to both the odor paired with sucrose and the one not paired with sucrose.

Second speaker was Kausik Si, talking about "The role of self-sustaining amyloidogenic state of Drosophila Orb2 in persistence of memory". This was a change in the program due to Si not being able to attend his scheduled session yesterday. He told us about his work on CPEB. CPEB in Aplysia is required for maintaining long term memory over several days. Interstingly, CPEB is a prion-like protein (it can switch from a monomeric to a oligomeric state) and forms a self-sustaining switch for protein synthesis at activated synapses. The oligomeric form of this protein is called the 'amyloidogenic state'. In flies, Orb2 is coding for a relevant CPEB. Activating mushroom-body neurons led to amyloidogenic oligomerization of Orb2. To test the relevance of Orb2 for memory, they used courtship conditioning and appetitive olfactory conditioning. In both experiments, mutants on Orb2 showed specific deficits in 3-day memory, but not earlier.

The final speaker of this long session was Li Liu. I've known Li for many years, as he was a psotdoc in Martin Heisenberg's lab when I was a graduate student there. He talked about visual leaning in Drosophila at the flight simulator. He presented his rescue experiments in which he drove rutabaga gene expression in various parts of the brain in a rutabaga mutant background. These experiments revealed that various populations of neurons in the central complex of the fly brain are potential stores of visual memory. Most prominently, different layers in the fan-shaped body support learning of different shapes and forms of the visual patterns used to condition the flies. He then went on to show that also populations in the ellipsoid-body can rescue the rutabaga mutation for T-shaped patterns. Driving rut-RNAi in either the fan-shaped body or the ellipsoid body showed that both structures are also necessary for this type of learning. Having two structures independently being both necessary and sufficient is logically impossible. This major problem was elegantly re-solved by doing the same experiments in the background not of the rut2080 allele, but the putative null-mutent, rut1. In that background, they needed to drive the rut-rescue construct both in the fan-shaped body AND the ellipsoid body to get the rescue, making much more sense.