Bertram was the first speaker after the student presentations at the summer school of the International Society for Neurochemistry. He started out with an attempt to convince people that there are cognitive processes in animals. He argued mainly from evolution, trying to pick a simple enough process in his model system, the fruit fly Drosophila, that we can understand, yet complicated enough to be interesting. Whithout explaining what is 'cognitive' about the process of Pavlovian (or classical) conditioning, he explained how groups of flies are conditioned to prefer different odors: the flies are presented with one odor and receive shock with this presentation. Then, they are presented with another odor without shock. After such training, the flies are placed in a choice chamber where they decide towards which odor they prefer to walk.

Bertram then went on to present a general, neurophysiological model of the structures whithin which this sort of learning is thought to take place: olfactory sensory neurons in the antenna project into the glomeruli of the antennal lobes. Projection neurons then transmit olfactory information to the mushroom bodies, where this information converges with probably dopaminergic neurons conveying information about the electric shock. At this point of convergence, a type I adenylyl cyclase (coded for by the gene rutabaga in Drosophila) serves as coincidence detector between olfactory information and electric shock. It is thought that one of the downstream targets of rutabaga is Synapsin which gets phosphorylated by PKA after Rutabaga activation. This phosphorylation then increases the number of vesicles in the synapse ready for transmitter release.

After this basic introduction, Bertram started to talk about the importance of the timing of the electric shock: if the shock comes at the same time as the odor or slightly afterwards, the animals avoid the odor paired with the shock. However, if the odor is presented at the offset of the electric shock, the odor becomes attractive for the flies. Bertram calls this situation 'Pain Relief Learning', as the odor signals the end of the aversive electric shock stimulus.

Synapsin null mutants, as well as RNAi-mediated synapsin knock-down, show normal shock avoidance, but impairments in both punishment and relief learning. Restoring synapsin function in only mushroom-body neurons (via GAL4 line MB247) restores both learning phenotypes, suggesting that these neurons subserve both memories, despite the animals behaving exactly opposite when they encounter the odor.
Expressing a mutated form of synapsin that lacks PKA phosphorylation sites only restores punishment learning but not relief learning, suggesting that relief learning differs from punishment learning upstream of synapsin.

Bertram then switched to experiments in humans. He started by explaining fear-potentiated startle experiments: a neutral stimulus is first paired with a mild electric shock. Then, the stimulus is presented just before a starteling stimulus, such as a loud noise. If the electric shock was presented together with, or slightly after the neutral stimulus, the startle is potentiated. If the neutral stimulus was presented at the offset of the electric shock, the startle is reduced. Thus, the timing effects of the neutral stimulus and the electric shock are nalogous to punishment and relief learning in flies. fMRI experiments in humans show that punishment experiments show activations in the amygdala, while relief experiment lead to activations in the striatum.

To gain causal insights into the amygdala/striatum connection to punishment and relief-learning, respectively, Betram collaborated with researchers who infused muscimol into the amygdala and striatum, respectively, in rats during punishment and relief learning. What they found was a double dissociation between the amygdala and striatum for punishment and relief learning, suggesting indeed a causal role of these two brain regions in the two learning experiments.

As there was some time left after he was done presenting his relief-learning data, Bertram added some results on Drosophila larval learning. Again, the experiment done was classical olfactory conditioning. Bertram found that 'arctic root' compound doubles the learning scores in the larvae, in a dose-dependent manner. Apparently, these increases are independent of any sensory or motor alterations in the animals. Interestingly, these effects can only be obtained with the original root extract, not with any commercially availalable material. At the moment they're trying to find out which compounds in the original extract are responsible for this memory increase as opposed to the commercially available products.