Project Summary
Our ability to learn and to remember countless experiences is essential for our daily lives. However, our
dependence on these processes also makes us vulnerable to disorders which affect the underlying neuronal
circuits. Notably, disorders causing cognitive impairments may broadly disrupt multiple forms of learning and
memory, even without directly affecting the underlying circuits themselves. This may be due to disruptions of
processes that coordinate learning across multiple learning systems. While the systems that form our different
kinds of memories, like motor skills or directions, are largely separate, they interact in various and surprising
ways. Memorizing facts can for example interfere with or boost the learning of motor skills. This suggests that
the continuous stream of experiences and learning events needs to be tightly controlled to prevent disruptions
and to take advantage of existing memories. We suggest that prefrontal cortex (PFC) exerts such cognitive
control and coordinates learning across multiple systems. Cognitive impairments disrupting PFC function may
therefore impair learning broadly. In particular, we suggest that PFC, to coordinate learning processes, extracts
common rules and task components. In the long term, this allows for identification of hidden relationships,
generalization across tasks and improved learning of new tasks. We hypothesize, however, that in the short
term, the underlying mechanisms interfere with and, under certain circumstances, also enhance learning. We
will test this using our cutting-edge custom infrastructure for high-throughput behavioral training and for
continuous (24/7) long-term in vivo electrophysiological recordings. With this, we will address three
interrelated, but fully independent aims. Aim 1 will identify the task components which, when shared across
tasks, drive interactions between learning systems. In large-scale behavioral experiments we will train animals
on pairings of procedural and declarative tasks and manipulate shared procedural or declarative task
components. This will reveal whether the two learning systems differentially affect the mechanisms underlying
learning interactions. Aim 2 will determine the neural activity dynamics in PFC during learning interactions,
track the development of representations of shared task components, and show how optogenetically identified
PFC projection neurons mediate learning interactions. Importantly, for this we will integrate high-density
Neuropixels probes with our system for long-term electrophysiological recordings. Finally, Aim 3 will causally
test if and at what times during training PFC and its projections are necessary for learning interactions, using
chemogenetic manipulations via implanted micro-infusion pumps over weeks of training. This multi-level,
orthogonal experimental approach promises vertical advances toward an understanding of the circuit
mechanisms underlying learning interactions and the relationships between distinct learning systems. Our
insights may also help to understand how cognitive impairments broadly affect multiple forms of learning and
memory and aid the improvement of existing and the development of novel diagnostic tools and treatments.