Dissecting the role of defined neuronal populations in fear learning

The astonishing plasticity of the brain and its ability for continuous learning and memory formation are among its most essential functions, since constant adaptation to the environment is critical for an animal’s survival. The individual brain areas mediating distinct types of learning and the cellular and molecular mechanisms underlying synaptic plasticity have been studied in great detail in the past. In order to understand learning and memory, though, it is necessary to bridge the gap between our insights on the macro- and the micro-scale. Neuronal circuits are the link between individual cells and entire brain areas and learning manifests as profound changes in their processing of information. However, the identity, function and mechanisms of neuronal circuits mediating learning are relatively unexplored. The study of neuronal circuits will not only reveal the actual sites of plasticity, but also the control mechanisms for the acquisition and expression of memories. To dissect the organization and function of neuronal circuits, both locally and across multiple brain areas, it is key to identify their fundamental building blocks – defined neuronal populations. Their specific roles and interplay will provide crucial insight into the function of neuronal circuits and the mechanisms of learning and memory.
During my PhD, I focused on the dissection of neuronal circuits underlying associative fear learning. Fear conditioning is one of the most powerful model systems to investigate plasticity of neuronal circuits and mechanisms of associative learning. The amygdala has been identified as a key brain structure where associative plasticity is induced through pairing of a neutral tone and a mild aversive footshock. Importantly, the amygdala is embedded in a vast network of brain structures which play distinct roles in fear learning, such as sensory cortices, the medial prefrontal cortex, and the hippocampus.
To determine the role of defined neuronal populations in these circuits, I established a broad portfolio of optogenetic techniques for monitoring and manipulating neuronal activity. Furthermore, I refined these methods, by developing hardware, genetic strategies, and viral approaches, and by synergistically integrating and combining electrophysiological recordings with light-induced perturbations of neuronal activity.
In three distinct projects, I applied these techniques to study different aspects of the neuronal fear circuitry. I revealed that distinct disinhibitory microcircuits both in the auditory cortex (AC) and the basolateral amygdala (BLA) control the acquisition of fear and that the function of BLA principal neurons (PNs) is related to the circuits in which they are embedded.
In the auditory cortex, a footshock leads to the acetylcholine-mediated excitation of interneurons (INs) in layer 1. Layer 1 interneurons, in turn, inhibit parvalbumin-expressing (PV+) INs in layer 2/3. This disinhibitory microcircuit allows for the enhanced excitation and plasticity of pyramidal neurons in response to the footshock, and is necessary for learning. The necessity of footshock-induced plasticity in AC and its disinhibitory control represent novel aspects of the circuitry underlying fear learning.
In the BLA, a different disinhibitory microcircuit controls the strength of acquired fear memories via stimulus-dependent mechanisms. During the shock, both PV+ and somatostatin-expressing (SOM+) INs are inhibited, which results in a general disinhibition of PNs along their somatodendritic axis, allowing for plasticity. During the tone however, PV+ INs are excited, while SOM+ INs are inhibited – most likely by directly connected PV+ INs. This causes dendritic disinhibition of PNs, which leads to a boosting of the impact of auditory inputs and enhanced plasticity and learning. This demonstrates that BLA PV+ and SOM+ INs exert bidirectional control over fear acquisition through differential changes in inhibition along the somatodendritic axis of PNs.
PNs in the BLA project to different target regions. This differential connectivity also relates to differences in their function. I demonstrated that BLA neurons projecting to the infralimbic or prelimbic cortex are oppositely involved in fear acquisition and extinction, and exhibit distinct plastic changes. This not only shows a relationship between function and connectivity, but also represents an example for the control of learning in neuronal circuits which span across several brain areas.
In summary, I established a combination of optogenetic and electrophysiological tools and applied them to dissect the neuronal circuits of fear. I revealed several fundamental mechanisms in fear learning and made significant contributions towards the understanding of defined inhibitory and excitatory neuronal subpopulations both in auditory cortex and amygdala.