Neural coding of fear memories formation and extinction in frontal association network of the behaving rodent – FEAR-FRA

Our project combine a wide range and unique set of optical, electrophysiological, genetic, and optogenetic tools in vivo in the behaving animal to understand how computations are performed in prefrontal neuronal microcircuit during fear lerning. We developped and implemented unique tools at the cutting edge of technology and innovation. Longitudinal in-vivo two-photon imaging and whole-cell recordings in the behaving animal are indeed at the forefront of the modern neuroscience techniques. When combined with the high temporal precision of optogenetics, calcium dynamics makes it possible to decipher the causal role of the interaction between different interconnected brain structures during behaviors.

Here, we first show that fear learning, a widely used form of predictive learning, is strongly affected by optogenetic inactivation of the dorsal prefrontal cortex (dPFC). Then, by combining in vivo two-photon large-scale neuronal calcium imaging and whole-cell recordings in behaving mice, we observe that layer II dPFC pyramidal neurons are activated upon sound presentation. This low and frequency-independent activation may act as an alert system with rapid habituation upon sound re-presentation. Interestingly, when the same sound is associated in time to an aversive stimulus (e.g. a foot-shock), the subsequent activation of dPFC neuronal network is strongly reinforced and synchronized upon sound presentation with increased spiking reliability across sessions and days suggesting that behavioraly-relevant pattern of activity have been stabilized following fear learning. Altogether our data show that the dPFC plays key role in the formation and expression of fear memory traces. In addition we hypothesize that the dPFC acts as a top-down warning structure allowing the animal to detect biologically important events such as downstream fear association by learning about signals of their occurrence.

Our “learning under the microscope” strategy tackles several outstanding issues that have never been addressed in the past most likely because of technical limitations. It will indeed give special insight into the higher-cognitive functions performed by the prefrontal cortex, including the cellular and synaptic mechanisms underlying fear memories formation, as well as the integration of cortical and subcortical valuation systems that may participate in fear learning. Studying such interactions is of crucial importance as many maladaptive behaviors in our daily life (e.g. impulsivity; anxiety, high-risk gambling…) may arise from their dysfunctional use. As a consequence it will certainly help to advance the design of new diagnosis framework and strategies towards behavioral enhancement in a wide range of neuropsychiatric diseases.

I created the new M.I.N.D. platform that brings together in the same place multiple in vivo skills (surgery, behaviors, 2P imaging, optogenetic, electrophysiology…). It offers a unique environment in Bordeaux and in France that will help us to decipher the neuronal underpinnings of complex behavior in the prefrontal cortex of mice.

The remodeling of the adult neuronal circuits is a fundamental property of the mammalian brain, enabling it, for example, to adapt to new environmental stimuli, to acquire new skills and to store information. Learning-dependent cortical plasticity likely relies on subtle changes in the spatial pattern and level of neuronal activity that may depend on connectivity rearrangement through synapses turnover and/or activity-dependent strengthening and weakening of pre-established synapses.
The idea that dendritic spines generation and loss are potential substrates for long-term storage of memory traces mostly comes from in vivo two-photon long-term structural and functional imaging of pyramidal neurons located in primary sensory cortices in response to gross manipulations of sensory inputs (e.g. monocular deprivation or whisker trimming). Under baseline conditions only a very small percentage (~5%) of newly generated spines become persistent, and a similar small fraction of the persistent spines is lost. Whisker trimming and monocular deprivation enhances both the gain and loss of persistent spines (~13-15%) leaving the total spine density unchanged. Accordingly, it has recently been reported that overall population activity in the somatosensory cortex of adult mice re-stabilized in favor of the spared sensory modality, indicating that: 1) spine turnover may lead to stable configuration changes in the neuronal network; 2) new persistent synapses are very likely to be associated with specific circuits that drive sensory experience. Whether this dendritic spine turnover-induced rewiring represents a ubiquitous correlate of learning has been recently strengthened by in vivo imaging evidences showing robust elimination and formation of clustered cortical dendritic spines upon motor skill training and associative fear conditioning.
The latter observation is at the basis of the questions that are raised in this grant application. We hypothesize indeed that the structural-functional cascade induced by drastic sensory deprivation is similar to those employed in higher cognitive cortical areas (e.g. frontal cortex) when an animal learns a specific motor task or a stimulus association. The prefrontal cortex of rodents and humans participates in the association of temporally separated events, with damage to this structure resulting in an inability to associate temporally disconnected stimuli. Accordingly, the prefrontal cortex has recently emerged as a strong fear modulatory system. Amongst all the prefrontal areas, mounting evidence implicates prelimbic and infralimbic areas in fear expression and extinction, respectively, supporting the notion that two distinct neuronal circuits might serve the consolidation and erasure of fear traces. However, it has recently been reported that inactivation of the frontal association cortex (FrA) alters both fear conditioning and extinction. By combining in vivo two-photon long-term structural imaging of dendritic spines, large-scale neuronal calcium imaging, and whole-cell recording in the behaving animal, we aim at understanding how the formation and extinction of associative fear memories are implemented by neuronal circuits in rodent frontal association cortex (FrA), which is reciprocally connected to the amygdala, participates in fear memories, and is accessible for in vivo two-photon imaging and whole-cell recordings.
By combining long-term structural imaging and large-scale neuronal activity recording in the behaving animal, we will come to the brick of understanding the causal relationship between structural synaptic network dynamics and functional brain plasticity, which is one of the major unresolved challenges in neurosciences. Studying such anatomical-functional adaptive plasticity in the neocortex might gain insight in general mechanisms learning, and potentially help to advance the design of new strategies towards memory enhancement and neuronal repair in a wide range of brain diseases.