Temporal organization of neuronal ensembles encoding behavioral states

Innate behaviors, such as feeding, social interaction, and exploratory behaviors, are instinctive actions essential for the survival and health of humans and animals. The mechanism that enables switching between these behaviors occurs on a short (seconds) timescale, allowing for timely transitions by processing information about the internal states and external sensory cues. This enables the prioritization of actions and switching between them. Innate behaviors are controlled by the hypothalamus, with its lateral aspects prominently interfacing with key brain regions that support adaptive behavior. In this thesis, I investigated the hitherto unknown underlying mechanisms of behavioral transitions. Applying advanced computational methods to the in vivo electrophysiological dataset obtained in freely behaving mice, I observed that the discharge of lateral hypothalamus (LH) cells increased and became entrained during beta (15-30 Hz) oscillations. Before the transition to a new innate behavior, a large proportion of LH cells with high firing rates during the three aforementioned behaviors became active near the peak of the beta oscillations, while typically LH cells have a tendency to fire near the trough of these oscillations. This prominent beta peak activity was evident exclusively during behavioral transitions, but not before or at the onset of behaviors, and it was not observed during gamma (30-60 Hz) oscillations. The beta phase-specific activity of LH cells during transitions was sufficient to decode both the timing and the type of upcoming behavior. Inhibition of LH cells during these phases, via the lateral preoptic area, disrupted the typical phase-specific activity associated with transitions, diminishing the occurrence and prolonging the duration of behaviors. A subset of LH cells with a high firing during all three behaviors strongly contributed to phase coding during transitions; referred as ‘transition’ cells. These cells spontaneously synchronized with medial prefrontal cortical (mPFC) cells at the peak of beta oscillations during transitions to feeding and social interaction but not at other times, suggesting that mPFC-LH signaling facilitates encoding of upcoming behaviors. Moreover, the discharge of putative dopamine (DA) neurons in the ventral tegmental area (VTA) increased during LH beta oscillations. DA cells with high firing rates during social and exploratory behaviors exhibited similar beta peak phase-specific activity to LH cells during transitions, while DA cells with high firing rates during feeding showed a half-cycle phase offset, i.e., firing near the trough of the beta oscillations. Thus, a transition state is encoded in the hypothalamus, enabling synchronization with reward-processing and cognitive circuits, signaling potential future behavior. Presented here results and other studies have shown changes in the activity of LH neurons depending on behaviors compared to baseline, but their firing rate dynamics within and between these behaviors have remained understudied. By analyzing the firing rate dynamics of LH cells during innate behaviors using an unsupervised approach without behavioral information, distinct groups of cells were revealed to be sequentially active at different times during feeding bouts. Other subpopulations of LH cells were found that are active temporally homogeneously throughout feeding, social interaction or novel object exploration. The feeding-related cells exhibited distinct electrophysiological properties, particularly those active in the second half of feeding bouts, which had lower average firing rates and longer action potentials compared to those more active in the first half. This suggests a possible enrichment of orexin and GABA cells during different phases of feeding. Moreover, the feeding-related cells showed enhanced coordination during slow gamma (30-60 Hz) oscillations compared to non-rhythmic epochs. These observations reveal the contribution of different LH populations during appetitive behaviors and stages of consummatory acts. Dysfunction in their interaction and plasticity during network oscillations may contribute to eating disorders.