ИСТИНА

The activity of human brain at rest is often suggested to reflect conscious processing of past experience. Here we show that past experience can shape the resting-state neuronal networks of a conscious mouse. Initially we characterized resting-state activity of 104 brain structures in naive mice by means of large-scale c-Fos cellular mapping. Then we modeled post-traumatic stress disorder (PTSD) in mice using a short footshock stress, and imaged resting-state networks in PTSD and naive (non-stressed) mice. The c-Fos activity of 42 selected brain areas (sensory and motor neocortical areas, hippocampus, parahippocampal cortex, amygdala, basal nuclei, associative and sensory thalamic nuclei, hypothalamic nuclei, and midbrain) was investigated in naive mice and mice that underwent footshock stress 7 days earlier. PTSD induction strongly affected subsequent resting-state brain activity: mice with prior traumatic experience had significantly more c-Fos+ cells in cingulate, retrosplenial, parietal associative and entorhinal cortices, basolateral and lateral amygdala, paraventricular thalamic nucleus and periaqueductal gray. Earlier these areas were shown to be involved in fear networks in humans and animals. Next, we used graph theory approach to reconstruct resting-state network connectivity of naive and PTSD mice and to define main clusters of resting-state networks. We compared experimentally identified networks with model networks: random, scale free and small world. Clustering of experimental networks of both groups of mice was at the same level as in scale free network - that is, the number of clusters exceeded the random level. At the same time, these clusters had weak interactions with each other: global efficiency of experimental networks was at the same level as of a random network. Resting-state networks of naive and PTSD mice were different: PTSD network was less clustered, and the clusters were divided by longer routes than in naive mice. Analysis of functional connectivity revealed that PTSD induction caused global changes in the resting-state network structure, which affected all studied brain areas. Overall, while naive mice had most of the connections between the cortical areas, PTSD mice had the majority of connections in areas of thalamus, striatum and amygdala. PTSD induction eliminated almost all functional connections present in naive mice; the only cluster to survive was the closely connected cluster of visual and auditory cortical areas. Furthermore, while cingulate and retrosplenial cortices were the main network hubs in naive mice, functional connectivity between those areas was lost in PTSD mice, and paraventricular thalamic nucleus became a hub. Conversely, while functional connections of amygdala were almost absent in naive mice, in PTSD mice there was a substantial number of connections between amygdala, cortical associative areas and striatum. Our data suggest that experience of stressful event can change both resting-state spontaneous activity and resting-state functional connectivity patterns in the mouse brain long after the traumatic episode. We hypothesize that these changes reflect a replay of neuronal assemblies involved in the states of past subjective experience. In the next series of experiments, we plan to test this hypothesis by using a targeted recombination in active populations (TRAP) technique. Supported by RSF #16-15-00300, 14-15-00685.