Brain network function during shifts in learning strategies in portal hypertension animals

Patients with minimal hepatic encephalopathy exhibit early impairments in their ability to shift attentional set. We employed a task-switching protocol to evaluate brain network changes. Strategy switching requires the modification of both the relevant stimulus dimension and the required memory system. Rats were trained in an allocentric (A) and a cue-guided (C) task using a four-arm maze. To examine priming, we changed the order in which the tasks were presented. Five groups of animals were used: a SHAM (sham-operated) A-C group (n=10), a SHAM C-A group (n=8), a PH (portal hypertension) A-C group (n=8), PH C-A group (n=8), and a naïve group (n=10). The triple portal vein ligation method was used to create an animal model of the early evolutive phase of PH. The animals were tested in the four-arm radial water maze in a single 10-trial session each day for six days (three days for the allocentric task and three days for the cue-guided task). The metabolic activities of the brains were studied with cytochrome oxidase histochemistry, and brain network changes were assessed with principal component analysis. The behavioural results revealed significant increases in the numbers of correct choices across training days in all groups studied, and facilitation of the acquisition of the second task was present in the C-A groups. Moreover, different brain network activities were found; in the experimental groups, the performance of A-C switch involved the prefrontal cortex, and the key structures involved in the C-A switch in the other groups were the dentate gyrus of the dorsal hippocampus and the basolateral and central amygdala. These networks have a common nucleus of structures (i.e., the parietal cortex and the dorsal and ventral striatum), whereas other structures were specifically involved in each type of strategy, suggesting that these regions are part of both circuits and may interact with one another during learning.