Modeling the cortical-hippocampal connectivity to resemble in vitro Multi-Electrode recordings.

There are experimental evidences that suggest that memory consolidation, retrieval and memory-guided decisions are reinforced by the flow of information between the prefrontal cortex and the hippocampus. In particular, it is thought that the cortex is involved in memory retrieval while the hippocampus has a stronger role in memory encoding.
The objective of this work is to further develop an existing computational model of the prefrontal cortex and the hippocampus. Subsequently, the model has been fit with data recorded in vitro from cortical and hippocampal rat tissues obtained in the Bioengineering Laboratory in the University of Genoa. In particular, two connectivity schemes were compared using the model to predict the types of connections that have formed in vitro. In one case the model’s connections were implemented using a gaussian probability distribution depending on the cell’s relative distance, favouring short-range connections and clustering. The other situation was to generate randomly distributed connections, thus allowing also long-range ones. The aim is to better understand the underlying mechanisms that generate the observed electrophysiological behaviour in vitro. This thesis falls into the domain of computational neuroscience.
It is believed that new memories are formed in the hippocampus and subsequently transferred to the cortex for long-term storage during slow oscillations in sleep. Therefore, the cortical network model was built in order to generate slow oscillations. On the other hand, sharp wave ripples are a characteristic electrophysiological hippocampal pattern. Recent understanding suggests that, during deep sleep, sharp-wave ripples participate in the mechanism of memory transfer from the hippocampus to the cortex for long-term storage. Thus, the hippocampal network developed aims to reproduce this type of activation.
It has been found that the random connectivity scheme is more adequate to represent the experimental data, characterized by rhythmic, synchronized firing patterns. This could indicate that in vitro long-range connections are preferred to near neighbours connectivity.