|M.Sc Student||Saar Sigal|
|Subject||Interaction between Adjacent Neurons in the Behaving Monkey|
|Department||Department of Biomedical Engineering||Supervisor||Professor Moshe Gur|
Neuronal assemblies are an important subject to study as a first step toward understanding the neural code. It is of interest to study neuronal interaction in V1 since they are easily observed and since V1 is the first cortical processing stage. The objective is to study small neuronal assemblies in V1 in order to gain a better understanding of more complex interactions.
To achieve those objectives, mathematical tools must be implemented according to the nature of the gathered data. Data that were gathered during experiments with behaving monkeys have distinguishing characteristics. Those features enable the optimization and minimization of known algorithms in order to get a better understanding of the neuronal interactions.
Tools are presented for assessing the connectivity of two neurons. Joint Pre-stimulus Histogram is enhanced by computing its singular value decomposition, thus achieving a better measurement of the amount of connectivity between the two neurons. Those tools were implemented on data gathered from a joint cell model and from V1 in the behaving monkey.
The validity of the tools presented here was tested using a Joint IPFM model that was developed for this purpose. Then the tools were used on experimental data streams.
Evidence of connectivity between many neuronal pairs was observed. Using the mathematical tools, developed especially for this work, effective connectivity scheme was evaluated. The most interesting finding was that the JPSTH figures originating from the experimental data streams were more interesting and meaningful than the JPSTH figures evaluated by the simulation. This was validated by the singular value decomposition analysis. It is an indication that even connections between adjacent cells are very complicated, involving many levels of neuronal processing. Figures from the Joint leaky IPFM model did not reach such a level of complexity.