Loading...

State Space Reconstruction with Application in Revealing the Nonlinear Dynamics of Brain

Heydari, Mohammad Reza | 2021

263 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54403 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Tavazoei, Mohammad Saleh; Ghazazideh, Ali
  7. Abstract:
  8. Learning is an essential mechanism for the survival of living things. There are different types of learning, and value learning is among the most important types. A child learns that water resolves the thirst need by repeatedly experiencing this situation. Eventually, the value of water, which has been valueless before that, increases gradually in his mind. How this concept is encoded in the brain? previous works reveal the role of different neurons and regions that are relevant to value learning. However, population analysis and dynamic modeling are less considered. Moreover, the links between different brain regions are unknown.Finding the relationship between two relevant regions of the brain is the main focus of this work. The task is a simple visual task that subjects gradually learn the value of fractals with receiving juice. The difficulty of the electrophysiological data recording makes it too problematic to answer this question, and we must first overcome the challenges of data. The results are so promising compared to the traditional methods. At first, we proposed a model that estimates the values of fractals at each time. Then, using these estimates and state-of-the-art data analysis tools, the dynamical model of the neurons is obtained that reveals some new aspects of the two regions. Finally, the relationship between two regions ventrolateral Prefrontal Cortex (vlPFC) and Substantia Nigra pars Reticulata (SNr) is obtained using the causal models, that is compatible with previous works. We used the following methods in our study: population analysis, reinforcement learning, deep learning, and causal inference
  9. Keywords:
  10. Time Series Analysis ; Reinforcement Learning ; Recurrent Neural Networks ; Causal Inference ; Population Analysis

 Digital Object List