About Me
Carmen Araujo
I’m a senior at Boston University majoring in Neuroscience and minoring in Computer Science, so I’m part of both the Neuroscience department and the Computer Science department.
email: sabrinaa “at” bu “dot” edu
About My Mentor
Dr. Sanmi Koyejo
Dr. Koyejo is an assistant professor and researcher at the University of Illinois at Urbana-Champaign.
Area of interest: Neuroinformatics, Neuroimaging, Machine Learning
About My Project
My research project is concerned with working on Hidden Markov models for analyzing ECOG data.
Phase Coupling (PhC) and Amplitude Coupling (AmpC) are two important brain connectivity measures used when analyzing electrophysiological data. PhC quantifies synchrony of phases of two oscillatory signals over time, while AmpC measures similarity of strength of those signals. A previous study showed that resting state AmpC and PhC only relate in baseline organization, but respond distinctly to stimuli. We wanted to investigate this further by focusing on dynamic Functional Connectivity. Functional Connectivity (FC) is defined as temporal dependence of brain signals between pairs of distinct brain areas, while FC dynamics relate to changes in FC over time. Studies have reported that FC dynamics are made up of a mixture of several hidden brain states components. We aimed to disentangle these components of AmpC and PhC dynamics and better understand how they relate.
The big goal was to implement a pipeline in MATLAB able to extract brain states from existing FC dynamics. In this project, the intention was to use the Hidden Markov Model method to disentangle FC dynamics into several perpendicular brain state dynamics.
This project entails:
- Literature review on the introduction to HMM and the concept of brain states
- Getting familiar with existing data preprocessing pipelines and with the structure of existing FC dynamics
- Implementing HMM to extract the brain states
- Visualizing said results and interpreting them
- Statistically analyzing the results