This course will present the basics of state space modeling to analyze time series data that are frequently encountered in neuroscience problems. The course lectures will cover basics of time series analysis, Markov chains, linear state space models, switching state space models, and algorithms for learning and inference. Students and instructors will work through practical data analysis exercises in Python in weekly labs.
- Students will learn the basics of frequency domain analysis methods (i.e., spectral estimation) and understand their limitation.
- Students will appreciate the complexity of analyzing neural oscillations in electrophysiological recordings.
- Students will have a working understanding of the nuts and bolts of fitting state space models, including Kalman filtering and smoothing, the EM algorithm, and applications to neuroscience.
- Students will be able to use the SOMATA python package that stream- lines analysis of neural oscillation using the abovementioned time-domain analysis tools.
You can install the requirered environment using the env-lab.yml file.
If you run into any error during installtions, that's due to conda's failure of dependency resolution.
In that case, we recommend to use mamba. To do so, simply download miniforge from their github page
https://github.com/conda-forge/miniforge, and install.
With Miniforge in your path, you can use mamba to create the environment from the env-lab.yml
file