Dimensionality reduction of large-scale neural recordings
Dimensionality reduction of large-scale neural recordings
Methods for dimensionality reduction of data, with focus on factor analysis.
Topics covered in this lesson
- Large noisy multi-channel datasets from new methods of large scale neural recording.
- Covariation among channels.
- Can we summarize the data using a smaller number of variables? Binning and resolution.
- Dimensionality reduction.
- Denoising.
- Dimensionality reduction models: principal component analysis (PCA), factor analysis (FA), Gaussian-process factor analysis (GPFA), latent dynamical systems, non-linear systems.
- Explicit noise models.
- Closer explanation of factor analysis, with computations.
External Links
Prerequisites
Linear algebra (matrices). Some basic statistics (probability and conditional probability).
Back to the course