Hitchhiker's guide to analyzing time-series data in neuroscience

Part I: Overview (slides)

Course audience
Why analyzing time-series data is important
Challenges of analyzing time-series data
Course goals
Course outline

Part II: Fundamental concepts (slides)

Data types in neuroscience
Mathematical models
Discriminative vs generative models
Parameters, fitting and overfitting
Model comparison
Random processes perspective
Dynamical systems perspective

Part III: Philosophy (slides)

Frameworks, theories, and models
Descriptive, mechanistic, and normative theories
Fundamental theorem of data analysis
Theory- vs data-driven approaches
The value of predictions
Transforming data into science
McNamara fallacy

Part IV: Common issues (slides)

Lack of trials
Controlled vs naturalistic data
Nonstationarity/long timescales
Trials/sessions/animals/conditions
Binning
Missing data
Too much data
Multicollinearity
Gotchas: normalization, correlated training/test data, whats is N?
Interpreting data-driven analyses

Part V: Practical tips and tricks (slides)

Approaching/vetting a new dataset
Data pre-processing
Data munging/storage
Control datasets
Desconstructing fit models
Statistical testing
How to not make mistakes
Leveraging LLMs
Designing custom analyses
Coding strategies
Reproducibility
Data sharing

Part VI: Canonical methods (slides)

Pre-processing:
    Filtering/smoothing
    Detrending
    Spike sorting
    Estimating firing rates
    Image processing/ROI extraction
    Behavior keypoint tracking
    Segmenting

Classical signal processing:
    Correlation functions
    Fourier transforms and power spectra
    Linear filters and impulse response

Neural data analysis:
    Raster plots and inter-spike interval distributions
    Peristimulus time histograms (PSTHs)
    Tuning curves
    Linear-nonlinear-Poisson (LNP) and generalized linear models (GLM)
    Spike-triggered average (STA)

Part VII: Modern methods (slides)

Dimensionality reduction
Clustering/segmentation
Statistical models
Latent variable models
Dynamics/state-space models
Mechanistic models
Connectome-based models
Spatiotemporal analyses
Artificial/deep neural networks
Variational inference/ELBO
Extended learning/inference techniques
Information theory
Validation metrics

Part VIII: Miscellaneous (slides)

Other methods
Common software
Frequently asked questions
Other courses/resources