Generalized Linear Model (GLM) |
| • | GLMspiketraintutorial
- tutorial code and slides from 2016 SFN short course, illustrating basics of Gaussian and Poisson GLMs for spike train
data. [zip
| readme]
|
| • | GLMspiketraintutorial_python
- python version of the tutorial above (NEW!)
|
| •
|
neuroGLM - Poisson GLM for
single-neuron trial-based data
[relevant pub: IM Park et al, Nat Neurosci 2014]
|
| •
|
GLM dynamical
behaviors - capturing repertoire of
Izhikevich neuron
[relevant pub: Weber
& Pillow, 2018]
|
| •
|
GLMspiketools - multi-neuron Poisson
GLM with spike-history and coupling between neurons
[relevant pub: Pillow et al, Nature 2008]
|
| •
|
polynomial
approximate GLM - fast, approximate inference for
Poisson GLM for large-scale datasets
[relevant pub: Zoltowski et al, NeurIPS 2018]
|
Receptive field estimation |
| • | fastASD
- fast, scalable estimation for high-D receptive fields via automatic
smoothness determination (ASD).
[relevant pub: Aoi & Pillow, bioRxiv 2017]
|
| •
|
Automatic Locality
Determination (ALD) - receptive field estimation under
localized priors
[relevant pub: M Park & Pillow, PLoS CB 2011]
|
Multi-filter LNP model |
| •
|
LNPfitting -
maximum likelihood / maximally informative dimensions (MID)
estimator for linear-nonlinear-Poisson (LNP) model
.
[relevant pub: Williamson
et al, PLoS Comp Biol 2015]
|
| •
|
convolutional
STC - convolutional spike-triggered covariance analysis for
subunit models
[relevant pub: Wu et al,
NeurIPS 2015]
|
| •
|
iSTAC - information-theoretic
spike-triggered average & covariance for multi-filter LNP model
[relevant pub: Pillow & Simoncelli, J Vision 2006]
|
non-Poisson spiking models |
| •
|
flexibleModulatedPoisson -
extends modulated Poisson model for over-dispersed spike count data [Goris
et, 2014] to account for flexible relationships between mean and variance
[relevant pub: Charles
et al, Neural Comp
2018]
|
| •
|
Generalized Integrate-and-Fire
model - ML fitting of spike-response model to spike train data via Fokker-Planck
[relevant pub: Pillow et al, J Neurosci 2005]
|
Stepping/Ramping model comparison |
| •
|
StepRampMCMC - Model comparison for stepping and
ramping spike train models
[relevant pubs: Latimer et al, Science
2015, Zoltowski et al,
Neuron 2019]
|
Structured sparsity / fMRI decoding |
| •
|
Dependent relevance
determination (DRD) - Inference for sparse, smooth, clustered support
regression weights.
[relevant pubs: Wu et al, JMLR 2019]
|
Supervised Dimensionality Reduction
(w/ applications to fMRI) |
| •
|
Class Factor-Analytic
Dimensions (CFAD) -
identify low-dimensional projection that preserves
information about an external covariate of interest. (Problem
setting also known as "sufficient dimension reduction").
[relevant pub: Jha,
Morais & Pillow, ICML 2021].
|
Adaptive stimulus selection / Closed-loop experiments |
| •
|
adaptivePsychophysicsToolbox -
adaptive stimulus selection for characterizing psychometric
functions, incorporating multiple response options and lapse
[relevant pub: Bak
& Pillow, JOV 2018]
|
| •
|
adapotiveOptimalTraining -
adaptive stimulus selection for optimal animal training ("AlignMax").
[relevant pub: Bak
et al, NeurIPS 2016]
|
| •
|
activelearningTCs -
adaptive stimulus selection for neural tuning curves with Poisson noise
[relevant pub: Park
& Pillow book chapter, 2016 ]
|
| •
|
active_learning_GLM_HMMs -
adaptive stimulus selection for discrete latent variable models,
including GLM-HMM
[relevant pub:
Jha,
Ashwood, & Pillow, Neural Comp 2024]
|
Psychophysics / Behavior modeling |
| •
| GLM-HMM -
Hidden Markov Model (HMM) with Bernoulli generalized linear
model (GLM) for identifying latent states from perceptual decision-making data
[relevant pub: Ashwood et al, Nat. Neurosci2022]
|
| •
| psyTrack -
dynamic psychophysical model for tracking weight changes during
learning.
collab
notebook - reproduces all figs from paper.
[relevant pubs:
Roy et al, Neuron 2020,
Roy et al, NeurIPS 2018]
|
| •
| Kalman
filter for target tracking psychophysics - continuous
psychophysics via target tracking
[relevant pub: Bonnen et al, Journal of Vision 2015]
|
Entropy estimation |
| •
|
CDMentropy - entropy estimation for binary
spike trains using centered-Dirichlet mixture (CDM) priors
[relevant pub: Archer,
Park, & Pillow, NeurIPS 2013]
|
| •
|
PYMentropy - entropy estimation
for discrete distributions with unknown support using Pitman-Yor Mixture (PYM) prior
[relevant pub: Archer,
Park & Pillow, JMLR 2014]
|
Spike sorting |
| •
|
BinaryPursuitSpikeSorting - Binary Pursuit spike sorting for
detecting synchronous and overlapping spikes
[relevant pub: Pillow et al, PLoS ONE 2013]
|
General-purpose utilities |
| •
| automaticRidgeRegression
- Empirical Bayes inference for ridge regression and smooth ridge
regression. Performs evidence-optimization for the ridge
parameter (and smoothness parameter, if desired) using fixed point
update.
|
| •
| multilinearRegression -
Bilinear and trilinear least-squares regression using alternating coordinate ascent or joint optimization.
|
| •
| GaussHermiteQuadrature
-
Performs Gauss-Hermite quadrature with a Gaussian density. Useful for
efficiently computing the integral of a smooth function multiplied by a Gaussian.
|
| •
| invBlockTriDiag -
Efficiently compute central and off-diagonal blocks of the inverse of a block-tridiagonal matrix.
|
| •
| raisedCosineBasis -
Makes raised cosine basis with log scaling of x axis (for
parametrizing stimulus or spike-history filters)
|
