Caveats with stochastic gradient and maximum likelihood based ICA for EEG

Abstract : Stochastic gradient (SG) is the most commonly used optimization technique for maximum likelihood based approaches to independent component analysis (ICA). It is in particular the default solver in public implementations of Infomax and variants. Motivated by experimental findings on electroencephalography (EEG) data, we report some caveats which can impact the results and interpretation of neuroscience findings. We investigate issues raised by controlling the step size in gradient updates combined with early stopping conditions, as well as initialization choices which can artificially generate biologically plausible brain sources, so called dipolar sources. We provide experimental evidence that pushing the convergence of Infomax using non stochastic solvers can reduce the number of highly dipolar components and provide a mathematical explanation of this fact. Results are presented on public EEG data.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [17 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01451432
Contributor : Jair Montoya Martinez <>
Submitted on : Wednesday, February 1, 2017 - 10:32:05 AM
Last modification on : Wednesday, February 20, 2019 - 1:28:51 AM

File

paper.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01451432, version 1

Citation

Jair Montoya-Martínez, Jean-François Cardoso, Alexandre Gramfort. Caveats with stochastic gradient and maximum likelihood based ICA for EEG. Latent Variable Analysis, Independent Component Analysis LVA-ICA International Conference, Feb 2017, Grenoble, France. ⟨hal-01451432⟩

Share

Metrics

Record views

366

Files downloads

443