Jaakko Särelä

Independent component approach to the analysis of EEG and MEG recordings

Multichannel recordings of the electromagnetic fields emerging from neural currents in the brain generate large amounts of data. Suitable feature extraction methods are, therefore, useful to facilitate the representation and interpretation of the data. Recently developed independent component analysis (ICA) has been shown to be an efficient tool for artifact identification and extraction from electroencephalographic (EEG) and magnetoencephalographic (MEG) recordings. In addition, ICA has been applied to the analysis of brain signals evoked by sensory stimuli. This paper reviews our recent results in this field.

Independent Component Analysis in Wave Decomposition of Auditory Evoked Fields

In this paper we introduce a novel approach to the problem of decomposition of auditory evoked fields (AEF), measured by magnetoencephalography (MEG), into basic components. This approach is based on independent component analysis (ICA), that separates components according to the kurtosis of their amplitude distribution over time. The fixed-point algorithm used extracts one independent component at a time, allowing the combination of a high resolution 122-channel whole-scalp neuromagnetometer, to a fast and very efficient implementation of ICA.

Separation of nonlinear image mixtures by denoising source separation

The denoising source separation framework is extended to nonlinear separation of image mixtures. MLP networks are used to model the nonlinear unmixing mapping. Learning is guided by a denoising function which uses prior knowledge about the sparsity of the edges in images. The main benefit of the method is that it is simple and computationally efficient. Separation results on a real-world image mixture proved to be comparable to those achieved with MISEP.

The Problem of Overlearning in High-Order ICA Approaches: Analysis and Solutions

We consider a type of overlearning typical of independent component analysis algorithms. These can be seen to minimize the mutual information between source estimates. The overlearning causes spikelike signals if there are too few samples or there is a considerable amount of noise present. It is argued that if the data has flicker noise the problem is more severe and is better characterized by bumps instead of spikes. The problem is demonstrated using recorded magnetoencephalographic signals. Several methods are suggested that attempt to solve the overlearning problem or, at least, diminishlreduce its effects.

Accurate, Fast and Stable Denoising Source Separation Algorithms

Denoising source separation is a recently introduced framework for building source separation algorithms around denoising procedures. Two developments are reported here. First, a new scheme for accelerating and stabilising convergence by controlling step sizes is introduced. Second, a novel signal-variance based denoising function is proposed. Estimates of variances of different source are whitened which actively promotes separation of sources. Experiments with artificial data and real magnetoencephalograms demonstrate that the developed algorithms are accurate, fast and stable.

Searching for Independence in Electromagnetic Brain Waves

Multichannel recordings of the electromagnetic fields emerging from neural currents in the brain generate large amounts of data. Suitable feature extraction methods are, therefore, useful to facilitate the representation and interpretation of the data.