The study of brain lesions can benefit from a clear identification of transitions between healthy and pathological tissues, through the analysis of brain imaging data. Current signal processing methods, able to address these issues, often rely on strong prior information. In this article, a new method for tissue segmentation is proposed. It is based on a discriminative strategy, in a self-supervised machine learning approach. This method avoids the use of prior information, which makes it very versatile, and able to cope with different tissue types. It also returns tissue probabilities for each voxel, crucial for a good characterization of the evolution of brain lesions. Simulated as well as real benchmark data were used to validate the accuracy of the method and compare it against other segmentation algorithms.

In this paper, we study the separation of synchronous sources (SSS) problem, which deals with the separation of sources whose phases are synchronous. This problem cannot be addressed through independent component analysis methods because synchronous sources are statistically dependent. We present a two-step algorithm, called phase locked matrix factorization (PLMF), to perform SSS. We also show that SSS is identifiable under some assumptions and that any global minimum of PLMFs cost function is a desirable solution for SSS. We extensively study the algorithm on simulated data and conclude that it can perform SSS with various numbers of sources and sensors and with various phase lags between the sources, both in the ideal (i.e., perfectly synchronous and nonnoisy) case, and with various levels of additive noise in the observed signals and of phase jitter in the sources.

It has been proven that there are synchrony (or phase-locking) phenomena present in multiple oscillating systems such as electrical circuits, lasers, chemical reactions, and human neurons. If the measurements of these systems cannot detect the individual oscillators but rather a superposition of them, as in brain electrophysiological signals (electro- and magneoencephalogram), spurious phase locking will be detected. Current source-extraction techniques attempt to undo this superposition by assuming properties on the data, which are not valid when underlying sources are phase-locked. Statistical independence of the sources is one such invalid assumption, as phase-locked sources are dependent. In this paper, we introduce methods for source separation and clustering which make adequate assumptions for data where synchrony is present, and show with simulated data that they perform well even in cases where independent component analysis and other well-known source-separation methods fail. The results in this paper provide a proof of concept that synchrony-based techniques are useful for low-noise applications.

According to a connectionist view, mental states consist of the activations of neural units in a connectionist network. We consider the similarity of representations that emerge in unsupervised, self-organization process of neural lattices when exposed to color spectrum stimuli. Self-organizing maps (SOM) are trained with color spectrum input, using various vectorial encodings for representation of the input. Further, the SOM is used for heteroassociative mapping to associate color spectrum with color names. Recall of association between the spectra and colors is assessed. It shows that the SOM learns representations for both stimuli and color names, and is able to associate them successfully. The resulting organization is compared through correlation of the activation patterns of the neural maps when responding to color spectrum stimuli. Experiments show that the emerged representations for stimuli are similar with respect to the partitioning-of-activation-space measure almost independently of the encoding used for input representation. This adds a new example in favour of the usability of the state space semantics.

We apply a blind source separation approach to the identification of statistically independent spatial patterns of brain activation to auditory stimulation. Stimuli consisted of spoken text. The data was collected via functional magnetic resonance imaging (fMRI). As expected from standard processing of fMRI, we observe that independent component analysis (ICA) reveals spatial patterns with similar temporal activation as the stimulus. In these, ICA further distinguishes between the primary auditory areas and Broca's and Wernicke's, which are associated with speech production and understanding, respectively. Furthermore, we observe the activation of the thalamus, with a time course unrelated to the stimulus, hence hard to detect in a classical manner. We observe as well a temporally evolving artifact, related to inefficient filtering of the fMRI scans. The consistency of the estimated signals is tested by running the algorithm with many different initial conditions. The solutions found are combined according to their similarities. Estimates that differ greatly from run to run are less likely to correspond to true components, whereas those that present small variances are considered reliable ones.

In blind source separation one tries to separate statistically independent unknown source signals from their linear mixtures without knowing the mixing coefficients. Such techniques are currently studied actively both in statistical signal processing and unsupervised neural learning. We apply neural blind separation techniques developed in our laboratory to the extraction of features from natural images and to the separation of medical EEG signals. The new analysis method yields features that describe the underlying data better than for example classical principal component analysis. We discuss difficulties related with real-world applications of blind signal processing, too.

Independent component analysis (ICA) is a recently developed, useful extension of standard principal component analysis (PCA). The ICA model is utilized mainly in blind separation of unknown source signals from their linear mixtures. In this application only the source signals which correspond to the coefficients of the ICA expansion are of interest. In this paper, we propose neural structures related to multilayer feedforward networks for performing complete ICA. The basic ICA network consists of whitening, separation, and basis vector estimation layers. It can be used for both blind source separation and estimation of the basis vectors of ICA. We consider learning algorithms for each layer, and modify our previous nonlinear PCA type algorithms so that their separation capabilities are greatly improved. The proposed class of networks yields good results in test examples with both artificial and real-world data.