W. El-Deredy

Selective smoothing of the generative topographic mapping

Generative topographic mapping is a nonlinear latent variable model introduced by Bishop et al. as a probabilistic reformulation of self-organizing maps. The complexity of this model is mostly determined by the number and form of basis functions generating the nonlinear mapping from latent space to data space, but it can be further controlled by adding a regularization term to increase the stiffness of the mapping and avoid data over-fitting. In this paper, we improve the map smoothing by introducing multiple regularization terms, one associated with each of the basis functions. A similar technique to that of automatic relevance determination, our selective map smoothing locally controls the stiffness of the mapping depending on length scales of the underlying manifold, while optimizing the effective number of active basis functions.

Assessment of statistical and neural networks methods in NMR spectral classification and metabolite selection

Magnetic resonance spectroscopy opens a window into the biochemistry of living tissue. However, spectra acquired from different tissue types in vivo or in vitro and from body fluids contain a large number of peaks from a range of metabolites, whose relative intensities vary substantially and in complicated ways even between successive samples from the same category. The realization of the full clinical potential of NMR spectroscopy relies, in part, on our ability to interpret and quantify the role of individual metabolites in characterizing specific tissue and tissue conditions. This paper addresses the problem of tissue classification by analysing NMR spectra using statistical and neural network methods. It assesses the performance of classification models from a range of statistical methods and compares them with the performance of artificial neural network models. The paper also assesses the consistency of the models in selecting, directly from the spectra, the subsets of metabolites most relevant for differentiating between tissue types. The analysis techniques are examined using in vitro spectra from eight classes of normal tissue and tumours obtained from rats. We show that, for the given data set, the performance of linear and non‐linear methods is comparable, possibly due to the small sample size per class. We also show that using a subset of metabolites selected by linear discriminant analysis for further analysis by neural networks improves the classification accuracy, and reduces the number of metabolites necessary for correct classification.

Changes in neural complexity of the EEG during a visual oddball task

OBJECTIVE Neural complexity (C(N)) was introduced by Tononi et al. in an information-theoretic framework to capture the balance between functional specialisation and integration in the brain. We hypothesised that C(N) should vary during cognitive processing, specifically during an oddball task. METHODS In 11 normal human subjects, we recorded from groups of EEG electrodes in the frontal (F), central-parietal (CP) and occipito-temporal (OT) regions during a visual oddball reward conditioning task and calculated C(N) in each region. Three types of visual stimulus (abstract shapes, called neutral, reward and penalty) were presented randomly in three blocks of trials. During the first block, subjects did not know the significance of the stimulus shapes. For the subsequent (conditioning) blocks, subjects were informed that whenever they saw reward or penalty patterns, they would win or lose money, respectively. RESULTS In regions CP and OT, C(N) was significantly larger in reward and penalty trials than in neutral during all blocks. During a trial, significant changes in C(N) occurred around the ERP peaks N1 and P300 and the effects of reward conditioning on C(N) could be distinguished from penalty. CONCLUSIONS Our findings support the above hypothesis, indicating that C(N) correlates with both the sensory and cognitive components of stimulus processing. SIGNIFICANCE This study extends the scope of C(N) in the analysis of cognitive processing.

Dipole source localisation using independent component analysis: single trial localisation of laser evoked pain

The accuracy of the inverse solution that finds the spatial location of the generating sources from averaged scalp-recorded event related potentials (ERPs) relies on assumptions about the ERP signals and the sources. We provide evidence that using independent component analysis (ICA) as a signal decomposition filter prior to applying the inverse solution reveals sources that cannot be detected by conventional source localisation methods. Five clusters of sources emerged: a single source cluster in caudal cingulate and bilateral sources in secondary somatosensory cortex (SII), inferior parietal cortex, premotor cortex and insular cortex. The locations of the source dipoles were consistent with findings using fMRI and PET but have not all been previously detected in a single electrophysiological study. In addition, the time-course of the activation of these dipoles was estimated. The results suggest that using ICA to localise single trial data is a powerful tool for exploring the spatiotemporal dynamics of rapid and complex brain processes.

Tissue characterisation with NMR spectroscopy: current state and future prospects for the application of neural networks analysis

Nuclear magnetic resonance (NMR) spectroscopy has considerable potential for non-invasive characterisation of tissue biochemistry and the diagnosis of tissue abnormalities, ranging from focal lesions in the brain, to tumours in any area of the body to assessing effect of HIV damage. However, the realisation of the full clinical potential NMR spectroscopy will depend on extracting information from the spectra directly and specifically related to the biochemistry of different tissue types under various normal and pathological circumstances. This paper reviews the progress made in the application of neural network analysis to the automatic characterisation of NMR data, raising some key issues and providing a perspective of the future of this technology.