- / -

- / -

No description available.

Characterising the relationships between conscious and unconscious processes is one of the most important goals in cognitive neuroscience. Behavioural studies as well as neuroimaging techniques have been conducted to understand the nature of conscious perception in the brain. Functional brain imaging and EEG (Electroencephalogram) methods allow for detailed exploration of neural and computational correlates of conscious and unconscious cognition. Using a high density EEG dataset, recorded from 129 electrodes over the scalp, we studied the neural responses of the brain to auditory stimuli. To this end, we employed an auditory oddball paradigm, called the local-global experiment. Bekinschtein et al (2009) designed this experiment to explore the neural dynamics at the early auditory cortex, associated with the MMN (mismatch negativity) component, generated by the local violation of auditory stimuli. They also investigated a later novelty response, associated with the P3 (a late positive response) component, which was generated by the global violation of auditory stimuli. Their findings suggest that the global response, corresponding to working memory updating, independently from the local response, is a signature of conscious processing. But our investigations shows that the local and global effects are not fully independent from one another. Therefore, we looked for other potential signatures of conscious processing. To do this, we studied 18 healthy participants who had been sedated. Using SPM (Statistical Parametric Mapping), which is a mass univariate approach, we analysed the sedation dataset in an omnibus statistical setting. We found an interaction between the local and global effects. In addition, we investigated the impact of sedation on both the early and late temporal components (i.e. the local and global effects), and their interaction. In addition to SPM analysis, we performed single-trial analysis. Unlike SPM analysis, which explores ERPs (average effect across replications) to assess significance, single-trial analysis looks for variation across replications, from one experimental level to another. More specifically, we looked at amplitude variation and temporal jitter when participants are sedated versus recovered. In the cases, when the null hypothesis is not rejected (i.e. there is no significant difference across different levels), we calculated Bayes factors to search for evidence in favour of the null hypothesis. With the exception of latency dispersion under dual (global and local) deviance, we could find no evidence for increased variability in single trial responses under sedation. This suggests the effects of reduced conscious level are systematic and can be summarised as an attenuation of dependency of (or interaction between) local and global processing.

Structural Equation Modeling is a common name for the statistical analysis of Structural Equation Models. Structural Equation Models are models that specify relationships between a set of variables and can be specified by means of path diagrams. A number of Structural Equation Modeling programs have been developed. These include, amongst others, AMOS, EQS, LISREL, Mx, RAMONA and SEPATH. A number of studies have been published on the use of some of the applications mentioned above. They include, amongst others, Brown (1986), Waller (1993) and Kano (1997). Structural Equation Models are increasingly being used in the social, economic and behavioral sciences. More and more people are therefore making use of one or more of the Structural Equation Modeling applications on the market. This study is performed with the aim of using each of the Structural Equation Modeling applications AMOS, EQS, LISREL, Mx, RAMONA and SEPATH for the first time and document the experience, joy and the difficulties encountered while using them. This treatise is different from the comparisons already published in that it is based on the use of AMOS, EQS, LISREL, Mx, RAMONA and SEPATH to fit a Structural Equation Model for peer influences on ambition, which is specified for data obtained by Duncan, Haller and Portes (1971), by myself as a first time user of each of the programs mentioned. The impressive features as well as the difficulties encountered are listed for each application. Recommendations for possible improvements to the various applications are also proposed. Finally, recommendations for future studies on the use of Structural Equation Modeling programs are made.

Uncertainty in structural dynamics is of growing concern to numerous industries. Significant attention has previously been devoted to the study of frequency response functions, however the uncertainty associated with excitation of structures by structureborne sound sources has received little attention. In this work, the uncertainty in the structure borne sound power transmitted from a vibration source into a receiving structure is considered. A method is presented whereby the uncertainties in the active and dynamic properties of a structure borne sound source and its receiver structure are propagated through to the injected power. Consideration is given to the case where the data describing the source and receiver is incomplete and is therefore termed ‘granular’. An approach for the estimation of the mean and uncertainty of granular variables is developed and it is shown that by estimating the mean and uncertainty of the missing elements the uncertainty propagation approach can be used for a ‘granular’ case. This approach is illustrated using an example in which the free velocity phase data is assumed to be unavailable. Idealised structure borne sound sources are created analytically in order to examine the validity of the presented methods. Good correlation is observed between the estimated uncertainties in the transmitted power and the uncertainties obtained through a Monte Carlo analysis. Insight into the frequency regions where large uncertainties can be expected in the transmitted structure borne sound power is obtained. It is argued that by providing estimates for the uncertainty of a prediction of the transmitted power, an insight into the reliability of the estimate is achieved, allowing engineering decisions to be made with greater confidence.

This thesis is concerned with developing techniques for analysing Raman spectroscopic images. A Raman spectroscopic image differs from a standard image as in place of red, green and blue quantities for each pixel a Raman image contains a spectrum of light intensities at each pixel. These spectra are used to identify the chemical components from which the image subject, for example a tablet, is comprised. The study of these types of images is known as chemometrics, with the majority of chemometric methods based on multivariate statistical and image analysis techniques. The work in this thesis has two main foci. The first of these is on the spectral decomposition of a Raman image, the purpose of which is to identify the component chemicals and their concentrations. The standard method for this is to fit a bilinear model to the image where both parts of the model, representing components and concentrations, must be estimated. As the standard bilinear model is nonidentifiable in its solutions we investigate the range of possible solutions in the solution space with a random walk. We also derive an improved model for spectral decomposition, combining cluster analysis techniques and the standard bilinear model. For this purpose we apply the expectation maximisation algorithm on a Gaussian mixture model with bilinear means, to represent our spectra and concentrations. This reduces noise in the estimated chemical components by separating the Raman image subject from the background. The second focus of this thesis is on the analysis of our spectral decomposition results. For testing the chemical components for uniform mixing we derive test statistics for identifying patterns in the image based on Minkowski measures, grey level co-occurence matrices and neighbouring pixel correlations. However with a non-identifiable model any hypothesis tests performed on the solutions will be specific to only that solution. Therefore to obtain conclusions for a range of solutions we combined our test statistics with our random walk. We also investigate the analysis of a time series of Raman images as the subject dissolved. Using models comprised of Gaussian cumulative distribution functions we are able to estimate the changes in concentration levels of dissolving tablets between the scan times. The results of which allowed us to describe the dissolution process in terms of the quantities of component chemicals.

M.Sc. / Please refer to full text to view abstract

No description available.

No description available.