J. M. Górriz

18F-FDG PET imaging analysis for computer aided Alzheimer's diagnosis

Finding sensitive and appropriate technologies for non-invasive observation and early detection of Alzheimers disease (AD) is of fundamental importance to develop early treatments. In this work we develop a fully automatic computer aided diagnosis (CAD) system for high-dimensional pattern classification of baseline ^1^8F-FDG PET scans from Alzheimers disease neuroimaging initiative (ADNI) participants. Image projection as feature space dimension reduction technique is combined with an eigenimage based decomposition for feature extraction, and support vector machine (SVM) is used to manage the classification task. A two folded objective is achieved by reaching relevant classification performance complemented with an image analysis support for final decision making. A 88.24% accuracy in identifying mild AD, with 88.64% specificity, and 87.70% sensitivity is obtained. This method also allows the identification of characteristic AD patterns in mild cognitive impairment (MCI) subjects.

Feature selection using factor analysis for Alzheimer's diagnosis using 18F-FDG PET images.

PURPOSEnThis article presents a computer-aided diagnosis technique for improving the accuracy of the early diagnosis of Alzheimers disease (AD). Two hundred and tenF18-FDG PET images from the ADNI initiative [52 normal controls (NC), 114 mild cognitive impairment (MCI), and 53 AD subjects] are studied.nnnMETHODSnThe proposed methodology is based on the selection of voxels of interest using the t-test and a posterior reduction of the feature dimension using factor analysis. Factor loadings are used as features for three different classifiers: Two multivariate Gaussian mixture model, with linear and quadratic discriminant function, and a support vector machine with linear kernel.nnnRESULTSnAn accuracy rate up to 95% when NC and AD are considered and an accuracy rate up to 88% and 86% for NC-MCI and NC-MCI, AD, respectively, are obtained using SVM with linear kernel.nnnCONCLUSIONSnResults are compared to the voxel-as-features and a PCA- based approach and the proposed methodology achieves better classification performance.

Projecting independent components of SPECT images for computer aided diagnosis of Alzheimer's disease

Finding sensitive and appropriate technologies for early detection of the Alzheimers disease (AD) are of fundamental importance to develop early treatments. Single Photon Emission Computed Tomography (SPECT) images are non-invasive observation tools to assist the diagnosis, commonly processed through unsupervised statistical tests, or assessed visually. In this work, we present a computer aided diagnosis system based on supervised learning methods, exploring two different novel approaches. Independent Component Analysis (ICA) was used within this work to extract the relevant features from the image database and reduce the feature space dimensionality, to build a SVM with the resulting data. The proposed approach led to an error estimation below the 9%, and was able to detect the AD perfusion pattern and classify new subjects in an unsupervised manner.

Accurate human limb angle measurement: sensor fusion through Kalman, least mean squares and recursive least-squares adaptive filtering

Inertial sensors are widely used in human body motion monitoring systems since they permit us to determine the position of the subjects limbs. Limb angle measurement is carried out through the integration of the angular velocity measured by a rate sensor and the decomposition of the components of static gravity acceleration measured by an accelerometer. Different factors derived from the sensors nature, such as the angle random walk and dynamic bias, lead to erroneous measurements. Dynamic bias effects can be reduced through the use of adaptive filtering based on sensor fusion concepts. Most existing published works use a Kalman filtering sensor fusion approach. Our aim is to perform a comparative study among different adaptive filters. Several least mean squares (LMS), recursive least squares (RLS) and Kalman filtering variations are tested for the purpose of finding the best method leading to a more accurate and robust limb angle measurement. A new angle wander compensation sensor fusion approach based on LMS and RLS filters has been developed.

Parameterization of the distribution of white and grey matter in MRI using the α-stable distribution

This work presents a study of the distribution of the grey matter (GM) and white matter (WM) in brain magnetic resonance imaging (MRI). The distribution of GM and WM is characterized using a mixture of α-stable distributions. A Bayesian α-stable mixture model for histogram data is presented and unknown parameters are sampled using the Metropolis-Hastings algorithm. The proposed methodology is tested in 18 real images from the MRI brain segmentation repository. The GM and WM distributions are accurately estimated. The α-stable distribution mixture model presented in this paper can be used as previous step in more complex MRI segmentation procedures using spatial information. Furthermore, due to the fact that the α-stable distribution is a generalization of the Gaussian distribution, the proposed methodology can be applied instead of the Gaussian mixture model, which is widely used in segmentation of brain MRI in the literature.

Comparison between Different Intensity Normalization Methods in 123I-Ioflupane Imaging for the Automatic Detection of Parkinsonism

Intensity normalization is an important pre-processing step in the study and analysis of DaTSCAN SPECT imaging. As most automatic supervised image segmentation and classification methods base their assumptions regarding the intensity distributions on a standardized intensity range, intensity normalization takes on a very significant role. In this work, a comparison between different novel intensity normalization methods is presented. These proposed methodologies are based on Gaussian Mixture Model (GMM) image filtering and mean-squared error (MSE) optimization. The GMM-based image filtering method is achieved according to a probability threshold that removes the clusters whose likelihood are negligible in the non-specific regions. The MSE optimization method consists of a linear transformation that is obtained by minimizing the MSE in the non-specific region between the intensity normalized image and the template. The proposed intensity normalization methods are compared to: i) a standard approach based on the specific-to-non-specific binding ratio that is widely used, and ii) a linear approach based on the α-stable distribution. This comparison is performed on a DaTSCAN image database comprising analysis and classification stages for the development of a computer aided diagnosis (CAD) system for Parkinsonian syndrome (PS) detection. In addition, these proposed methods correct spatially varying artifacts that modulate the intensity of the images. Finally, using the leave-one-out cross-validation technique over these two approaches, the system achieves results up to a 92.91% of accuracy, 94.64% of sensitivity and 92.65 % of specificity, outperforming previous approaches based on a standard and a linear approach, which are used as a reference. The use of advanced intensity normalization techniques, such as the GMM-based image filtering and the MSE optimization improves the diagnosis of PS.

Functional brain image classification using association rules defined over discriminant regions

This letter shows a novel computer aided diagnosis (CAD) system for the early diagnosis of Alzheimers Disease (AD). The proposed method evaluates the reliability of association rules (ARs) aiming to discover interesting associations between attributes in functional brain imaging, i.e. single photon emission computed tomography (SPECT) and positron emission tomography (PET). AR mining firstly requires a masking process for reducing the computational cost, which is based on Fisher discriminant ratio (FDR), in order to identify transactions or relationships among discriminant brain areas. Once the activation map is achieved by means of activation estimation (AE), the resulting regions of interest (ROIs) are subjected to AR discovery with a specified minimum support and confidence. Finally, the proposed CAD system performs image classification by evaluating the number of previously mined rules from controls that are verified by each subject. Several experiments were carried out on two different image modalities (SPECT and PET) in order to highlight the generalization ability of the proposed method. The AR-based method yields an accuracy up to 92.78% (with 87.5% sensitivity and 100% specificity) and 91.33% (with 82.67% sensitivity and 100% specificity) for SPECT and PET, respectively, thus outperforming recently developed methods for early diagnosis of AD.

Voice activity detection using higher order statistics

A robust and effective voice activity detection (VAD) algorithm is proposed for improving speech recognition performance in noisy environments. The approach is based on filtering the input channel to avoid high energy noisy components and then the determination of the speech/non-speech bispectra by means of third order auto-cumulants. This algorithm differs from many others in the way the decision rule is formulated (detection tests) and the domain used in this approach. Clear improvements in speech/non-speech discrimination accuracy demonstrate the effectiveness of the proposed VAD. It is shown that application of statistical detection test leads to a better separation of the speech and noise distributions, thus allowing a more effective discrimination and a tradeoff between complexity and performance. The algorithm also incorporates a previous noise reduction block improving the accuracy in detecting speech and non-speech.

Independent Component Analysis-Based Classification of Alzheimer’s Disease from Segmented MRI Data

An accurate and early diagnosis of the Alzheimer’s disease (AD) is of fundamental importance to improve diagnosis techniques, to better understand this neurodegenerative process and to develop effective treatments. In this work, a novel classification method based on independent component analysis (ICA) and supervised learning methods is proposed to be applied on segmented brain magnetic resonance imaging (MRI) from Alzheimer’s disease neuroimaging initiative (ADNI) participants for automatic classification task. The ICA-based method is composed of three step. First, MRI are normalized and segmented by the Statistical Parametric Mapping (SPM8) software. After that, average image of normal (NC), mild cognitive impairment (MCI) or AD subjects are computed. Then, FastICA is applied to these different average images for extracting a set of independent components (IC) which symbolized each class characteristics. Finally, each brain image from the database was projected onto the space spanned by this independent components basis for feature extraction, a support vector machine (SVM) is used to manage the classification task. A 87.5% accuracy in identifying AD from NC, with 90.4% specificity and 84.6% sensitivity is obtained. According to the experimental results, we can see that this proposed method can successfully differentiate AD, MCI and NC subjects. So, it is suitable for automatic classification of sMRI images.

Optimizing blind source separation with guided genetic algorithms

This paper proposes a novel method for blindly separating unobservable independent component (IC) signals based on the use of a genetic algorithm. It is intended for its application to the problem of blind source separation (BSS) on post-nonlinear mixtures. The paper also includes a formal proof on the convergence of the proposed algorithm using guiding operators, a new concept in the GA scenario. This approach is very useful in many fields such as forecasting indexes in financial stock markets, where the search for independent components is the major task to include exogenous information into the learning machine; or biomedical applications which usually use a high number of input signals. The guiding GA (GGA) presented in this work, is able to extract IC with faster rate than the previous ICA algorithms, as input space dimension increases. It shows significant accuracy and robustness than the previous approaches in any case. In addition, we present a simple though effective contrast function which evaluates individuals of each population (candidate solutions) based (a) on estimating the probability densities of the outputs through histogram approximation and (b) evaluating higher-order statistics of the outputs.