Kernel-Based Relevance Analysis with Enhanced Interpretability for Detection of Brain Activity Patterns (original) (raw)
Related papers
Stepwise Feature Selection by Cross Validation for EEG-based Brain Computer Interface
The 2006 IEEE International Joint Conference on Neural Network Proceedings, 2006
The potential of brain-computer interfaces (BCI) in serving a useful purpose, e.g., supporting communication in paralyzed patients, hinges on the quality of the classification of the brain waves. This paper proposes a novel method to construct a classifier with improved generalization performance. A feature selection method is applied to features calculated from the EEG signals so that unnecessary or redundant features can be removed and only effective features are left for the classification task. Kernel support vector machines (kernel SVM) were used as a classifier and the best combinations of features were searched by backward stepwise selection, i.e., by eliminating unnecessary features one by one, and by evaluating the resulting generalization performance through cross validation. Experiments showed that the generalization performance of the classifier constructed from the best set of features was higher than that of the classifier using all features.
Automatic EEG signal classification for epilepsy diagnosis with Relevance Vector Machines
In this paper, we investigate the potentials of applying a kernel-based learning machine, the Relevance Vector Machine (RVM), to the task of epilepsy detection through automatic electroencephalogram (EEG) signal classification. For this purpose, some experiments have been conducted over publicly available data, contrasting the performance levels exhibited by RVM models with those achieved with Support Vector Machines (SVMs), both in terms of predictive accuracy and sensitivity to the choice of the kernel function. Four settings of both types of kernel machine were considered in this study, which vary in accord with the type of input data they receive, either raw EEG signal or some statistical features extracted from the wavelet-transformed data. The empirical results indicate that: (1) in terms of accuracy, the best-calibrated RVM models have shown very satisfactory performance levels, which are rather comparable to those of SVMs; (2) an increase of accuracy is sometimes accompanied by loss of sparseness in the resulting RVM models; (3) both types of machines present similar sensitivity profiles to the kernel functions considered, having some kernel parameter values clearly associated with better accuracy rate; (4) when not making use of a feature extraction technique, the choice of the kernel function seems to be very relevant for significantly leveraging the performance of RVMs; and (5) when making use of derived features, the choice of the feature extraction technique seems to be an important factor to one take into account.
BioMedical Engineering OnLine, 2014
Background: The extraction of physiological rhythms from electroencephalography (EEG) data and their automated analyses are extensively studied in clinical monitoring, to find traces of interictal/ictal states of epilepsy. Methods: Because brain wave rhythms in normal and interictal/ictal events, differently influence neuronal activity, our proposed methodology measures the contribution of each rhythm. These contributions are measured in terms of their stochastic variability and are extracted from a Short Time Fourier Transform to highlight the non-stationary behavior of the EEG data. Then, we performed a variability-based relevance analysis by handling the multivariate short-time rhythm representation within a subspace framework. This maximizes the usability of the input information and preserves only the data that contribute to the brain activity classification. For neural activity monitoring, we also developed a new relevance rhythm diagram that qualitatively evaluates the rhythm variability throughout long time periods in order to distinguish events with different neuronal activities. Results: Evaluations were carried out over two EEG datasets, one of which was recorded in a noise-filled environment. The method was evaluated for three different classification problems, each of which addressed a different interpretation of a medical problem. We perform a blinded study of 40 patients using the support-vector machine classifier cross-validation scheme. The obtained results show that the developed relevance analysis was capable of accurately differentiating normal, ictal and interictal activities. Conclusions: The proposed approach provides the reliable identification of traces of interictal/ictal states of epilepsy. The introduced relevance rhythm diagrams of physiological rhythms provides effective means of monitoring epileptic seizures; additionally, these diagrams are easily implemented and provide simple clinical interpretation. The developed variability-based relevance analysis can be translated to other monitoring applications involving time-variant biomedical data.
Kernel regression for fMRI pattern prediction
NeuroImage, 2011
This paper introduces two kernel-based regression schemes to decode or predict brain states from functional brain scans as part of the Pittsburgh Brain Activity Interpretation Competition (PBAIC) 2007, in which our team was awarded first place. Our procedure involved image realignment, spatial smoothing, detrending of low-frequency drifts, and application of multivariate linear and non-linear kernel regression methods: namely kernel ridge regression (KRR) and relevance vector regression (RVR). RVR is based on a Bayesian framework, which automatically determines a sparse solution through maximization of marginal likelihood. KRR is the dual-form formulation of ridge regression, which solves regression problems with high dimensional data in a computationally efficient way. Feature selection based on prior knowledge about human brain function was also used. Post-processing by constrained deconvolution and re-convolution was used to furnish the prediction. This paper also contains a detailed description of how prior knowledge was used to fine tune predictions of specific "feature ratings," which we believe is one of the key factors in our prediction accuracy. The impact of pre-processing was also evaluated, demonstrating that different pre-processing may lead to significantly different accuracies. Although the original work was aimed at the PBAIC, many techniques described in this paper can be generally applied to any fMRI decoding works to increase the prediction accuracy.
Novel Optimization Models for Abnormal Brain Activity Classification
Operations Research, 2008
This paper proposes a new classification technique, called support feature machine (SFM), for multidimensional time-series data. The proposed technique was applied to the classification of abnormal brain activity represented in electroencephalograms (EEGs). First, the dynamical properties of EEGs from each electrode were extracted. These dynamical profiles were put in SFM, which is an optimization model that maximizes classification accuracy by selecting electrodes (features) that correctly classify unlabeled EEG samples based on the nearest-neighbor classification rule. The empirical studies were performed on the EEG data sets collected from 10 subjects. The performance of SFM was assessed and compared with the ones achieved by the traditional k-nearest-neighbor classifier and support vector machines (SVMs). The results show that SFM achieved, on average, over 90% correct classification and outperformed other classification techniques. In the validation step, SFM correctly classified unseen preseizure and normal EEGs with over 73% accuracy.
Classification of EEG signals using a multiple kernel learning support vector machine
Sensors (Basel, Switzerland), 2014
In this study, a multiple kernel learning support vector machine algorithm is proposed for the identification of EEG signals including mental and cognitive tasks, which is a key component in EEG-based brain computer interface (BCI) systems. The presented BCI approach included three stages: (1) a pre-processing step was performed to improve the general signal quality of the EEG; (2) the features were chosen, including wavelet packet entropy and Granger causality, respectively; (3) a multiple kernel learning support vector machine (MKL-SVM) based on a gradient descent optimization algorithm was investigated to classify EEG signals, in which the kernel was defined as a linear combination of polynomial kernels and radial basis function kernels. Experimental results showed that the proposed method provided better classification performance compared with the SVM based on a single kernel. For mental tasks, the average accuracies for 2-class, 3-class, 4-class, and 5-class classifications we...
Kernel PCA feature extraction of event-related potentials for human signal detection performance
Artificial neural networks in …, 2000
In this paper, we propose the application of the Kernel PCA technique for feature selection in high-dimensional feature space where input variables are mapped by a Gaussian kernel. The extracted features are employed in the regression problem of estimating human signal detection performance from brain event-related potentials elicited by task relevant signals. We report the superiority of Kernel PCA for feature extraction over linear PCA.
Multiple kernel learning for brain-computer interfacing
Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, 2013
Combining information from different sources is a common way to improve classification accuracy in Brain-Computer Interfacing (BCI). For instance, in small sample settings it is useful to integrate data from other subjects or sessions in order to improve the estimation quality of the spatial filters or the classifier. Since data from different subjects may show large variability, it is crucial to weight the contributions according to importance. Many multi-subject learning algorithms determine the optimal weighting in a separate step by using heuristics, however, without ensuring that the selected weights are optimal with respect to classification. In this work we apply Multiple Kernel Learning (MKL) to this problem. MKL has been widely used for feature fusion in computer vision and allows to simultaneously learn the classifier and the optimal weighting. We compare the MKL method to two baseline approaches and investigate the reasons for performance improvement.
arXiv (Cornell University), 2017
Alzheimer's disease is a major cause of dementia. Its diagnosis requires accurate biomarkers that are sensitive to disease stages. In this respect, we regard probabilistic classification as a method of designing a probabilistic biomarker for disease staging. Probabilistic biomarkers naturally support the interpretation of decisions and evaluation of uncertainty associated with them. In this paper, we obtain probabilistic biomarkers via Gaussian Processes. Gaussian Processes enable probabilistic kernel machines that offer flexible means to accomplish Multiple Kernel Learning. Exploiting this flexibility, we propose a new variation of Automatic Relevance Determination and tackle the challenges of high dimensionality through multiple kernels. Our research results demonstrate that the Gaussian Process models are competitive with or better than the well-known Support Vector Machine in terms of classification performance even in the cases of single kernel learning. Extending the basic scheme towards the Multiple Kernel Learning, we improve the efficacy of the Gaussian Process models and their interpretability in terms of the known anatomical correlates of the disease. For instance, the disease pathology starts in and around the hippocampus and entorhinal cortex. Through the use of Gaussian Processes and Multiple Kernel Learning, we have automatically and efficiently determined those portions of neuroimaging data. In addition to their interpretability, our Gaussian Process models are competitive with recent deep learning solutions under similar settings.