Spatio-Temporal EEG Models for Brain Interfaces.

Multichannel electroencephalography (EEG) is widely used in non-invasive brain computer interfaces (BCIs) for user intent inference. EEG can be assumed to be a Gaussian process with unknown mean and autocovariance, and the estimation of parameters is required for BCI inference. However, the relatively high dimensionality of the EEG feature vectors with respect to the number of labeled observations lead to rank deficient covariance matrix estimates. In this manuscript, to overcome ill-conditioned covariance estimation, we propose a structure for the covariance matrices of the multichannel EEG signals. Specifically, we assume that these covariances can be modeled as a Kronecker product of temporal and spatial covariances. Our results over the experimental data collected from the users of a letter-by-letter typing BCI show that with less number of parameter estimations, the system can achieve higher classification accuracies compared to a method that uses full unstructured covariance estimation. Moreover, in order to illustrate that the proposed Kronecker product structure could enable shortening the BCI calibration data collection sessions, using Cramer-Rao bound analysis on simulated data, we demonstrate that a model with structured covariance matrices will achieve the same estimation error as a model with no covariance structure using fewer labeled EEG observations.

Assessing subsets of analytes in context of detecting laboratory errors.

Laboratory error detection is a hard task yet plays an important role in efficient care of the patients. Quality controls are inadequate in detecting pre-analytic errors and are not frequent enough. Hence population- and patient-based detectors are developed. However, it is not clear what set of analytes leads to the most efficient error detectors. Here, we use three different scoring functions that can be used in detecting errors, to rank a set of analytes in terms of their strength in distinguishing erroneous measurements. We also observe that using evaluations of larger subsets of analytes in our analysis does not necessarily lead to a more accurate error detector. In our data set obtained from renal kidney disease inpatients, calcium, potassium, and sodium, emerged as the top-3 indicators of an erroneous measurement. Using the joint likelihood of these three analytes, we obtain an estimated AUC of 0.73 in error detection.

Dealing with inter-expert variability in retinopathy of prematurity: A machine learning approach.

BACKGROUND AND OBJECTIVE: Understanding the causes of disagreement among experts in clinical decision making has been a challenge for decades. In particular, a high amount of variability exists in diagnosis of retinopathy of prematurity (ROP), which is a disease affecting low birth weight infants and a major cause of childhood blindness. A possible cause of variability, that has been mostly neglected in the literature, is related to discrepancies in the sets of important features considered by different experts. In this paper we propose a methodology which makes use of machine learning techniques to understand the underlying causes of inter-expert variability. METHODS: The experiments are carried out on a dataset consisting of 34 retinal images, each with diagnoses provided by 22 independent experts. Feature selection techniques are applied to discover the most important features considered by a given expert. Those features selected by each expert are then compared to the features selected by other experts by applying similarity measures. Finally, an automated diagnosis system is built in order to check if this approach can be helpful in solving the problem of understanding high inter-rater variability. RESULTS: The experimental results reveal that some features are mostly selected by the feature selection methods regardless the considered expert. Moreover, for pairs of experts with high percentage agreement among them, the feature selection algorithms also select similar features. By using the relevant selected features, the classification performance of the automatic system was improved or maintained. CONCLUSIONS: The proposed methodology provides a handy framework to identify important features for experts and check whether the selected features reflect the pairwise agreements/disagreements. These findings may lead to improved diagnostic accuracy and standardization among clinicians, and pave the way for the application of this methodology to other problems which present inter-expert variability.

Analysis of underlying causes of inter-expert disagreement in retinopathy of prematurity diagnosis. Application of machine learning principles.

OBJECTIVE: Inter-expert variability in image-based clinical diagnosis has been demonstrated in many diseases including retinopathy of prematurity (ROP), which is a disease affecting low birth weight infants and is a major cause of childhood blindness. In order to better understand the underlying causes of variability among experts, we propose a method to quantify the variability of expert decisions and analyze the relationship between expert diagnoses and features computed from the images. Identification of these features is relevant for development of computer-based decision support systems and educational systems in ROP, and these methods may be applicable to other diseases where inter-expert variability is observed. METHODS: The experiments were carried out on a dataset of 34 retinal images, each with diagnoses provided independently by 22 experts. Analysis was performed using concepts of Mutual Information (MI) and Kernel Density Estimation. A large set of structural features (a total of 66) were extracted from retinal images. Feature selection was utilized to identify the most important features that correlated to actual clinical decisions by the 22 study experts. The best three features for each observer were selected by an exhaustive search on all possible feature subsets and considering joint MI as a relevance criterion. We also compared our results with the results of Cohen's Kappa [36] as an inter-rater reliability measure. RESULTS: The results demonstrate that a group of observers (17 among 22) decide consistently with each other. Mean and second central moment of arteriolar tortuosity is among the reasons of disagreement between this group and the rest of the observers, meaning that the group of experts consider amount of tortuosity as well as the variation of tortuosity in the image. CONCLUSION: Given a set of image-based features, the proposed analysis method can identify critical image-based features that lead to expert agreement and disagreement in diagnosis of ROP. Although tree-based features and various statistics such as central moment are not popular in the literature, our results suggest that they are important for diagnosis.

Improved accuracy using recursive bayesian estimation based language model fusion in ERP-based BCI typing systems.

RSVP Keyboard™ is an electroencephalography (EEG) based brain computer interface (BCI) typing system, designed as an assistive technology for the communication needs of people with locked-in syndrome (LIS). It relies on rapid serial visual presentation (RSVP) and does not require precise eye gaze control. Existing BCI typing systems which uses event related potentials (ERP) in EEG suffer from low accuracy due to low signal-to-noise ratio. Henceforth, RSVP Keyboard™ utilizes a context based decision making via incorporating a language model, to improve the accuracy of letter decisions. To further improve the contributions of the language model, we propose recursive bayesian estimation, which relies on non-committing string decisions, and conduct an offline analysis, which compares it with the existing naïve bayesian fusion approach. The results indicate the superiority of the recursive bayesian fusion and in the next generation of RSVP Keyboard™ we plan to incorporate this new approach.

OBSERVER AND FEATURE ANALYSIS ON DIAGNOSIS OF RETINOPATHY OF PREMATURITY.

Retinopathy of prematurity (ROP) is a disease affecting low-birth weight infants and is a major cause of childhood blindness. However, human diagnoses is often subjective and qualitative. We propose a method to analyze the variability of expert decisions and the relationship between the expert diagnoses and features. The analysis is based on Mutual Information and Kernel Density Estimation on features. The experiments are carried out on a dataset of 34 retinal images diagnosed by 22 experts. The results show that a group of observers decide consistently with each other and there are popular features that have a high correlation with labels.

Optimizing the P300-based brain-computer interface: current status, limitations and future directions.

This paper summarizes the presentations and discussions at a workshop held during the Fourth International BCI Meeting charged with reviewing and evaluating the current state, limitations and future development of P300-based brain-computer interface (P300-BCI) systems. We reviewed such issues as potential users, recording methods, stimulus presentation paradigms, feature extraction and classification algorithms, and applications. A summary of the discussions and the panel's recommendations for each of these aspects are presented.

On Visually Evoked Potentials in EEG Induced by Multiple Pseudorandom Binary Sequences for Brain Computer Interface Design.

Visually evoked potentials have attracted great attention in the last two decades for the purpose of brain computer interface design. Visually evoked P300 response is a major signal of interest that has been widely studied. Steady state visual evoked potentials that occur in response to periodically flickering visual stimuli have been primarily investigated as an alternative. There also exists some work on the use of an m-sequence and its shifted versions to induce responses that are primarily in the visual cortex but are not periodic. In this paper, we study the use of multiple m-sequences for intent discrimination in the brain interface, as opposed to a single m-sequence whose shifted versions are to be discriminated from each other. Specifically we used four different m-sequences of length 31. Our main goal is to study if the bit presentation rate of the m-sequences have an impact on classification accuracy and speed. In this initial study, where we compared two basic classifier schemes using EEG data acquired with 15Hz and 30Hz bit presentation rates, our results are mixed; while on one subject, we got promising results indicating bit presentation rate could be increased without decrease in classification accuracy; thus leading to a faster decision-rate in the brain interface, on our second subject, this conclusion is not supported. Further detailed experimental studies as well as signal processing methodology design, especially for information fusion across EEG channels, will be conducted to investigate this question further.

Speech enhancement in discontinuous transmission systems using the constrained-stability least-mean-squares algorithm.

In this paper a novel constrained-stability least-mean-squares (LMS) algorithm for filtering speech sounds is proposed in the adaptive noise cancellation (ANC) problem. It is based on the minimization of the squared Euclidean norm of the weight vector change under a stability constraint over the a posteriori estimation errors. To this purpose, the Lagrangian methodology has been used in order to propose a nonlinear adaptation in terms of the product of differential input and error. Convergence analysis is also studied in terms of the evolution of the natural modes to the optimal Wiener-Hopf solution so that the stability performance depends exclusively on the adaptation parameter mu and the eigenvalues of the difference matrix DeltaR(1). The algorithm shows superior performance over the referenced algorithms in the ANC problem of speech discontinuous transmission systems, which are characterized by rapid transitions of the desired signal. The experimental analysis carried out on the AURORA 3 speech databases provides an extensive performance evaluation together with an exhaustive comparison to the standard LMS algorithms, i.e., the normalized LMS (NLMS), and other recently reported LMS algorithms such as the modified NLMS, the error nonlinearity LMS, or the normalized data nonlinearity LMS adaptation.

A comparison of optimal MIMO linear and nonlinear models for brain-machine interfaces.

The field of brain-machine interfaces requires the estimation of a mapping from spike trains collected in motor cortex areas to the hand kinematics of the behaving animal. This paper presents a systematic investigation of several linear (Wiener filter, LMS adaptive filters, gamma filter, subspace Wiener filters) and nonlinear models (time-delay neural network and local linear switching models) applied to datasets from two experiments in monkeys performing motor tasks (reaching for food and target hitting). Ensembles of 100-200 cortical neurons were simultaneously recorded in these experiments, and even larger neuronal samples are anticipated in the future. Due to the large size of the models (thousands of parameters), the major issue studied was the generalization performance. Every parameter of the models (not only the weights) was selected optimally using signal processing and machine learning techniques. The models were also compared statistically with respect to the Wiener filter as the baseline. Each of the optimization procedures produced improvements over that baseline for either one of the two datasets or both.

Generalized information potential criterion for adaptive system training.

We have previously proposed the quadratic Renyi's error entropy as an alternative cost function for supervised adaptive system training. An entropy criterion instructs the minimization of the average information content of the error signal rather than merely trying to minimize its energy. In this paper, we propose a generalization of the error entropy criterion that enables the use of any order of Renyi's entropy and any suitable kernel function in density estimation. It is shown that the proposed entropy estimator preserves the global minimum of actual entropy. The equivalence between global optimization by convolution smoothing and the convolution by the kernel in Parzen windowing is also discussed. Simulation results are presented for time-series prediction and classification where experimental demonstration of all the theoretical concepts is presented.