Clinician's road map to wavelet EEG as an Alzheimer's disease biomarker.

Alzheimer's disease (AD) is considered the main cause of dementia in Western countries. Consequently, there is a need for an accurate, universal, specific and cost-effective biomarker for early AD diagnosis, to follow disease progression and therapy response. This article describes a new diagnostic approach to quantitative electroencephalogram (QEEG) diagnosis of mild and moderate AD. The data set used in this study was composed of EEG signals recorded from 2 groups: (S1) 74 normal subjects, 33 females and 41 males (mean age 67 years, standard deviation = 8) and (S2) 88 probable AD patients (NINCDS-ADRDA criteria), 55 females and 33 males (mean age 74.7 years, standard deviation = 7.8) with mild to moderate symptoms (DSM-IV-TR). Attention is given to sample size and the use of state of the art open source tools (LetsWave and WEKA) to process the EEG data. This innovative technique consists in associating Morlet wavelet filter with a support vector machine technique. A total of 111 EEG features (attributes) were obtained for 162 probands. The results were accuracy of 92.72% and area under the curve of 0.92 (percentage split test). Most important, comparing a single patient versus the total data set resulted in accuracy of 84.56% (leave-one-patient-out test). Particular emphasis was on clinical diagnosis and feasibility of implementation of this low-cost procedure, because programming knowledge is not required. Consequently, this new method can be useful to support AD diagnosis in resource-limited settings.

Improving Alzheimer's disease diagnosis with machine learning techniques.

There is not a specific test to diagnose Alzheimer's disease (AD). Its diagnosis should be based upon clinical history, neuropsychological and laboratory tests, neuroimaging and electroencephalography (EEG). Therefore, new approaches are necessary to enable earlier and more accurate diagnosis and to follow treatment results. In this study we used a Machine Learning (ML) technique, named Support Vector Machine (SVM), to search patterns in EEG epochs to differentiate AD patients from controls. As a result, we developed a quantitative EEG (qEEG) processing method for automatic differentiation of patients with AD from normal individuals, as a complement to the diagnosis of probable dementia. We studied EEGs from 19 normal subjects (14 females/5 males, mean age 71.6 years) and 16 probable mild to moderate symptoms AD patients (14 females/2 males, mean age 73.4 years. The results obtained from analysis of EEG epochs were accuracy 79.9% and sensitivity 83.2%. The analysis considering the diagnosis of each individual patient reached 87.0% accuracy and 91.7% sensitivity.

EEG spectro-temporal modulation energy: a new feature for automated diagnosis of Alzheimer's disease.

There is recent indication that Alzheimer's disease (AD) can be characterized by atypical modulation of electrophysiological brain activity caused by fibrillar amyloid deposition in specific regions of the brain, such as those related to cognition and memory. In this paper, we propose to objectively characterize EEG sub-band modulation in an attempt to develop an automated noninvasive AD diagnostics tool. First, multi-channel full-band EEG signals are decomposed into five well-known frequency sub-bands: delta, theta, alpha, beta, and gamma. The temporal amplitude envelope of each sub-band is then computed via a Hilbert transformation. The proposed 'spectro-temporal modulation energy' feature measures the rate with which each sub-band is modulated. Modulation energy features are computed for 19 referential EEG signals and seven bipolar signals. Salient features are then selected and used to train four different classifiers, namely, support vector machines, logistic regression, classification and regression trees, and neural networks. Experiments with a database of 34 participants, 22 of which have been clinically diagnosed with probable-AD, show a neural network classifier achieving over 91% accuracy, thus significantly outperforming a classifier trained with conventional spectral-based features.

Evaluation of noise reduction techniques in the splice junction recognition problem

The Human Genome Project has generated a large amount of sequence data. A number of works are currently concerned with analyzing these data. One of the analyses carried out is the identification of genes' structures on the junctions represent a type of signal present on eukariot genes. Many studies have appied Machine Learning techniques in the recognition of such regions. However, most of the genetic databases are characterized y the presence of noise data, which can affect the performance of the learning techniques. This paper evaluates the effectiveness of five data pre-processing algorithms in the elimination of noisy instances from two splice junction recognition datasets. After the pre-processing phase, two learning techniques, Decision Trees and Support Vector Machines, are employed in the recognition process.