Numerical recipes in C (2nd ed.): the art of scientific computing
Numerical recipes in C (2nd ed.): the art of scientific computing
A combined neural network and decision trees model for prognosis of breast cancer relapse
Artificial Intelligence in Medicine
| Hi-index | 0.00 |
Event-Related Potentials (ERPs) provide non invasive measurements of the electrical activity on the scalp that are linked to the presentation of stimuli and events. Brain mapping techniques are able to provide evidence on the solution of debatable issues in cognitive science. In this paper, an effective signal classification approach is proposed, extending the use of two inversion techniques: the Brain Electrical Tomography using Algebraic Reconstruction Technique (BET-ART) and the Low Resolution Brain Electromagnetic Tomography (LORETA). The first step of the methodology applied is the feature extraction, which is based on the combination of the Multivariate Autoregressive model with the Simulated Annealing technique, in order to extract optimum features, in terms of classification rate. The classification, as the second step of the methodology, is implemented by means of an Artificial Neural Network (ANN) trained with the back-propagation algorithm under the "leave-one-out cross-validation" scenario. The ANN is a multi-layer perceptron, the architecture of which is selected after a detailed search. The proposed methodology has been applied for the classification of First Episode Schizophrenic (FES) patients and normal controls using the intracranial activity distributions obtained by ERPs. A comparative analysis was performed using BET-ART and LORETA inversion methods. Implementation of the proposed methodology provided classification rates of up to 93.1%, for both types of input signals. Additionally, for both BET-ART and LORETA signals, the brain regions that differentiate FES patients from normal controls are located in the frontal brain area, in accordance to the related literature. The proposed methodology may be used for the design of more robust classifiers based on intracranial source distributions, which are more closely related to the underlying cognitive mechanisms responsible for the generation of the scalp-recorded biosignals.