Bidirectional associative memories
IEEE Transactions on Systems, Man and Cybernetics
Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations
Design of a general purpose meta-assembler for parallel processor environment in ISPS
ANSS '89 Proceedings of the 22nd annual symposium on Simulation
| Hi-index | 0.00 |
Earlier research at the Navy Personnel Research and Development Center revealed that measures of the brain response to sensory stimuli, known as Event Related Potentials (ERP) may be used to assess unique process-related variance that is dependent upon human performance. For example, it was found that the sensitivity of individual subjects to dynamic color contrast in computer displays can be assessed by visual ERP's. It has also been observed that RMS measures of the P1-N1-P2 complex and the P300 component of the ERP are related to signal detection and classification measures of performance. The present effort is to classify the ERP response to stimuli followed by a motor action of the subject, using neural networks. The stimuli consist of flashing a light which could be either one of two distinct colors. The following motor-action corresponds to pushing one of the two available buttons representing the two colors. The features will be selected from the Fast Walsh Transformation (FWT) of ERP observations and will be applied to appropriate Neural Networks to obtain a prediction of a possible response. Since a small number of features is derived from a large section of available processed data and since the ERP data is diluted with noise and artifacts, there is a need to evaluate the effectiveness of this data compression towards efficient prediction of the future action of the subject. The present paper deals with this study.