Related Experiment Videos
Neural network approach in multichannel auditory event-related potential analysis
F Y Wu1, J D Slater, R E Ramsay
1Department of Electrical and Computer Engineering, University of Miami, Coral Gables, FL.
International Journal of Bio-Medical Computing
|April 1, 1994
Summary
Artificial neural networks (ANNs) can classify P300 event-related potentials (ERPs) in multiple sclerosis (MS) patients. This study enhances ANN methods for automated P300 assessment, standardizing criteria and aiding computer-aided diagnosis.
Area of Science:
- Neuroscience
- Medical Imaging
- Artificial Intelligence
Background:
- P300 event-related potentials (ERPs) lack defined abnormality criteria for neuropsychological impairment.
- Statistical analysis suggests criteria exist for differentiating control subjects from patients with conditions like multiple sclerosis (MS).
- Previous studies showed artificial neural network (ANN) feasibility in single-channel ERP classification for MS patients.
Purpose of the Study:
- To report multichannel P300 ERP analysis results.
- To introduce a modified ANN methodology for automated classification rule extraction.
- To enhance the standardization and computer-aided analysis of P300 ERPs in neuropsychological assessments.
Main Methods:
- Multichannel P300 event-related potential (ERP) data acquisition from control subjects and MS patients.
- Application of artificial neural network (ANN) analysis to ERP waveforms.
- Development of a modified ANN methodology for automated classification rule extraction.
Main Results:
- Demonstrated feasibility of ANN methods for classifying ERP waveforms.
- Successfully applied multichannel ERP analysis.
- Introduced a modified ANN methodology that significantly reduces statistical analysis workload.
- Enhanced automation of classification rule extraction for P300 ERPs.
Conclusions:
- The proposed methodology standardizes P300 ERP assessment criteria.
- Facilitates computer-aided analysis of neuropsychological functions.
- Offers a more automated and efficient approach to classifying neurological impairment using ERPs.