Paraconsistent artificial neural networks and Alzheimer disease: a preliminary study

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Diagnostics

Background:

• Electroencephalogram (EEG) visual analysis aids Alzheimer's Disease (AD) diagnosis but suffers from imprecision.
• Factors like equipment variability and human interpretation introduce uncertainty in traditional EEG analysis.

Purpose of the Study:

• To utilize a Paraconsistent Artificial Neural Network (PANN) for determining the certainty of probable dementia diagnosis.
• To develop an auxiliary method for Alzheimer's Disease diagnosis using a specialized Artificial Neural Network (ANN).

Main Methods:

• EEG records from 10 probable AD patients and 10 controls were analyzed.
• A normal EEG background pattern was defined as 8-12 Hz with a 0.5 Hz variance.
• A PANN architecture was employed to recognize specific EEG patterns.

Main Results:

• The PANN accurately recognized Alpha band waves with favorable evidence (0.30) versus contrary evidence (0.19).
• For non-Alpha waves, favorable evidence averaged 0.19 and contrary evidence 0.32.
• The PANN demonstrated 80% efficiency in recognizing Alpha waves in the evaluated cases.

Conclusions:

• PANN shows potential as a tool for EEG analysis, offering quantitative insights.
• The network's ability to handle imprecise and inconsistent data is valuable for EEG interpretation.
• This approach addresses the need for more objective methods in AD diagnosis.