Prediction of Cognitive Decline in Temporal Lobe Epilepsy and Mild Cognitive Impairment by EEG, MRI, and Neuropsychology

This study explores how well EEG, MRI, and standard cognitive tests can predict cognitive decline in people with temporal lobe epilepsy (TLE) or mild cognitive impairment (MCI). The researchers used machine learning to combine these different types of data to improve prediction accuracy. The study found that combining data from EEG, MRI, and neuropsychological tests could predict changes in cognitive performance, such as executive functions, visual-verbal memory, and divided attention. This suggests a more holistic approach to understanding and predicting cognitive decline. The findings indicate that there may be common biomarkers across different neurological conditions that contribute to cognitive decline. Identifying these biomarkers could lead to more general models for predicting cognitive decline.

• 1
  The best sensitivity/specificity for decline was 72%/82% for executive functions based on a feature combination from MRI volumetry and EEG partial coherence during recall of memories.
• 2
  For visual-verbal memory, a combination of MRI-wavelet features and neuropsychology achieved 95% sensitivity and 74% specificity in predicting decline.
• 3
  Combining EEG partial directed coherence at rest and neuropsychology predicted an increase in depression scores with 81% sensitivity and 90% specificity.