Normalized Power Variance: A new Field Orthogonal to Power in EEG Analysis

Area of Science:

• Neuroscience
• Biophysics
• Mathematical Neuroscience

Background:

• Electroencephalogram (EEG) is crucial for diagnosing neurological conditions like epilepsy and dementia.
• Traditional EEG analysis focuses on wave inspection and power analysis, often using methods like exact low-resolution brain electromagnetic tomography (eLORETA).
• Information geometry introduces a new dimension to EEG analysis: normalized power variance (NPV), orthogonal to power.

Purpose of the Study:

• To investigate the utility of NPV analysis combined with eLORETA in idiopathic normal pressure hydrocephalus (iNPH) patients.
• To determine if NPV analysis can detect differences related to cerebrospinal fluid (CSF) shunt operation outcomes where traditional methods fail.
• To explore NPV as a sensitive indicator of cortical impairment.

Main Methods:

• Applied normalized power variance (NPV) analysis using eLORETA to EEG data from iNPH patients.
• Compared NPV values between patients who responded to CSF shunt surgery (n=17) and those who did not (n=19).
• Focused on the beta frequency band and high convexity areas of the brain.

Main Results:

• NPV analysis of eLORETA revealed significantly higher NPV values in the high convexity area (beta frequency band) in shunt responders compared to non-responders.
• Traditional power analysis did not detect significant differences between the groups regarding shunt operation outcome.
• The findings suggest NPV is a sensitive metric for detecting subtle cortical changes.

Conclusions:

• EEG possesses a previously unrecognized dimension, normalized power variance (NPV), offering valuable information about cortical electrical activity.
• NPV analysis, particularly when combined with eLORETA, shows promise as a sensitive biomarker for predicting clinical outcomes in iNPH.
• This novel approach has the potential to serve as an early warning signal for cortical impairment in various neurological conditions.