Use of brain electrophysiological quantitative data to classify and subtype an individual into diagnostic categories by discriminant and cluster analysis
First Claim
Patent Images
1. A quantitative process of using brain electro-physiological data (BE) to classify an individual in one of more than two diagnostic categories comprising the machine-implemented steps of:
- providing BE data for a selected individual;
processing said BE data by applying thereto selected spectral analysis and statistical procedures to extract a set of desired features of said BE data;
deriving a respective discriminant score for each of a set of more than two diagnostic categories by selectively weighting and combining a number of selected ones of said features of said BE data;
combining selected discriminant scores to derive respective probabilities that the individual belongs to selected diagnostic categories;
applying selected guardbands or rule-out levels to said probabilities to enhance the reliability of classifying said individual into a diagnostic category on the basis of said probabilities; and
classifying the individual into one of more than two diagnostic categories on the basis of said probabilities and said guardbands.
1 Assignment
0 Petitions
Accused Products
Abstract
A system for using discriminant analysis of EEG data to automatically evaluate the probability that an individual patient belongs to specified diagnostic categories or a subtype within a category where there are more than two categories or subtypes, and the system automatically places the patient into one of those more than two categories or subtypes.
-
Citations
21 Claims
-
1. A quantitative process of using brain electro-physiological data (BE) to classify an individual in one of more than two diagnostic categories comprising the machine-implemented steps of:
-
providing BE data for a selected individual; processing said BE data by applying thereto selected spectral analysis and statistical procedures to extract a set of desired features of said BE data; deriving a respective discriminant score for each of a set of more than two diagnostic categories by selectively weighting and combining a number of selected ones of said features of said BE data; combining selected discriminant scores to derive respective probabilities that the individual belongs to selected diagnostic categories; applying selected guardbands or rule-out levels to said probabilities to enhance the reliability of classifying said individual into a diagnostic category on the basis of said probabilities; and classifying the individual into one of more than two diagnostic categories on the basis of said probabilities and said guardbands. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A quantitative EEG method comprising:
-
deriving quantitative EEG data for an individual and subjecting said EEG data to preprocessing including artifact rejection to reduce said EEG data from artifact-free segments, subjecting said artifact-free segments to spectral processing to extract desired features thereof, applying transforms to said features as needed to ensure Gaussianity and Z-transforming said features on the basis of normative age regression data for a population of individuals assumed to be normal to thereby derive Z-scores for selected features and any desired combinations of features; and deriving discriminant scores for selected diagnostic categories from Z-scores derived from an individual'"'"'s EEG data by combining for each respective discriminant function selected ones of the individual'"'"'s Z-scores weighted by selected coefficients specific to the respective diagnostic category, combining selected discriminant scores to evaluate the probability that the individual belongs to a particular one of two or more selected diagnostic categories and applying guardbands to the evaluated probabilities to ascertain the level of confidence that the probability correctly classifies the individual into a diagnostic category.
-
Specification