Detection of epileptogenic brains with non-linear analysis of electromagnetic signals
First Claim
Patent Images
1. A method of treating a patient having epilepsy based on an quantitative assessment of an epileptogenicity level of the patient'"'"'s brain, the method comprising:
- storing, in a database, a multi-dimensional dataset including non-linear feature set values for a first plurality of patients diagnosed with epilepsy and a second plurality of patients without an epilepsy diagnosis;
receiving electroencephalography (EEG) data recorded from the brain of the patient, wherein the received EEG data includes a time series of EEG data that does not include and is not associated with epileptiform activity;
generating, for the time series of EEG data that does not include and is not associated with epileptiform activity, a plurality of additional time series of EEG data by applying, using at least one computer processor, a multiscale algorithm to the time series of EEG data;
calculating, for the time series of EEG data and each of the generated plurality of additional time series of EEG data, values for each of plurality of non-linear features in a set of non-linear features, wherein the set of non-linear features includes sample entropy and a plurality of recurrence quantitative analysis features;
determining, using at least some of the calculated values for the plurality of non-linear features in the set of non-linear features, the multi-dimensional dataset stored in the database, and a statistical learning algorithm, the epileptogenicity level of the patient'"'"'s brain; and
treating the patient as a result of the determined epileptogenicity level.
1 Assignment
0 Petitions
Accused Products
Abstract
Methods and apparatus for identifying and using at least one nonlinear feature determined from multiscale electroencephalography (EEG) data to evaluate an epileptogenicity level of a patient is described. A multiscale algorithm is applied to EEG data recorded from the patient to produce scaled EEG data. At least one nonlinear feature value for the received EEG data and/or the scaled EEG data is determined and the at least one nonlinear feature value is used, at least in part, to evaluate the epileptogenicity level of the patient.
4 Citations
17 Claims
-
1. A method of treating a patient having epilepsy based on an quantitative assessment of an epileptogenicity level of the patient'"'"'s brain, the method comprising:
-
storing, in a database, a multi-dimensional dataset including non-linear feature set values for a first plurality of patients diagnosed with epilepsy and a second plurality of patients without an epilepsy diagnosis; receiving electroencephalography (EEG) data recorded from the brain of the patient, wherein the received EEG data includes a time series of EEG data that does not include and is not associated with epileptiform activity; generating, for the time series of EEG data that does not include and is not associated with epileptiform activity, a plurality of additional time series of EEG data by applying, using at least one computer processor, a multiscale algorithm to the time series of EEG data; calculating, for the time series of EEG data and each of the generated plurality of additional time series of EEG data, values for each of plurality of non-linear features in a set of non-linear features, wherein the set of non-linear features includes sample entropy and a plurality of recurrence quantitative analysis features; determining, using at least some of the calculated values for the plurality of non-linear features in the set of non-linear features, the multi-dimensional dataset stored in the database, and a statistical learning algorithm, the epileptogenicity level of the patient'"'"'s brain; and treating the patient as a result of the determined epileptogenicity level. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A non-transitory computer-readable medium, encoded with a plurality of instructions that, when executed by a computer, performs a method of selecting a treatment for a patient, the method comprising:
-
storing, in a database, a multi-dimensional dataset including non-linear feature set values for a first plurality of patients diagnosed with epilepsy and a second plurality of patients without an epilepsy diagnosis; generating, for a time series of electroencephalograpy (EEG) data recorded from the brain of the patient that does not include and is not associated with epileptiform activity, a plurality of additional time series of EEG data by applying a multiscale algorithm to the time series of EEG data; calculating, for the time series of EEG data and each of the generated plurality of additional time series of EEG data, values for each of plurality of non-linear features in a set of non-linear features, wherein the set of non-linear features includes sample entropy and a plurality of recurrence quantitative analysis features; determining, using at least some of the calculated values for the plurality of non-linear features in the set of non-linear features, the multi-dimensional dataset stored in the database, and a statistical learning algorithm, the epileptogenicity level of the patient'"'"'s brain; and selecting a treatment for the patient as a result of the determined epileptogenicity level.
-
-
17. A computer system, comprising:
-
a database having stored thereon, a multi-dimensional dataset including non-linear feature set values for a first plurality of patients diagnosed with epilepsy and a second plurality of patients without an epilepsy diagnosis; an input interface configured to receive electroencephalography (EEG) data recorded from a brain of a patient, wherein the received EEG data includes a time series of EEG data that does not include and is not associated with epileptiform activity; at least one computer processor programmed to; generate, for the time series of EEG data that does not include and is not associated with epileptiform activity, a plurality of additional time series of EEG data by applying a multiscale algorithm to the-time series of EEG data; calculate, for the time series of EEG data and each of the generated plurality of additional time series of EEG data, values for each of plurality of non-linear features in a set of non-linear features, wherein the set of non-linear features includes sample entropy and a plurality of recurrence quantitative analysis features; determine, using at least some of the calculated values for the plurality of non-linear features in the set of non-linear features, the multi-dimensional dataset stored in the database, and a statistical learning algorithm, the epileptogenicity level of the patient'"'"'s brain; and select a treatment for the patient as a result of the determined epileptogenicity level; and an output interface configured to output an indication of the selected treatment.
-
Specification