System and method for prediction of adverse events during treatment of psychological and neurological disorders
The present invention is a system and method of deriving and computing features and indices that predict the likelihood of psychological and neurological adverse events such as suicidal thoughts and/or actions. The method of the present invention further predicts the likelihood of suicidal thoughts and/or actions prior to and or during treatment for psychological disease. To obtain such features and indices, power spectrum and time domain values are derived from biopotential signals acquired from the subject being tested. The system and method identify people who are likely to experience changing, especially worsening, symptoms of psychological and neurological adverse events such as suicidal thoughts or actions and who therefore may be at risk (e.g. suicide).
1. A system for predicting a psychological and neurological adverse event from electroencephalographic (EEG) signals comprising:
a data acquisition unit for acquiring electroencephalographic signals from a subject being evaluated,
a processor for:
computing at least one parameter from said electroencephalographic signals,
deriving a first feature from said at least one parameter at a pre-treatment condition,
deriving a second feature from said at least one parameter after an initiation of treatment of the subject,
creating an index by combining said first feature and said second feature, and
predicting the adverse event through the use of said index,
wherein said treatment is pharmacological treatment, psychotherapy, electroconvulsive therapy, transmagnetic stimulation or neurostimulation and wherein said adverse event is suicidal thoughts, suicidal actions, somatic symptoms, sexual side-effects, nausea or vomiting.
2. The system for predicting a psychological and neurological adverse event from electroencephalographic signals of claim 1 , wherein said index combining said first and second features is a distance from an origin to a point represented by said first and second features.
3. The system for predicting a psychological and neurological adverse event from electroencephalographic signals of claim 1 wherein said at least one parameter is computed using a combination of relative EEG theta power and relative combined theta+alpha asymmetry.
4. The system for predicting a psychological and neurological adverse event from electroencephalographic signals of claim 1 , wherein said first feature and said second feature are relative theta frequency band power.
5. The system for predicting a psychological and neurological adverse event from electroencephalographic signals of claim 1 , wherein said first feature and said second feature are relative theta+alpha frequency band asymmetry.
6. A method of predicting a psychological and neurological adverse event from electroencephalographic (EEG) signals, comprising the steps of:
acquiring electroencephalographic signals from a subject being evaluated,
computing with a processor at least one parameter from said electroencephalographic signals,
deriving a first feature from said at least one parameter at a pre-treatment condition,
deriving a second feature from said at least one parameter after an initiation of treatment of said subject,
creating an index by combining said first feature and said second feature,
predicting the psychological and neurological adverse event through the use of said index, and
wherein said treatment is pharmacological treatment, psychotherapy, electroconvulsive therapy, transmagnetic stimulation or neurostimulation and wherein said adverse event is suicidal thoughts, suicidal actions, somatic symptoms, sexual side-effects, nausea or vomiting.
7. The method of predicting a psychological and neurological adverse event from electroencephalographic signals of claim 6 , wherein said first feature and said second feature are relative theta frequency band power.
8. The method of predicting a psychological and neurological adverse event from electroencephalographic signals of claim 6 , wherein said first feature and said second feature are relative theta+alpha frequency band asymmetry.
9. The method of predicting a psychological and neurological adverse event from electroencephalographic signals of claim 6 , wherein said index combining said first and second features is a distance from an origin to a point represented by said first and second features.
10. The method of predicting a psychological and neurological adverse event from electroencephalographic signals of claim 6 wherein said at least one parameter is computed using a combination of relative EEG theta power and relative combined theta+alpha asymmetry.