Systems and methods for determining the coupling response of a non-linear variant system
An exemplary system and method are disclosed that can detect unobserved and unobservable signals in a subject, or system, using an auxiliary, secondary signal measured from the subject while the unobserved and unobservable signals are known to be present. In particular, the exemplary system and method can used to measure observed signals, e.g., in a nonlinear time-variant system (NLTV system) such as the brain, to measure the unobserved or unobservable signal of interest via response coupling that employs a test instrument configured to facilitate interactions between the primary and the auxiliary signal.
1 . A system comprising:
a processor; and
a memory having instructions stored thereon, wherein execution of the instructions by the processor causes the processor to:
receive measurements of an observed signal that is a physiological response evoked by presentation of a stimulus;
determine, via response coupling that employs a test instrument configured to select a property of the observed signal to amplify an unobserved signal and facilitate interactions between the observed signal and the unobserved signal, a probability value of a presence or non-presence of the unobserved signal via a trained classifier, wherein: (i) the unobserved signal is associated with a non-linear time-variant system, (ii) the observed signal and the unobserved signal are concurrently measured and originate from different sources such that the unobserved signal is not a component of the stimulus, and (iii) the trained classifier was trained based on training data comprising a plurality of measured unobserved signals acquired when a corresponding observed signal is present;
in response to determining a probability value indicating the presence of the unobserved signal, determine one or more inferred properties of the unobserved signal based on one or more observed signal properties; and
control a computing system or perform a diagnostic operation based, at least in part, on the one or more inferred properties.
2 . The system of claim 1 , wherein the observed signal is acquired via a brain-computer interface (BCI).
3 . The system of claim 1 , wherein the instructions to determine the probability value of a presence or non-presence of an unobserved signal comprises:
instructions to generate (i) a first band signal having frequency range from a first frequency to a second frequency of the observed signal, (ii) a second band signal having frequency range from the second frequency to a third frequency of the observed signal, (iii) a third band signal having frequency range from the third frequency to a fourth frequency of the observed signal; and
instructions to provide the first band signal, the second band signal, the third band signal, and a source signal of the first band signal, the second band signal, the third band signal to the trained classifier.
4 . The system of claim 1 , wherein the instructions to determine the probability value of a presence or non-presence of an unobserved signal comprises:
instructions to generate a probability density distribution curve from a difference between a first window of signal frequency of the observed signal and a second window of signal frequency of the observed signal, wherein the first window and the second window have different frequencies, and
instructions to determine the probability value of the presence or non-presence of the unobserved signal via a Bayesian classifier.
5 . A method comprising:
receiving, by a processor, measurements of an observed signal that is a physiological response evoked by presentation of a stimulus;
determining, by the processor via response coupling, that employs a test instrument configured to select a property of the observed signal to amplify an unobserved signal and facilitate interactions between the observed signal and the unobserved signal, a probability value of a presence or non-presence of the unobserved signal via a trained classifier, wherein: (i) the unobserved signal is associated with a non-linear time-variant system, (ii) the observed signal and the unobserved signal are concurrently measured and originate from different sources such that the unobserved signal is not a component of the stimulus, and (iii) the trained classifier was trained based on training data comprising a plurality of measured unobserved signals acquired when a corresponding observed signal is present;
in response to determining, by the processor, a probability value indicating the presence of the unobserved signal, determining one or more inferred properties of the unobserved signal based on one or more observed signal properties; and
controlling, by the processor, a computing system or performing a diagnostic operation based, at least in part, on the one or more inferred properties.
6 . The method of claim 5 , wherein the measurements of the observed signal are acquired as a physiological response to a stimulus.
7 . The method of claim 5 , wherein the observed signal comprises an offset-onset SSVEP signal or a pattern reversal SSVEP.
8 . The method of claim 5 , wherein the observed signal is acquired via a brain-computer interface (BCI).
9 . The method of claim 5 , wherein the observed signal was acquired while a display is contemporaneously presenting an observable error in movement or action of a graphical element.
10 . The method of claim 5 , wherein the probability value of the presence of the unobserved signal, or a value derived therefrom, is employed as a human-in-the-loop input signal for an AI training system.
11 . The method of claim 10 , wherein the AI training system is configured to generate a trained AI system for control of a robot.
12 . The method of claim 10 , wherein the AI training system is configured to generate a trained AI system for control of a computing task.
13 . The method of claim 5 , wherein determining the probability value of a presence or non-presence of an unobserved signal comprises:
generating a probability density distribution curve from a difference between a first window of signal frequency of the observed signal and a second window of signal frequency of the observed signal, wherein the first window and the second window have different frequencies, and
determining the probability value of the presence or non-presence of the unobserved signal via a Bayesian classifier.
14 . The method of claim 13 , wherein the probability value of presence or non- presence of the unobserved signal is a composite score from two or more scores, including a score from the Bayesian classifier.
15 . A non-transitory computer-readable medium having instructions stored thereon, wherein execution of the instructions by a processor causes the processor to:
receive measurements of an observed signal that is a physiological response evoked by presentation of a stimulus;
determine, via response coupling that employs a test instrument configured to select a property of the observed signal to amplify an unobserved signal and facilitate interactions between the observed signal and the unobserved signal, a probability value of a presence or non-presence of the unobserved signal via a trained classifier, wherein: (i) the unobserved signal is associated with a non-linear time-variant system, (ii) the observed signal and the unobserved signal are concurrently measured and originate from different sources such that the unobserved signal is not a component of the stimulus, and (iii) the trained classifier was trained based on training data comprising a plurality of measured unobserved signals acquired when a corresponding observed signal is present;
in response to determining a probability value indicating the presence of the unobserved signal, determining one or more inferred properties of the unobserved signal based on one or more observed signal properties; and
control a computing system or perform a diagnostic operation based, at least in part, on the one or more inferred properties.
16 . The non-transitory computer-readable medium of claim 15 , wherein the observed signal is acquired via a brain-computer interface (BCI).
17 . The non-transitory computer-readable medium of claim 15 , wherein the observed signal has a simultaneous combined response in a subject with the unobserved signal.
18 . The non-transitory computer-readable medium of claim 15 , wherein the probability value of the presence of the unobserved signal, or a value derived therefrom, is employed as a human-in-the-loop control signal.