Systems and methods for seizure detection
Systems and methods to detect seizures using analog circuitry. One example method generally includes obtaining, at a seizure detection system, one or more electroencephalogram (EEG) signals, detecting a plurality of features associated with each of the one or more EEG signals, generating a bitstream indicating a seizure probability associated with each feature of the plurality of features to yield a plurality of bitstreams indicating a plurality of seizure probabilities, and generating a seizure detection output based on the plurality of bitstreams indicating the plurality of seizure probabilities of the plurality of features.
1 . A method to detect seizures comprising:
obtaining, at a seizure detection system, one or more electroencephalogram (EEG) signals;
detecting a plurality of features associated with each of the one or more EEG signals;
generating a bitstream indicating a seizure probability associated with each feature of the plurality of features to yield a plurality of bitstreams indicating a plurality of seizure probabilities; and
generating, via feature evaluation circuitry, a seizure detection output based on the plurality of bitstreams indicating the plurality of seizure probabilities of the plurality of features, the feature evaluation circuitry including:
one or more p-channel metal-oxide semiconductor (PMOS) transistors coupled in series between a voltage rail and a common node, wherein a gate of the one or more PMOS transistors is configured to receive at least one of the plurality of bitstreams,
one or more n-channel metal-oxide semiconductor (NMOS) transistors coupled in series between the common node and a reference potential node, wherein a gate of the one or more NMOS transistors is configured to receive the at least one of the plurality of bitstreams, and
a latch comprising one or more inverters coupled between the common node and an output of the feature evaluation circuitry.
2 . The method of claim 1 , further comprising:
generating an analog signal for each feature of the plurality of features,
wherein,
the bitstream is generated based on the analog signal, and
the analog signal includes a voltage corresponding to the seizure probability.
3 . The method of claim 1 , wherein a quantity of bits, in a specific time period, of the bitstream having a specific logic state indicates the seizure probability.
4 . The method of claim 1 , wherein a probability that each bit of the bitstream has a specific logic state is equal to the seizure probability.
5 . The method of claim 1 , wherein the seizure detection output is generated via a Muller C-element circuit through stochastic computing.
6 . The method of claim 1 , further comprising:
generating a feature evaluation signal based on a comparison of the plurality of bitstreams,
wherein,
the seizure detection output is generated based on the feature evaluation signal.
7 . The method of claim 6 , wherein the generating of the feature evaluation signal includes:
setting the feature evaluation signal to a specific logic state when corresponding bits of the plurality of bitstreams have a same logic state; and
setting the feature evaluation signal to a previous logic state when the corresponding bits of the plurality of bitstreams having different logic states.
8 . The method of claim 6 , further comprising:
filtering the feature evaluation signal to yield a filtered feature evaluation signal,
wherein,
the seizure detection output is generated based on the filtered feature evaluation signal.
9 . The method of claim 1 , further comprising:
selecting, based on measurements associated with a patient, the plurality of features from candidate features; and
before implanting the seizure detection system for the patient, configuring the seizure detection system to use the plurality of features for seizure detection based on the selection.
10 . The method of claim 1 , further comprising:
performing measurements associated with the plurality of features,
wherein,
detecting the plurality of features comprises detecting whether each of the measurements meets a measurement threshold or is within a measurement range.
11 . The method of claim 10 , further comprising:
selecting, based on measurements associated with a patient, the measurement threshold or the measurement range associated with each of the plurality of features; and
before implanting the seizure detection system for the patient, configuring the seizure detection system to use the measurement threshold or the measurement range for detecting each of the plurality of features based on the selection.
12 . An apparatus to detect seizures comprising:
a feature detection circuit configured to detect a plurality of features associated with one or more electroencephalogram (EEG) signals;
a bitstream generator configured to generate a bitstream indicating a seizure probability associated with each feature of the plurality of features to yield a plurality of bitstreams indicating a plurality of seizure probabilities; and
feature evaluation circuitry configured to generate a seizure detection output based on the plurality of bitstreams indicating the plurality of seizure probabilities of the plurality of features, the feature evaluation circuitry including:
a set of p-channel metal-oxide semiconductor (PMOS) transistors coupled in series between a voltage rail and a common node, wherein a gate of each of the set of PMOS transistors are configured to receive a respective one of the plurality of bitstreams,
a set of n-channel metal-oxide semiconductor (NMOS) transistors coupled in series between the common node and a reference potential node, wherein a gate of each of the set of NMOS transistors are configured to receive the respective one of the plurality of bitstreams, and
a latch comprising a pair of inverters coupled between the common node and an output of the feature evaluation circuitry.
13 . The apparatus of claim 12 , further comprising:
an analog voltage generation circuit configured to generate an analog signal for each feature of the plurality of features,
wherein,
the bitstream is generated based on the analog signal, and
the analog signal includes a voltage corresponding to the seizure probability.
14 . The apparatus of claim 13 ,
wherein,
the bitstream generator is configured to receive the analog signal and generate the bitstream based on the analog signal,
a quantity of bits, in a specific time period, of the bitstream having a specific logic state indicates the seizure probability.
15 . The apparatus of claim 12 ,
wherein,
the feature evaluation circuitry comprises a Muller c-element circuit configured to generate a feature evaluation signal, and
the seizure detection output is generated based on the feature evaluation signal.
16 . The apparatus of claim 12 ,
wherein,
the feature evaluation circuitry is configured to generate a feature evaluation signal based on a comparison of the plurality of bitstreams, and
the seizure detection output is generated based on the feature evaluation signal.
17 . The apparatus of claim 16 , wherein the feature evaluation circuitry is configured to generate the feature evaluation signal by:
setting the feature evaluation signal to a specific logic state when corresponding bits of the plurality of bitstreams have a same logic state; and
setting the feature evaluation signal to a previous logic state when the corresponding bits of the plurality of bitstreams having different logic states.
18 . The apparatus of claim 16 ,
wherein,
the feature evaluation circuitry includes a low-pass filter configured to filter the feature evaluation signal and yield a filtered feature evaluation signal, and
the seizure detection output is generated based on the filtered feature evaluation signal.
19 . The apparatus of claim 12 ,
wherein,
the feature detection circuit is configured to perform measurements associated with each of the plurality of features, and
to detect the plurality of features, the feature detection circuit is configured to detect whether each of the measurements meets a measurement threshold or is within a measurement range.
20 . The apparatus of claim 12 , further comprising:
a signal generator configured to:
generate an electrical stimulation signal in response to the seizure detection output; and
apply the electrical stimulation signal to a patient.
21 . The apparatus of claim 20 , further comprising:
a wireless receiver coupled to the signal generator and configured to receive a wireless signal from a wireless transmitter,
wherein,
the signal generator is configured to tune the electrical stimulation signal based on the wireless signal.
22 . A seizure detection system comprising:
one or more electrodes configured to generate one or more electroencephalogram (EEG) signals;
an amplification circuit configured to amplify the one or more EEG signals to yield one or more amplified EEG signals;
a feature detection circuit configured to detect a plurality of features associated with each of the one or more amplified EEG signals;
a bitstream generator configured to generate a bitstream indicating a seizure probability associated with each feature of the plurality of features to yield a plurality of bitstreams indicating a plurality of seizure probabilities; and
feature evaluation circuitry configured to generate a seizure detection output based on the plurality of bitstreams indicating the plurality of seizure probabilities of the plurality of features, the feature evaluation circuitry including:
one or more p-channel metal-oxide semiconductor (PMOS) transistors coupled in series between a voltage rail and a common node, wherein a gate of the one or more PMOS transistors is configured to receive at least one of the plurality of bitstreams,
one or more n-channel metal-oxide semiconductor (NMOS) transistors coupled in series between the common node and a reference potential node, wherein a gate of the one or more NMOS transistors is configured to receive the at least one of the plurality of bitstreams, and
a latch comprising one or more inverters coupled between the common node and an output of the feature evaluation circuitry.