Real-time neural spike detection

1 . A method performed by a sensor, including an electrode and a computer chip, implanted in a cranium of a subject, the method comprising:

receiving, by the electrode implanted in the cranium of the subject, a neurological voltage signal, the electrode being connected to an input of the computer chip implanted in the cranium of the subject;

filtering, by bandpass filter circuitry in the implanted computer chip, the received neurological voltage signal to generate a filtered signal;

fitting, by circuitry of the implanted computer chip, the filtered signal to a model;

identifying, by the circuitry of the implanted computer chip, a set of fit values based on the model, the set of fit values comprising a first sample voltage value corresponding to a local minimum voltage value, a second sample voltage value corresponding to local maximum voltage value temporally following the local minimum voltage value; a third sample voltage value corresponding to a local maximum voltage value temporally preceding the local minimum voltage value, a first time value associated with the first sample voltage value, a second time value associated with the second sample voltage value, and a third time value associated with the third sample voltage value;

based on the set of fit values, computing, by the circuitry of the implanted computer chip, a set of characteristic values using fixed-point arithmetic calculation circuitry that uses fewer bits compared to floating-point calculations, wherein the set of characteristic values comprises a difference between the second time value and the first time value, a difference between the third time value and the second time value, a ratio of the third sample voltage value and an absolute value of the second sample voltage value, and a ratio of the first sample voltage value and the third sample voltage value;

comparing, by the circuitry of the implanted computer chip, the set of characteristic values to a corresponding set of threshold values;

based on the comparison, determining, by the circuitry of the implanted computer chip substantially in real time, whether the received neurological voltage signal comprises a neural spike; and

based on determining that the received neurological voltage signal comprises the neural spike, transmitting, by a transmitter of the computer chip, an indication of the neurological voltage signal to a receiver remote from the sensor for further analysis, the indication being smaller in size than the received neurological voltage signal, and based on determining that the received neurological voltage signal does not comprise the neural spike, discarding and refraining from transmitting the received neurological voltage signal.

2 . The method of claim 1 , further comprising:

based on the comparison, classifying the neural spike as corresponding to a spike category, of a plurality of spike categories; and

the transmitting step further comprises transmitting an indicator of the spike category and a timestamp.

3 . The method of claim 1 , wherein the filtering is executed via the bandpass filter circuitry in the computer chip, and the identifying, comparing, and determining steps are executed via a set of logic gates comprised in the computer chip.

4 . The method of claim 1 ,

wherein the transmitting step further comprises transmitting, by the transmitter the indication of the neurological voltage signal wirelessly, and

wherein the receiver is external to the cranium of the subject.

5 . The method of claim 1 , further comprising:

based on determining that the received neurological voltage signal corresponds to the neural spike, packetizing and transmitting, by the computer chip, a timestamp corresponding to the neurological voltage signal.

6 . The method of claim 1 , wherein the computing the set of characteristic values further comprises performing, via dedicated logic circuits of the circuitry of the computer chip, the fixed-point arithmetic calculations.

7 . The method of claim 1 , wherein the computing further comprises:

assigning a first characteristic value, of the set of characteristic values, a first fixed-point number representation; and

assigning a second characteristic value, of the set of characteristic values, a second fixed-point number representation, the number representations being different from each other.

8 . The method of claim 1 , wherein the receiving the neurological voltage signal further comprises receiving a neural signal via a plurality of electrodes, implanted in the cranium of the subject.

9 . A sensor, comprising:

an electrode implantable in a cranium of a subject, the electrode being configured to receive a neurological voltage signal from the cranium of the subject;

a computer chip implantable in the cranium, the computer chip being connected to the electrode and being configured to receive the neurological voltage signal from the electrode, the computer chip including

bandpass filter circuitry configured to filter the received neurological voltage signal to generate a filtered signal;

control circuitry configured to:

fit the filtered signal to a model;

identify a set of fit values based on the model, the set of fit values comprising a first sample voltage value corresponding to a local minimum voltage value, a second sample voltage value corresponding to local maximum voltage value temporally following the local minimum voltage value; a third sample voltage value corresponding to a local maximum voltage value temporally preceding the local minimum voltage value, a first time value associated with the first sample voltage value, a second time value associated with the second sample voltage value, and a third time value associated with the third sample voltage value;

based on the set of fit values, compute a set of characteristic values using fixed-point arithmetic calculation circuitry that uses fewer bits compared to floating-point calculations, wherein the set of characteristic values comprises a difference between the second time value and the first time value, a difference between the third time value and the second time value, a ratio of the third sample voltage value and an absolute value of the second sample voltage value, and a ratio of the first sample voltage value and the third sample voltage value;

compare the set of characteristic values to a corresponding set of threshold values;

based on the comparison, determine, substantially in real time, whether the received neurological voltage signal comprises a neural spike; and

based on determining that the received neurological voltage signal comprises the neural spike, cause a transmitter of the computer chip to transmit an indication of the neurological voltage signal from the sensor to a receiver that is remote from the sensor for further analysis, the indication being smaller in size than the received neurological voltage signal, and based on determining that the received neurological voltage signal does not comprise the neural spike, discard and refrain from transmitting the received neurological voltage signal; and

the transmitter configured to transmit the indication from the sensor to the receiver.

10 . The sensor of claim 9 , wherein the control circuitry of the computer chip is further configured to:

classify the neural spike as corresponding to a spike category, of a plurality of configured spike categories, based on the comparison, and

transmit output comprising an indicator of the spike category and a timestamp.

11 . The sensor of claim 9 , wherein the transmitter is further configured to transmit the indication of the neurological voltage signal wirelessly, and wherein the receiver is external to the cranium of the subject.

12 . The sensor of claim 9 , wherein the control circuitry of the computer chip is further configured to, based on determining that the received neurological voltage signal comprises the neural spike, cause the transmitter to packetize and transmit a timestamp corresponding to the neurological voltage signal.

13 . The sensor of claim 9 , wherein dedicated logic circuits of the control circuitry of the computer chip is further are configured to compute the set of characteristic values using the fixed-point arithmetic calculations.

14 . The sensor of claim 9 , wherein the control circuitry of the computer chip is further configured to:

assign a first characteristic value, of the set of characteristic values, a first fixed-point number representation; and

assign a second characteristic value, of the set of characteristic values, a second fixed-point number representation, the number representations being different from each other.

15 . A non-transitory computer-readable media storing a program that, when executed, causes a sensor to perform a method, the sensor including an electrode and a computer chip implanted in a cranium of a subject, the method comprising:

receiving, by the electrode implanted in the cranium of the subject, a neurological voltage signal, the electrode being connected to an input of the computer chip implanted in the cranium of the subject;

filtering, by bandpass filter circuitry in the implanted computer chip, the received neurological voltage signal to generate a filtered signal;

fitting, by circuitry of the implanted computer chip, the filtered signal to a model;

identifying, by the circuitry of the implanted computer chip, a set of fit values based on the model, the set of fit values comprising a first sample voltage value corresponding to a local minimum voltage value, a second sample voltage value corresponding to local maximum voltage value temporally following the local minimum voltage value; a third sample voltage value corresponding to a local maximum voltage value temporally preceding the local minimum voltage value, a first time value associated with the first sample voltage value, a second time value associated with the second sample voltage value, and a third time value associated with the third sample voltage value;

based on the set of fit values, computing, by the circuitry of the implanted computer chip, a set of characteristic values using fixed-point arithmetic calculation circuitry that uses fewer bits compared to floating-point calculations, wherein the set of characteristic values comprises a difference between the second time value and the first time value, a difference between the third time value and the second time value, a ratio of the third sample voltage value and an absolute value of the second sample voltage value, and a ratio of the first sample voltage value and the third sample voltage value;

comparing, by the circuitry of the implanted computer chip, the set of characteristic values to a corresponding set of threshold values;

based on the comparison, determining, by the circuitry of the implanted computer chip substantially in real time, whether the received neurological voltage signal comprises a neural spike; and

based on determining that the received neurological voltage signal comprises the neural spike, transmitting, by a transmitter of the computer chip, an indication of the neurological voltage signal to a receiver remote from the sensor for further analysis, the indication being smaller in size than the received neurological voltage signal, and based on determining that the received neurological voltage signal does not comprise the neural spike, discarding and refraining from transmitting the received neurological voltage signal.