Apparatus, chip, device and readable storage medium for closed-loop deep brain stimulation

Abstract

The present disclosure discloses a device, a chip, an apparatus and a readable storage medium for closed-loop deep brain stimulation, the device comprising: a controller configured to generate a control signal for deep brain stimulation based on a preset initial signal and a brain electrical feedback signal; a waveform generation module configured to generate a stimulation waveform based on the control signal; and an operation circuit configured to perform model operation for simulating physiological characteristics of a condition based on the stimulation waveform to generate a stimulation signal for deep brain stimulation. The device according to the embodiments of the present disclosure can automatically adjust and adaptively adjust the control signal to meet the needs of individualized stimulation programs.

Description

Technical Field

[0001] This disclosure generally relates to the field of medical device technology. More specifically, this disclosure relates to a device, chip, apparatus, and readable storage medium for closed-loop deep brain stimulation. Background Technology

[0002] Parkinson's disease is a degenerative neurodegenerative disorder caused by the loss of dopaminergic neurons in the substantia nigra pars compacta. Patients primarily present with motor impairments such as resting tremor, muscle rigidity, bradykinesia, and postural instability, as well as cognitive impairments such as speech difficulties and depression. Deep brain stimulation (DBS) is an important brain modulation technique for treating Parkinson's disease. It modulates pathological brain rhythms and alleviates symptoms such as motor impairments by applying high-frequency (approximately 130Hz to 180Hz) electrical pulses to deep brain tissue. DBS addresses the issues of poor efficacy and drug resistance in some patients due to drug nonspecificity, and avoids the potential for additional symptoms and brain damage caused by nucleus lesioning surgery. Furthermore, DBS can also be used to treat epilepsy, essential tremor, obsessive-compulsive disorder, anorexia, Alzheimer's disease, and depression.

[0003] However, current deep brain stimulation techniques often employ an open-loop stimulation model, requiring doctors to control stimulation parameters based on the patient's condition until the next follow-up visit. In this model, the stimulation parameters cannot be adjusted in real-time according to the patient's symptoms, failing to adapt to changes in the patient's condition and potentially affecting the stimulation's effectiveness.

[0004] In view of this, there is an urgent need to provide a technical solution for closed-loop deep brain stimulation that can be adaptively adjusted. Summary of the Invention

[0005] In order to at least address one or more of the technical problems mentioned above, this disclosure provides a device, chip, apparatus, and readable storage medium for closed-loop deep brain stimulation in several aspects.

[0006] In a first aspect, this disclosure provides an apparatus for closed-loop deep brain stimulation, comprising: a controller for generating a control signal for deep brain stimulation based on a preset initial signal and an EEG feedback signal; a waveform generation module for generating a stimulation waveform based on the control signal; and a computing circuit for performing model calculations to simulate the physiological characteristics of a disease based on the stimulation waveform, so as to generate a stimulation signal for deep brain stimulation.

[0007] In some embodiments, the controller is further configured to: determine the current oscillation amplitude error based on a preset initial signal and the EEG feedback signal generated after the previous stimulus; determine the current control signal increment based on the current oscillation amplitude error; and generate the current control signal based on the current control signal increment and the control signal output previously.

[0008] In other embodiments, the device further includes: an analog-to-digital converter for performing analog-to-digital conversion on the EEG feedback signal generated after the previous stimulus to obtain a corresponding feedback digital signal; and the controller is further configured to determine the current oscillation amplitude error based on the difference between a preset initial signal and the feedback digital signal.

[0009] In some other embodiments, the controller further comprises, in determining the current control signal increment,: determining a first feature value based on the current oscillation amplitude error; determining a second feature value based on the difference between the current oscillation amplitude error and the previous oscillation amplitude error; and determining the current control signal increment based on the first feature value and the second feature value.

[0010] In some embodiments, the controller further comprises, in determining the current control signal increment based on a first feature value and a second feature value, performing a normalization operation on the first feature value and the second feature value to obtain a normalization result; and determining the current control signal increment based on the normalization result.

[0011] In other embodiments, the controller, when performing normalization, is configured to: determine the normalized neuron weight corresponding to each feature value in the current control operation based on the previous oscillation amplitude error, the previous output control signal, and the feature value determined in the previous control operation; and perform a weighted summation of the first feature value and the second feature value based on the normalized neuron weights corresponding to the first feature value and the second feature value, respectively.

[0012] In some other embodiments, the controller obtains the current control signal increment based on the following logical operation: Where Δg represents the current control signal increment, and K N Indicates the calibration coefficient. x represents the normalization result. i (n) represents the i-th eigenvalue, ω′ i (n) represents the normalized neuron weights, i = 1, 2.

[0013] In some embodiments, the controller obtains the normalized neuron weights according to the following logical operation: Where, ω′ i(n) represents the normalized neuron weights, n represents time point n, n-1 represents the previous time point n, and η i Let x represent the integral and proportional learning rates, e(n-1) represent the previous oscillation amplitude error, g(n-1) represent the previous output control signal, and x represent the integral and proportional learning rates. i (n-1) represents the i-th feature value determined in the previous control operation, where i = 1, 2.

[0014] In other embodiments, the controller is also configured to: obtain the control signal of the previous output through a delay operation.

[0015] In some other embodiments, the arithmetic circuit includes an arctangent circuit and a division circuit.

[0016] In some embodiments, the arctangent operation circuit is used to perform an arctangent operation on the input signal generated based on the stimulation waveform and the EEG feedback signal to obtain an arctangent operation result; and the division operation circuit is used to perform a division operation on the arctangent operation result to generate the stimulation signal.

[0017] In other embodiments, the arctangent circuit performs the following arctangent operation: Where u represents the result of the arctangent operation, y represents the input signal of the arctangent operation circuit, and h represents the dopamine parameter.

[0018] In some other embodiments, the division circuit performs the following division operation: Where G(u) represents the stimulus signal, u represents the result of the arctangent operation, k represents the gain, and -b represents the open-loop heavy pole.

[0019] In a second aspect, this disclosure provides a chip including the apparatus described in any of the first aspects of this disclosure.

[0020] In some embodiments, the chip further includes: a clock management unit for controlling the operating time of the device; and a processing unit for controlling the clock switch of the clock management unit.

[0021] In a third aspect, this disclosure provides an apparatus for closed-loop deep brain stimulation, comprising: a processor for executing program instructions; and a memory storing the program instructions, which, when loaded and executed by the processor, cause the processor to perform the following closed-loop deep brain stimulation method: generating a control signal for deep brain stimulation based on a preset initial signal and an EEG feedback signal; generating a stimulation waveform based on the control signal; and inputting the stimulation waveform into a mathematical model for simulating the physiological characteristics of a disease to generate a stimulation signal for deep brain stimulation.

[0022] In some embodiments, when the program instructions are executed by the processor, the device also performs the following operations in generating the control signal: determining the current oscillation amplitude error based on a preset initial signal and the EEG feedback signal generated after the previous stimulus; determining the current control signal increment based on the current oscillation amplitude error; and determining the current control signal based on the current control signal increment and the previously output control signal.

[0023] In other embodiments, when the program instructions are executed by the processor, the device also performs the following operations in determining the current oscillation amplitude error: performs analog-to-digital conversion on the EEG feedback signal generated after the previous stimulus to obtain a corresponding feedback digital signal; and determines the current oscillation amplitude error based on the difference between a preset initial signal and the feedback digital signal.

[0024] In some other embodiments, when the program instructions are executed by the processor, the device also performs the following operations in determining the current control signal increment: determining a first feature value based on the current oscillation amplitude error; determining a second feature value based on the difference between the current oscillation amplitude error and the previous oscillation amplitude error; and determining the current control signal increment based on the first feature value and the second feature value.

[0025] In some embodiments, when the program instructions are executed by the processor, the device further causes to perform the following operations in determining the current control signal increment based on a first feature value and a second feature value: normalizing the first feature value and the second feature value to obtain a normalized result; and determining the current control signal increment based on the normalized result.

[0026] In other embodiments, when the program instructions are executed by the processor, the device also performs the following operations during the normalization operation: determining the normalized neuron weight corresponding to each feature value in the current control operation based on the previous oscillation amplitude error, the previous output control signal, and the feature value determined when the control signal was generated last time; and performing a weighted summation of the first feature value and the second feature value based on the normalized neuron weights corresponding to the first feature value and the second feature value, respectively.

[0027] In some other embodiments, when the program instructions are executed by the processor, the device also calculates the current control signal increment in the following manner: Where Δg represents the current control signal increment, K N Indicates the calibration coefficient. x represents the normalization result. i (n) represents the i-th eigenvalue, ω′ i (n) represents the normalized neuron weights, i = 1, 2.

[0028] In some embodiments, when the program instructions are executed by the processor, the device also calculates the normalized neuron weights in the following manner: Where, ω′ i (n) represents the normalized neuron weights, n represents time point n, n-1 represents the previous time point n, and η i The integral and proportional learning rates, e(n-1) represents the previous oscillation amplitude error, g(n-1) represents the previous output control signal, and x i (n-1) represents the i-th feature value determined when the control signal was generated last time, i = 1, 2.

[0029] In other embodiments, when the program instructions are run by the processor, the device also performs the following operation: obtaining the control signal of the last output by delaying the operation.

[0030] In some other embodiments, the mathematical model includes arctangent and division operations.

[0031] In some embodiments, when the program instructions are executed by the processor, the device further causes to perform the following operations in generating the stimulation signal: generating an input signal for input into a mathematical model based on the stimulation waveform and the EEG feedback signal; performing an arctangent operation on the input signal to obtain an arctangent result; and performing a division operation on the arctangent result to generate the stimulation signal.

[0032] In other embodiments, when the program instructions are executed by the processor, the device also performs the following arctangent operation: Where u represents the result of the arctangent operation, y represents the input signal of the arctangent operation, and h represents the dopamine parameter.

[0033] In some other embodiments, when the program instructions are executed by the processor, the device also performs the following division operation: Where G(u) represents the stimulus signal, u represents the result of the arctangent operation, k represents the gain, and -b represents the open-loop heavy pole.

[0034] In a fourth aspect, this disclosure provides a computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by one or more processors, implement the method performed by the device as described in any of the third aspects of this disclosure.

[0035] Using the technical solution for closed-loop deep brain stimulation provided above, the solution in this disclosure embodiment generates control signals by combining EEG feedback signals, enabling automated and adaptive adjustment of the control signals to meet the needs of personalized stimulation programs. Simultaneously, by performing model calculations to simulate the physiological characteristics of the disease, stimulation signals corresponding to the patient's disease characteristics and recognizable by the patient's brain can be generated. Attached Figure Description

[0036] The above and other objects, features, and advantages of exemplary embodiments of this disclosure will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of this disclosure are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0037] Figure 1 This is a schematic block diagram illustrating an apparatus for closed-loop deep brain stimulation according to an embodiment of this disclosure;

[0038] Figure 2 This is a schematic diagram illustrating the stimulation waveform according to an embodiment of this disclosure;

[0039] Figure 3 This is a schematic diagram illustrating the control flow of a controller according to an embodiment of this disclosure;

[0040] Figure 4 This is a schematic diagram illustrating the operation flow of the arithmetic circuit according to an embodiment of this disclosure;

[0041] Figure 5 This is a function graph illustrating the arctangent operation according to an embodiment of this disclosure;

[0042] Figure 6 This is a function graph illustrating a division operation according to an embodiment of this disclosure;

[0043] Figure 7 This is a schematic block diagram illustrating a chip according to an embodiment of this disclosure; and

[0044] Figure 8 This is a schematic block diagram illustrating a device for closed-loop deep brain stimulation according to an embodiment of this disclosure. Detailed Implementation

[0045] The technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, not all of them. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0046] It should be understood that the terms “comprising” and “including” used in this disclosure and claims indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0047] It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure. As used in this disclosure and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this disclosure and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0048] As used in this specification and claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [described condition or event] is detected" may be interpreted, depending on the context, as "once determined," "in response to determination," "once [described condition or event] is detected," or "in response to detection of [described condition or event]."

[0049] The specific embodiments disclosed herein will now be described in detail with reference to the accompanying drawings.

[0050] Figure 1 This is a schematic block diagram illustrating an apparatus for closed-loop deep brain stimulation according to an embodiment of this disclosure. Figure 1 As shown, the device 100 may include a controller 110, a waveform generation module 120, and a computing circuit 130. The controller 110 can be used to generate a control signal for deep brain stimulation based on a preset initial signal and an EEG feedback signal; the waveform generation module 120 can be used to generate a stimulation waveform based on the control signal; and the computing circuit 130 can be used to perform model calculations to simulate the physiological characteristics of the disease based on the stimulation waveform, so as to generate a stimulation signal for deep brain stimulation.

[0051] In some embodiments, the preset initial signal can be set based on experience. In other embodiments, the preset initial signal can be an expected value of multiple EEG feedback signals. For example, it could be the expected value of EEG feedback signals received within a preset time period, or the expected value of all EEG feedback signals received throughout the patient's medical history, and so on. The EEG feedback signal x can include feedback information generated by brain nuclei after receiving a stimulation signal. In some embodiments, the preset initial signal can be updated in real time based on the EEG feedback signals received by the device 100. In some embodiments, the EEG feedback signal x can be detected via EEG electrodes.

[0052] In some embodiments, the controller 110 can generate a control signal based on a preset initial signal and the result of sampling the EEG feedback signal. In other embodiments, the device 100 may further include an analog-to-digital converter (ADC), which can be used to perform analog-to-digital conversion on the EEG feedback signal x generated after the previous stimulus to obtain a corresponding feedback digital signal. An analog-to-digital converter (ADC) is an electronic component capable of converting analog signals into digital signals. In still other embodiments, the controller 110 may be further used to determine the current oscillation amplitude error based on the difference between the preset initial signal and the feedback digital signal; and to generate a control signal based on the current oscillation amplitude error. That is, the controller 110 can generate a control signal based on the difference between the preset initial signal and the EEG feedback signal generated after the previous stimulus.

[0053] In some embodiments, the control signal may include control parameters for generating a stimulation waveform, such as at least one of amplitude, frequency (or pulse width), and period. In other embodiments, the waveform generation module 120 may be used to generate, for example, a waveform based on the control signal. Figure 2 The stimulation waveform is shown in the figure. Specifically, as shown in the figure... Figure 2 As shown, the waveform generation module can generate a square-wave-shaped stimulation waveform based on the amplitude a, frequency d, and period T in the control signal. In some embodiments, the waveform generation module 120, or waveform generation module, can be implemented in software and / or hardware. In some embodiments, the waveform generation module 120 can be implemented as an integrated circuit. The waveform generation module 120 can be implemented using existing or future implementable modules, devices, circuits, or algorithms for generating stimulation waveforms, such as the waveform generation module used to drive a brain pacemaker in a conventional DBS system, and is not limited herein.

[0054] Return below Figure 1Continuing the description, further, the arithmetic circuit 130 can perform model calculations to simulate the physiological characteristics of a symptom. These model calculations can include: constructing a mathematical model to simulate the physiological characteristics of the symptom to be stimulated, and performing calculations using this mathematical model. Specifically, the stimulation waveform generated by the waveform generation module 120 can be input into the mathematical model simulated by the arithmetic circuit 130 for calculation, and a stimulation signal that can be recognized by the brain can be output.

[0055] In some applications, the conditions to be stimulated can include those that can be treated with deep brain stimulation, such as Parkinson's disease, epilepsy, and essential tremor. The physiological characteristics can be the pathophysiological mechanisms of the corresponding condition during an attack, such as the resulting neural responses in the brain. Taking Parkinson's disease as an example, research has shown that Parkinson's disease attacks cause abnormal synchronous oscillations of the basal ganglia. Therefore, a mathematical model can be constructed to simulate the hypersynchronous oscillation behavior of the basal ganglia during a Parkinson's attack, and the stimulation waveform can be calculated based on this model. With this setup, the generated stimulation signals can be recognized by the brain, and the brain can produce a stimulus response specific to the corresponding condition.

[0056] The above combination Figure 1 An exemplary description of a device for closed-loop deep brain stimulation according to embodiments of this disclosure has been provided. It is understood that by generating control signals based on preset initial signals and EEG feedback signals, a closed-loop deep brain stimulation scheme containing feedback information can be realized. The output stimulation signal can be adjusted in a timely manner according to the feedback EEG signals, thereby improving the adaptability and automated control of the device, so as to better adapt to the transient or long-term changes of the disease, without reducing or eliminating the reliance on manual adjustment by doctors.

[0057] Figure 3 This is a schematic diagram illustrating the control flow of a controller according to an embodiment of this disclosure. As will be seen from the following description, Figure 3 The control scheme shown can be Figure 1 The controller 110 shown is a specific manifestation of the control operation of the controller, therefore, in conjunction with the preceding text Figure 1 The description of controller 110 also applies to the following... Figure 3 In the description.

[0058] like Figure 3 As shown, in some embodiments, the controller according to embodiments of this disclosure may be further configured to: based on a preset initial signal Based on the EEG feedback signal x generated after the previous stimulus, determine the current oscillation amplitude error e(n); based on the current oscillation amplitude error e(n), determine the current control signal increment Δg; and based on the current control signal increment Δg and the previous output control signal g(n-1), generate the current control signal g(n).

[0059] In the above text, 'n' can represent a time point n, and 'n-1' can represent the previous time point n. Figure 3 In this context, g(n) can represent the current control signal output by the controller, and g(n-1) can represent the control signal output by the controller during the previous control operation. In some embodiments, the controller can generate the current control signal g(n) based on the sum of the current control signal increment Δg and the previously output control signal g(n-1). In other embodiments, the controller may include an adder (or adding circuit) for performing the summation operation of the current control signal increment Δg and the previously output control signal g(n-1). In still other embodiments, the controller may include a filter to output the current control signal g(n) based on the input current control signal increment Δg and the previously output control signal g(n-1).

[0060] In other embodiments, the controller may further be used to determine the current control signal increment Δg by: determining a first characteristic value based on the current oscillation amplitude error e(n); determining a second characteristic value based on the difference between the current oscillation amplitude error e(n) and the previous oscillation amplitude error e(n-1); and determining the current control signal increment Δg based on the first characteristic value and the second characteristic value. In some embodiments, the controller may include a rectifier that can be used to output the first characteristic value and the second characteristic value based on the input current oscillation amplitude error.

[0061] In some embodiments, the first characteristic value and the second characteristic value can be expressed as: x1(n) = e(n); x2(n) = e(n) - e(n-1); where x1(n) represents the first characteristic value generated in the current control operation, x2(n) represents the second characteristic value generated in the current control operation, e(n) represents the current oscillation amplitude error, and e(n-1) represents the previous oscillation amplitude error. e(n) and e(n-1) can both be discrete-time signals. e(n-1) can be the oscillation amplitude error determined by the controller during the previous generation of the control signal.

[0062] In other embodiments, the controller may further be configured to: normalize the first and second feature values ​​to obtain a normalized result; and determine the current control signal increment based on the normalized result. In some embodiments, the normalization operation may include a weighted summation operation. In other embodiments, the first and second feature values ​​may be weighted summed according to the normalized neuron weights corresponding to the first and second feature values, respectively.

[0063] In some other embodiments, the normalization result can be directly determined as the current control signal increment. In some embodiments, the normalization result can be adjusted to determine the current control signal increment. For example, the product of the normalization result and the calibration coefficient can be used as the current control signal increment, where the calibration coefficient can be a positive number and its value can be set as needed.

[0064] In one specific implementation, the controller can obtain the current control signal increment based on the following logical operation:

[0065]

[0066] Where Δg can represent the current control signal increment, K N Indicates the calibration coefficient. x represents the normalization result. i (n) represents the i-th eigenvalue, ω′ i (n) represents the normalized neuron weights corresponding to the i-th feature value, where i = 1, 2.

[0067] In some embodiments, the controller, when performing normalization, may further be configured to: determine the normalized neuron weights corresponding to each feature value in the current control operation based on the previous oscillation amplitude error, the previous output control signal, and the feature values ​​determined in the previous control operation; and perform a weighted summation of the first feature value and the second feature value based on the normalized neuron weights corresponding to the first feature value and the second feature value, respectively. In other embodiments, the controller may also be configured to: obtain the previous output control signal through a delay operation. In some embodiments, the controller may include a delay circuit to perform the delay operation.

[0068] In other embodiments, the controller can determine the weight coefficient corresponding to each feature value in the current control operation based on the previous oscillation amplitude error, the previous output control signal, the feature value determined in the previous control operation, and the weight coefficient determined in the previous control operation. Then, the controller can determine the normalized neuron weight corresponding to each feature value based on the weight coefficient corresponding to each feature value.

[0069] Furthermore, in some embodiments, the controller can obtain normalized neuron weights based on the following logical operations:

[0070]

[0071]

[0072] Where, ω′ i (n) represents the normalized neuron weights corresponding to the i-th feature value, n represents time point n, n-1 represents the previous time point n, and η i ω represents the integral and proportional learning rates corresponding to the i-th feature value. i (n) represents the weight coefficient corresponding to the i-th eigenvalue, ω i (n-1) represents the weight coefficient corresponding to the i-th eigenvalue determined in the previous control operation, e(n-1) represents the previous oscillation amplitude error, g(n-1) represents the previous output control signal, and x i (n-1) represents the i-th characteristic value determined in the previous control operation, where i = 1, 2. The previous control operation is the controller operation performed when the control signal was generated last time. In some other embodiments, ω i The initial value ω of (n) i (0) can be set as needed; there are no restrictions here. In some embodiments, η i It can be a positive number, and its specific value can be set as needed.

[0073] The above combination Figure 3 The control operations of the controller according to embodiments of this disclosure have been described exemplary. It is understood that the controller according to embodiments of this disclosure can implement the above functions through logic circuits or through microprogrammed control instructions, and those skilled in the art can choose as needed. Furthermore, to better understand the operational circuits in the embodiments of this disclosure, the following will be combined with… Figure 4 An exemplary description is provided.

[0074] Figure 4 This is a schematic diagram illustrating the operational flow of the arithmetic circuit according to an embodiment of this disclosure. Figure 4As shown in the illustration, the computational circuit of this embodiment may include an arctangent operation circuit 410 and a division operation circuit 420. The arctangent operation circuit 410 and the division operation circuit 420 can be used to simulate the hypersynchronous oscillation behavior of the basal ganglia in Parkinson's disease. In some embodiments, the stimulation waveform generated in the waveform generation module can be directly input to the arctangent operation circuit 410 for arctangent operation. In other embodiments, the arctangent operation circuit 410 can be used to perform arctangent operation on the input signal generated based on the stimulation waveform and the EEG feedback signal x to obtain the arctangent operation result; and the division operation circuit 420 can be used to perform division operation on the arctangent operation result to generate a stimulation signal.

[0075] In some embodiments, the input signal to the arctangent circuit 410 can be generated by fusing the stimulus waveform and the EEG feedback signal x. In other embodiments, the arctangent circuit 410 can be implemented by running a coordinate rotation digital computer (Cordic) method. In one specific embodiment, Figure 5 This is a function graph illustrating the arctangent operation according to an embodiment of this disclosure, such as... Figure 5 As shown, the arctangent circuit 410 can perform the following arctangent operation:

[0076]

[0077] Where u can represent the result of the arctangent operation, y can represent the input signal of the arctangent operation circuit 410, and h can represent the dopamine parameter. In some embodiments, y can be used to simulate the local field potential (LFP) signal of a nerve nucleus, and u can represent the synaptic current of the nerve nucleus, which is used as input to the division operation circuit 420.

[0078] In other embodiments, the division circuit 420 may include a divider to perform division on the result of the arctangent operation. Figure 6 This is a function graph illustrating a division operation according to an embodiment of this disclosure. For example... Figure 6 As shown, in one specific embodiment, the division circuit 420 can perform the following division operation on the input arctangent result:

[0079]

[0080] Where G(u) represents the stimulus signal, u represents the arctangent result, k represents the gain, and -b represents the open-loop heavy pole. In some embodiments, the arctangent result u can be expressed in complex form, for example, u = σ + jω, where σ represents the real part, ω represents the imaginary part, and j 2=-1. The open-loop multiple pole -b can be used to determine the oscillation frequency of the neural network, and the gain k can be used to determine the value of the dopamine parameter h when oscillation occurs. For example, when k = b, the stimulation signal output by the device will only cause the brain's neural nuclei to produce the desired constant-amplitude oscillation state when h ≤ 1 / π ≈ 0.3183. It can be understood that the neural network here refers to the network composed of neurons in the brain connected by synapses. The open-loop multiple pole -b, or open-loop pole / multiple pole, is a concept in complex functions. Open-loop means that the input and output are decoupled, that is, G(u) no longer participates in the next calculation.

[0081] In some other embodiments, the arctangent operation circuit 410 and the division operation circuit 420 described above can be implemented using Verilog to perform fast arctangent and division operations. Verilog generally refers to Verilog HDL, which is a hardware description language that can describe the structure and behavior of digital system hardware in text form. It can be used to represent logic circuit diagrams, logic expressions, and the logical functions performed by digital logic systems.

[0082] The above combination Figures 4-6 The computational circuit according to the embodiments disclosed herein has been described exemplarily. It is understood that the operations of the arctangent and division circuits can simulate the hypersynchronous oscillation behavior of the basal ganglia in Parkinson's disease. This allows the stimulus waveform generated by the waveform generation module to be processed by the computational circuits to generate a stimulus signal that elicits a desired response from the brain. This stimulus signal can be directly transmitted to the brain electrodes for corresponding stimulation. It is also understood that the arctangent and division circuits are exemplary and not limiting. In other application scenarios, mathematical model computational circuits targeting the pathophysiological mechanisms of other diseases can be set up to generate targeted stimulus signals for the corresponding diseases.

[0083] Furthermore, according to the technical solution disclosed herein, a chip is also provided in the second aspect, which may include the chip described above in conjunction with... Figures 1-6 The device described herein may be any of the aforementioned devices. In some embodiments, the chip may include, for example, a brain-computer interface (BCI) chip. The BCI chip may be a system-on-a-chip (SoC) designed to integrate most of the components required for a BCI, thereby effectively reducing the size and power consumption of the BCI system. It may include the device of the embodiments disclosed herein to achieve real-time monitoring of EEG feedback signals and to perform closed-loop adaptive adjustment of the stimulation signal based on the received EEG feedback signals after stimulation for a specific symptom. In yet other embodiments, the device according to the embodiments disclosed herein may be implemented in the form of an application-specific integrated circuit (ASIC) within the chip.

[0084] Figure 7This is a schematic block diagram illustrating a chip according to an embodiment of this disclosure. Figure 7 As shown, in some other embodiments, chip 700 may include device 710, clock management unit 720 and processing unit 730, wherein clock management unit 720 may be used to control the running time of device 710 and processing unit 730 may be used to control the clock switch of clock management unit 720.

[0085] In some embodiments, the clock management unit 720 can generate a clock signal to control the on or off of the device 710. That is, the device 710 will only operate when the clock management unit 720 inputs a clock signal to the device 710; when the clock signal input by the clock management unit 720 to the device 710 is turned off, the device 710 stops operating. Therefore, the timing of the clock signal input by the clock management unit 720 to the device 710 determines the operating time of the device 710.

[0086] With this configuration, device 710 can be activated only when stimulation signal calculation is required. For example, in some applications where the stimulation scheme involves one minute of stimulation over ten minutes, a clock signal can be input to device 710 via clock management unit 720 during the one minute of stimulation required to activate device 710. During the remaining nine minutes of the ten minutes, the clock signal input from clock management unit 720 to device 710 is disabled to keep device 710 off. Since device 710 does not consume power when off, the clock management function of clock management unit 720 reduces the overall power consumption of device 710, thereby reducing the overall power consumption of chip 700 and ensuring stable, low-power operation of both device 710 and chip 700 over extended periods.

[0087] In particular, in the application scenario of chip implantation in the skull, since the chip also needs a battery to power it, the clock switch function of the clock management unit 720 can reduce the overall operating power consumption of the device 710, thereby maximizing the battery life of the treatment device and extending the charging cycle for the patient.

[0088] In some embodiments, the device 710 needs to perform arctangent and division operations. Since the computational complexity of arctangent and division operations is high, the power consumption of the operations is high. Therefore, by integrating the device 710 onto the chip 700 and utilizing the clock management function of the clock management unit 720, the power consumption of arctangent and division operations can be reduced to the greatest extent, thereby extending the battery life of the treatment device (e.g., including brain electrodes and brain-computer interface chips).

[0089] Furthermore, the processing unit 730 can control the start and stop of the device 710 by controlling the clock switch of the clock management unit 720. In other embodiments, the processing unit 730 can control the clock switch of the clock management unit 720 through its register. In some application scenarios, the processing unit 730 can turn on the clock switch of the clock management unit 720 when data processing (e.g., EEG feedback signals) is required, and can control the device 710 to read the data to be processed to calculate the stimulation signal; when the device 710 has completed the calculation, the processing unit 730 can control the clock management unit 720 to turn off the clock switch, and perform subsequent operations on the stimulation signal calculated by the device 710, such as controlling the output to the EEG electrodes.

[0090] According to the technical solution disclosed herein, a device for closed-loop deep brain stimulation is also provided in the third aspect, which will be discussed below. Figure 8 Please provide an explanation.

[0091] Figure 8 This is a schematic block diagram illustrating a device for closed-loop deep brain stimulation according to an embodiment of this disclosure. Figure 8 As shown, the device 800 may include: a processor 810 for executing program instructions; and a memory 820 for storing the program instructions, which, when loaded and executed by the processor 810, enable the processor 810 to perform the following closed-loop deep brain stimulation method: generating a control signal for deep brain stimulation based on a preset initial signal and an EEG feedback signal; generating a stimulation waveform based on the control signal; and inputting the stimulation waveform into a mathematical model for simulating the physiological characteristics of the disease to generate a stimulation signal for deep brain stimulation.

[0092] In some embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to perform the following operations in generating the control signal: determining the current oscillation amplitude error based on a preset initial signal and the EEG feedback signal generated after the previous stimulus; determining the current control signal increment based on the current oscillation amplitude error; and determining the current control signal based on the current control signal increment and the control signal output previously.

[0093] In other embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to perform the following operations in determining the current oscillation amplitude error: perform analog-to-digital conversion on the EEG feedback signal generated after the previous stimulus to obtain a corresponding feedback digital signal; and determine the current oscillation amplitude error based on the difference between the preset initial signal and the feedback digital signal.

[0094] In some other embodiments, when the program instructions are executed by the processor 810, the device 800 is also caused to perform the following operations in determining the current control signal increment: determining a first feature value based on the current oscillation amplitude error; determining a second feature value based on the difference between the current oscillation amplitude error and the previous oscillation amplitude error; and determining the current control signal increment based on the first feature value and the second feature value.

[0095] In some embodiments, when the program instructions are executed by the processor 810, the device 800 is also caused to perform the following operations in determining the current control signal increment based on the first feature value and the second feature value: normalizing the first feature value and the second feature value to obtain a normalized result; and determining the current control signal increment based on the normalized result.

[0096] In other embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to perform the following operations during the normalization operation: determine the normalized neuron weight corresponding to each feature value in the current control operation based on the previous oscillation amplitude error, the previous output control signal, and the feature value determined when the control signal was generated last time; and perform a weighted summation of the first feature value and the second feature value based on the normalized neuron weights corresponding to the first feature value and the second feature value, respectively.

[0097] In some other embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to calculate the current control signal increment in the following manner: Where Δg represents the current control signal increment, K N Indicates the calibration coefficient. x represents the normalization result. i (n) represents the i-th eigenvalue, ω′ i (n) represents the normalized neuron weights, i = 1, 2.

[0098] In some embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to calculate the normalized neuron weights in the following manner: Where, ω′ i (n) represents the normalized neuron weights, n represents time point n, n-1 represents the previous time point n, and η i The integral and proportional learning rates, e(n-1) represents the previous oscillation amplitude error, g(n-1) represents the previous output control signal, and x i (n-1) represents the i-th feature value determined when the control signal was generated last time, i = 1, 2.

[0099] In other embodiments, when the program instructions are run by the processor 810, the device 800 is also enabled to perform the following operation: obtain the control signal of the last output by means of a delay.

[0100] In some other embodiments, the mathematical model may include arctangent and division operations.

[0101] In some embodiments, when the program instructions are executed by the processor 810, the device 800 is also enabled to perform the following operations in generating the stimulus signal: generating an input signal for input into a mathematical model based on the stimulus waveform and the EEG feedback signal; performing an arctangent operation on the input signal to obtain an arctangent operation result; and performing a division operation on the arctangent operation result to generate the stimulus signal.

[0102] In other embodiments, when the program instructions are run by the processor 810, the device 800 also performs the following arctangent operation: Where u represents the result of the arctangent operation, y represents the input signal of the arctangent operation, and h represents the dopamine parameter.

[0103] In some other embodiments, when the program instructions are run by the processor 810, the device 800 is also enabled to perform the following division operation: Where G(u) represents the stimulus signal, u represents the result of the arctangent operation, k represents the gain, and -b represents the open-loop heavy pole.

[0104] It is understood that the specific implementation of the device according to the embodiments of this disclosure has been described in detail above in conjunction with the implementation of the apparatus, and will not be repeated here.

[0105] Furthermore, according to the technical solution disclosed herein, a computer-readable storage medium is also provided in the fourth aspect, having stored thereon computer-readable instructions that, when executed by one or more processors, implement the method performed by the device as described in any of the third aspects of this disclosure.

[0106] Computer-readable storage media can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this disclosure can be implemented by computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0107] While numerous embodiments of this disclosure have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and intent of this disclosure. It should be understood that various alternatives to the embodiments of this disclosure described herein may be employed in the practice of this disclosure. The appended claims are intended to define the scope of this disclosure and therefore cover equivalents or alternatives within the scope of these claims.

Claims

1. A device for closed-loop deep brain stimulation, comprising: The controller is used to generate control signals for deep brain stimulation based on preset initial signals and EEG feedback signals; A waveform generation module is used to generate a stimulation waveform based on the control signal; as well as An arithmetic circuit is used to perform model calculations based on the stimulation waveform to simulate the physiological characteristics of the disease, thereby generating a stimulation signal for deep brain stimulation. The arithmetic circuit includes an arctangent circuit and a division circuit, which are used to simulate the hypersynchronous oscillation behavior of the basal ganglia in Parkinson's disease. The arctangent circuit performs an arctangent operation on the input signal generated based on the stimulation waveform and the EEG feedback signal to obtain an arctangent operation result. The division circuit performs a division operation on the arctangent operation result to generate the stimulation signal. The arctangent circuit described above performs the following arctangent operation: ； Where u represents the result of the arctangent operation, y represents the input signal of the arctangent operation circuit, and h represents the dopamine parameter; The division circuit performs the following division operation: ； Where G(u) represents the stimulus signal, u represents the result of the arctangent operation, k represents the gain, and -b represents the open-loop heavy pole.

2. The apparatus of claim 1, wherein the controller is further configured to: Based on the preset initial signal and the EEG feedback signal generated after the previous stimulus, the current oscillation amplitude error is determined. Based on the current oscillation amplitude error, determine the current control signal increment; and The current control signal is generated based on the current control signal increment and the previous output control signal.

3. The apparatus according to claim 2, further comprising: An analog-to-digital converter is used to convert the EEG feedback signal generated after the previous stimulus into a digital signal to obtain the corresponding feedback signal. as well as The controller is further configured to determine the current oscillation amplitude error based on the difference between the preset initial signal and the feedback digital signal.

4. The apparatus according to claim 2 or 3, wherein the controller further comprises, in determining the current control signal increment,: Based on the current oscillation amplitude error, determine the first characteristic value; The second characteristic value is determined based on the difference between the current oscillation amplitude error and the previous oscillation amplitude error; as well as The current control signal increment is determined based on the first feature value and the second feature value.

5. The apparatus of claim 4, wherein the controller further configures, in determining the current control signal increment based on the first characteristic value and the second characteristic value, to: Normalize the first eigenvalue and the second eigenvalue to obtain a normalized result; and Based on the normalization result, the current control signal increment is determined.

6. The apparatus of claim 5, wherein the controller, during the normalization operation, is configured to: Based on the previous oscillation amplitude error, the previous output control signal, and the feature values ​​determined in the previous control operation, determine the normalized neuron weights corresponding to each feature value in the current control operation; and Based on the normalized neuron weights corresponding to the first and second eigenvalues, respectively, the first and second eigenvalues ​​are summed using a weighted average.

7. The apparatus of claim 5, wherein the controller obtains the current control signal increment according to the following logical operation: ； in, K represents the increment of the current control signal. N Indicates the calibration coefficient. This represents the normalization result. Represents the i-th eigenvalue. This represents the normalized neuron weights, i=1, 2.

8. The apparatus of claim 6, wherein the controller obtains the current control signal increment according to the following logical operation: ； in, K represents the increment of the current control signal. N Indicates the calibration coefficient. This represents the normalization result. Represents the i-th eigenvalue. This represents the normalized neuron weights, i=1, 2.

9. The apparatus according to any one of claims 6-8, wherein the controller obtains the normalized neuron weights according to the following logical operation: ； ； in, This represents the normalized neuron weights, where n represents time point n, and n-1 represents the previous time point n. Representing integral and proportional learning rates, This represents the weight coefficient corresponding to the i-th eigenvalue. This represents the weight coefficient corresponding to the i-th eigenvalue determined in the previous control operation. This indicates the error in the amplitude of the previous oscillation. This indicates the control signal output last time. This represents the i-th feature value determined in the previous control operation, where i = 1 or 2.

10. The apparatus according to any one of claims 2, 3, or 5-8, wherein the controller is further configured to: The control signal from the previous output is obtained through a delay operation.

11. The apparatus of claim 4, wherein the controller is further configured to: The control signal from the previous output is obtained through a delay operation.

12. The apparatus of claim 9, wherein the controller is further configured to: The control signal from the previous output is obtained through a delay operation.

13. A chip comprising the apparatus according to any one of claims 1-12.

14. The chip according to claim 13, further comprising: A clock management unit is used to control the operating time of the device; and The processing unit is used to control the clock switch of the clock management unit.

15. A device for closed-loop deep brain stimulation, comprising: A processor is used to execute program instructions; as well as A memory storing the program instructions, which, when loaded and executed by the processor, cause the processor to perform the following closed-loop deep brain stimulation method: Based on preset initial signals and EEG feedback signals, control signals for deep brain stimulation are generated. Based on the control signal, a stimulation waveform is generated; and The stimulation waveform is input into a mathematical model used to simulate the physiological characteristics of the disease to generate a stimulation signal for deep brain stimulation; wherein the mathematical model includes arctangent operation and division operation, the arctangent operation and the division operation are used to simulate the hypersynchronous oscillation behavior of the basal ganglia in Parkinson's disease state; When the program instructions are executed by the processor, the device also performs the following operations in generating the stimulus signal: Based on the stimulation waveform and the EEG feedback signal, an input signal is generated for input into the mathematical model; Perform an arctangent operation on the input signal to obtain the arctangent result; as well as The result of the arctangent operation is divided to generate the stimulus signal; The arctangent operation includes: ； Where u represents the result of the arctangent operation, y represents the input signal of the arctangent operation, and h represents the dopamine parameter; The division operation includes: ； Where G(u) represents the stimulus signal, u represents the result of the arctangent operation, k represents the gain, and -b represents the open-loop heavy pole.

16. The device of claim 15, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations in generating control signals: Based on the preset initial signal and the EEG feedback signal generated after the previous stimulus, the current oscillation amplitude error is determined. Based on the current oscillation amplitude error, determine the current control signal increment; and The current control signal is determined based on the current control signal increment and the previous output control signal.

17. The device of claim 16, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations in determining the current oscillation amplitude error: The EEG feedback signal generated after the previous stimulus is converted from analog to digital to obtain the corresponding feedback digital signal; and The current oscillation amplitude error is determined based on the difference between the preset initial signal and the feedback digital signal.

18. The device of claim 16 or 17, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations in determining the current control signal increment: Based on the current oscillation amplitude error, determine the first characteristic value; The second characteristic value is determined based on the difference between the current oscillation amplitude error and the previous oscillation amplitude error; as well as The current control signal increment is determined based on the first feature value and the second feature value.

19. The device of claim 18, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations in determining the current control signal increment based on the first feature value and the second feature value: Normalize the first eigenvalue and the second eigenvalue to obtain a normalized result; and Based on the normalization result, the current control signal increment is determined.

20. The device of claim 19, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations during a normalization operation: Based on the previous oscillation amplitude error, the previous output control signal, and the feature values ​​determined when the control signal was generated, determine the normalized neuron weights corresponding to each feature value in the current control operation; and Based on the normalized neuron weights corresponding to the first and second eigenvalues, respectively, the first and second eigenvalues ​​are summed using a weighted average.

21. The device of claim 19, wherein when the program instructions are executed by the processor, the device further causes to calculate the current control signal increment in such a manner as follows: ； in, K represents the increment of the current control signal. N Indicates the calibration coefficient. This represents the normalization result. Represents the i-th eigenvalue. This represents the normalized neuron weights, i=1, 2.

22. The device of claim 20, wherein when the program instructions are executed by the processor, the device further causes to calculate the current control signal increment in such a manner as follows: ； in, K represents the increment of the current control signal. N Indicates the calibration coefficient. This represents the normalization result. Represents the i-th eigenvalue. This represents the normalized neuron weights, i=1, 2.

23. The device according to any one of claims 20-22, wherein when the program instructions are executed by the processor, the device further causes the normalized neuron weights to be calculated in the following manner: ； ； in, This represents the normalized neuron weights, where n represents time point n, and n-1 represents the previous time point n. Representing integral and proportional learning rates, This represents the weight coefficient corresponding to the i-th eigenvalue. This represents the weight coefficient corresponding to the i-th eigenvalue determined in the previous control operation. This indicates the error in the amplitude of the previous oscillation. This indicates the control signal output last time. This represents the i-th feature value determined when the control signal was generated last time, where i = 1 or 2.

24. The device according to any one of claims 16, 17, or 19-22, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations: The control signal from the previous output is obtained through a delay operation.

25. The device of claim 18, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations: The control signal from the previous output is obtained through a delay operation.

26. The device of claim 23, wherein when the program instructions are executed by the processor, the device further causes to perform the following operations: The control signal from the previous output is obtained through a delay operation.

27. A computer-readable storage medium having stored thereon computer-readable instructions that, when executed by one or more processors, implement the method performed by the device as described in any one of claims 15-26.

Images