An Adaptive DBS Parameter Adjustment Method and Device

Abstract

The present invention relates to an adaptive DBS parameter adjustment method and device, belonging to the field of biomedical signal processing technology. The method comprises: decoding real-time neural signals into real-time behavioral signals based on a behavioral decoding model; and adjusting DBS parameters based on the real-time behavioral signals. The behavioral decoding model is constructed based on the training results of a linear discrete state-space model, which is obtained by constructing a state equation based on neural signals, behavioral signals, and state signals, wherein the state signal is a potential state that simultaneously drives neural signals and behavioral signals. The present invention achieves DBS parameter adjustment through symptom markers while avoiding interference from normal behavior and reducing the complexity of regulation.

Description

Technical Field

[0001] The present invention relates to the technical field of biomedical signal processing, and in particular to a method and device for adaptive DBS parameter adjustment. Background Art

[0002] Deep brain stimulation (DBS) is a neuromodulation technology primarily used to treat intractable neurological and psychiatric disorders, such as Parkinson's disease, epilepsy, obsessive-compulsive disorder, and depression. Based on a deep understanding of brain function and neural circuits, DBS targets specific diseases by implanting electrodes in specific brain regions and applying precise electrical stimulation to regulate abnormal neural circuit activity and alleviate movement symptoms.

[0003] Despite significant progress in DBS technology, precise and robust adaptive DBS closed-loop control still faces significant challenges. The selection and acquisition of optimal neural markers is fundamental and prerequisite for achieving successful results with adaptive DBS closed-loop control. However, existing DBS technologies, for example, mostly target different disease types, using a single, generally accepted or widely used, empirically and statistically sound general indicator as a neural marker. This approach struggles to meet the robustness requirements for optimal neural markers required for precise adaptive DBS closed-loop control.

[0004] To circumvent the selection of optimal neural markers, there are currently options for directly measuring symptoms through peripheral sensors and adjusting stimulation parameters based on symptom markers. For example, inertial measurement units equipped with gyroscopes and accelerometers have been successfully used to detect and measure tremors, gait freezing, bradykinesia, and dyskinesia. However, a patient's normal behavior may interfere with symptom markers, leading to misjudgment. Furthermore, adding symptom measurement sensors and supporting data transmission and processing systems to a DBS system will increase the overall system complexity. Summary of the Invention

[0005] In view of this, it is necessary to provide an adaptive DBS parameter adjustment method and device to solve the problem that the existing solution of adjusting DBS stimulation parameters based on symptom markers is easily interfered with by the patient's normal behavior and has a high overall complexity.

[0006] In order to solve the above problems, in a first aspect, the present invention provides an adaptive DBS parameter adjustment method, comprising:

[0007] Decoding real-time neural signals into real-time behavioral signals based on a behavioral decoding model;

[0008] adjusting DBS parameters based on the real-time behavioral signal;

[0009] Among them, the behavioral decoding model is constructed based on the training results of the linear discrete state space model, and the linear discrete state space model is obtained by constructing a state equation based on neural signals, behavioral signals and state signals, and the state signal is a potential state that simultaneously drives the neural signals and behavioral signals.

[0010] In one possible implementation, the training process of the linear discrete state space model includes:

[0011] Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals;

[0012] The construction process of the behavior decoding model includes:

[0013] A recursive Kalman filter is constructed based on the solution result of the linear discrete state space model, and the recursive Kalman filter is used as the behavior decoding model.

[0014] In one possible implementation, the linear discrete state space model is expressed as:

[0015]

[0016]

[0017] in, Represents the time index, Represents neural signals, Indicates behavioral signals, Indicates status signal, and Represents Independent zero-mean white noise, Indicates interference behavior signal, represents the cross-correlation operation, for The autocovariance matrix of for The autocovariance matrix of for and The covariance matrix between represents the output matrix of the state signal, represents the output matrix of the neural signal, Output matrix representing the behavioral signal.

[0018] In one possible implementation, solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals includes:

[0019] The neural signal and behavioral signal at the target moment are projected onto the neural signal at the previous moment of the target moment, and a linear least squares solution of the orthogonal projection is performed to determine the linear discrete state space model. 、 、 、 、 、 .

[0020] In a possible implementation, the method further includes:

[0021] After completing the construction of the behavior decoding model, taking the historical neural signal as the input of the behavior decoding model, and calculating the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model;

[0022] The decoding effect of the behavior decoding model is evaluated based on the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model.

[0023] In a possible implementation, adjusting DBS parameters based on the real-time behavior signal includes:

[0024] determining a symptom marker value based on the real-time behavioral signal;

[0025] When the symptom marker value is not within a preset threshold range, the DBS parameter is adjusted based on the symptom marker value.

[0026] In a possible implementation, adjusting DBS parameters based on the symptom marker values includes:

[0027] Based on the symptom marker value, determining a stimulation parameter value corresponding to the symptom marker value in a stimulation parameter table, wherein the stimulation parameter table is determined based on clinical data, and the stimulation parameters include stimulation amplitude, stimulation frequency, and stimulation bandwidth;

[0028] The DBS parameters are adjusted according to the stimulation parameter values corresponding to the symptom marker values.

[0029] On the other hand, the present invention also provides an adaptive DBS parameter adjustment device, comprising:

[0030] A decoding module, configured to decode the real-time neural signal into a real-time behavioral signal based on a behavioral decoding model;

[0031] an adjustment module, configured to adjust DBS parameters based on the real-time behavioral signal;

[0032] Among them, the behavioral decoding model is constructed based on the training results of the linear discrete state space model, and the linear discrete state space model is obtained by constructing a state equation based on neural signals, behavioral signals and state signals, and the state signal is a potential state that simultaneously drives the neural signals and behavioral signals.

[0033] In a second aspect, the present invention further provides a DBS parameter adjustment device, comprising a memory and a processor, wherein:

[0034] The memory is used to store programs;

[0035] The processor is coupled to the memory and is configured to execute the program stored in the memory to implement the steps of the adaptive DBS parameter adjustment method described in any of the above implementations.

[0036] In a third aspect, the present invention further provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps of the adaptive DBS parameter adjustment method described in any of the above implementations.

[0037] The beneficial effects of the present invention are as follows: the adaptive DBS parameter adjustment method and device provided by the present invention decode neural signals into behavioral signals by constructing a behavioral decoding model, thereby avoiding the selection of neural markers, and instead adjusting DBS parameters through behavioral signals, thereby improving the reliability of DBS parameter adjustment. At the same time, behavioral signals are obtained by decoding neural signals, and there is no need to add additional behavioral signal acquisition equipment, thereby reducing the complexity of DBS parameter adjustment. Moreover, the behavioral signals obtained by decoding neural signals do not contain normal behavioral information, thereby avoiding interference with normal behavior. The present invention avoids interference with normal behavior while achieving DBS parameter adjustment through symptom markers, and reduces the complexity of regulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic flow chart of an embodiment of the adaptive DBS parameter adjustment method provided by the present invention;

[0039] Figure 2 A schematic diagram of an embodiment of the DBS regulation process provided by the present invention;

[0040] Figure 3 A schematic diagram of an embodiment of the process of constructing a behavior decoding model provided by the present invention;

[0041] Figure 4 A schematic flow chart of an embodiment of the DBS control strategy provided by the present invention;

[0042] Figure 5A schematic structural diagram of an embodiment of the adaptive DBS parameter adjustment device provided by the present invention;

[0043] Figure 6 This is a schematic structural diagram of an embodiment of the DBS parameter adjustment device provided by the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] In the description of the embodiments of the present invention, unless otherwise specified, "plurality" means two or more. "And / or" describes the association relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone.

[0046] The terms "first," "second," and so on, used in the embodiments of the present invention are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, technical features designated as "first" or "second" may explicitly or implicitly include at least one such feature.

[0047] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0048] Before presenting the embodiments, the following terms are explained.

[0049] DBS includes intracerebral stimulation electrodes, subcutaneous wires, pulse generators, and magnetic switches. The stimulation electrodes are inserted into specific brain nuclei through stereotactic technology and microelectrode recording technology, and continuous high-frequency pulse electrical stimulation is used to suppress abnormal discharges of brain nuclei to achieve therapeutic effects. DBS is mainly used to treat intractable neurological and psychiatric diseases such as Parkinson's disease, epilepsy, obsessive-compulsive disorder, and depression. DBS technology is based on a deep understanding of brain region functions and neural circuits. For specific diseases, it implants electrodes into specific areas of the brain and applies precise electrical stimulation to regulate abnormal neural circuit activity and relieve movement symptoms.

[0050] The present invention provides an adaptive DBS parameter adjustment method and device, which are described below.

[0051] Figure 1 A flow chart of an embodiment of the adaptive DBS parameter adjustment method provided by the present invention is shown as follows: Figure 1 As shown, the adaptive DBS parameter adjustment method includes:

[0052] S101. Decode the real-time neural signal into a real-time behavior signal based on a behavior decoding model.

[0053] It should be noted that real-time neural signals can be those from the brain region corresponding to the disease type. For example, in Parkinson's disease, the corresponding brain region may be the subthalamic nucleus, and the neural signals may be local field potential (LFP) signals. Real-time neural signals can be acquired using DBS electrodes implanted in the target patient's brain region corresponding to the disease type. A behavioral decoding model can be used to decode neural signals into behavioral signals, which can then be used as symptom markers to control DBS stimulation parameters, thus avoiding the need for selecting neural markers.

[0054] S102: Adjust DBS parameters based on the real-time behavior signal.

[0055] It should be noted that certain numerical values of the symptoms corresponding to the target patient's disease type can be determined through real-time behavioral signals. Taking Parkinson's disease as an example, the corresponding symptom marker may be tremor, and the symptom marker numerical value may be tremor frequency.

[0056] In some embodiments of the present invention, adjusting DBS parameters based on the real-time behavior signal includes:

[0057] determining a symptom marker value based on the real-time behavioral signal;

[0058] When the symptom marker value is not within a preset threshold range, the DBS parameter is adjusted based on the symptom marker value.

[0059] It should be noted that: after the symptom marker value is determined through real-time behavioral signals, it can be determined whether the target patient is in an abnormal state based on the preset threshold range, and then determine whether DBS parameter adjustment is needed. The preset threshold range can be set based on clinical data in the clinical stage, or it can be determined based on current disease-related research data. When the symptom marker value is not within the preset threshold range, it can be determined that the target patient is in an abnormal state. When it is determined that the target patient is in an abnormal state, the DBS parameters can be adjusted according to the symptom marker value. For example, the DBS can be adjusted through a comparison table of symptom marker values and stimulation parameters, or a regulatory model can be constructed through a neural network, and then the DBS is adjusted.

[0060] Among them, the behavioral decoding model is constructed based on the training results of the linear discrete state space model, and the linear discrete state space model is obtained by constructing a state equation based on neural signals, behavioral signals and state signals, and the state signal is a potential state that simultaneously drives the neural signals and behavioral signals.

[0061] In some embodiments of the present invention, Figure 2 As shown, the behavioral decoding model training process can be carried out during the target patient's clinical follow-up debugging phase, and the DBS adaptive parameter adjustment process can be carried out during the target patient's daily use phase. That is, after completing clinical follow-up debugging, the patient does not need to wear peripheral sensors to measure symptoms in real time. In addition, because normal behavior is not reflected in the neural activity of the brain area corresponding to disease symptoms, the disease symptom behavior obtained through neural signal decoding will not contain normal behavior information, thus avoiding interference from normal behavior.

[0062] In summary, the adaptive DBS parameter adjustment method provided by the embodiment of the present invention decodes neural signals into behavioral signals by constructing a behavioral decoding model, thereby avoiding the selection of neural markers. Instead, DBS parameters are adjusted through behavioral signals, thereby improving the reliability of DBS parameter adjustment. At the same time, behavioral signals are obtained by decoding neural signals, and there is no need to add additional behavioral signal acquisition equipment, thereby reducing the complexity of DBS parameter adjustment. Moreover, the behavioral signals obtained by decoding neural signals do not contain normal behavioral information, thereby avoiding interference with normal behavior. The present invention avoids interference with normal behavior while realizing DBS parameter adjustment through symptom markers, and reduces the complexity of regulation.

[0063] In some embodiments of the present invention, Figure 3 As shown, the behavior decoding model construction process provided by the present invention includes:

[0064] 1. Data preprocessing: Process the LFP signals collected by the DBS electrodes into data that is more relevant to behavior and suitable for model training. The specific process of data preprocessing includes:

[0065] (1) Frequency band decomposition: The collected LFP signal can be decomposed into frequency bands based on the frequency band information of the common neural markers of the disease to obtain the energy information within each frequency band. For example, for Parkinson's disease, the main relevant frequency bands are the high and low beta bands (low beta: 13-20Hz, high beta: 20-30Hz) and the gamma band (gamma: 30-90Hz). The energy information can be obtained by filtering and Hilbert transform or short-time Fourier transform. The energy value obtained can be optionally logarithmized to change its numerical distribution.

[0066] (2) Data / time alignment: Neural signals and behavioral signals are collected by two systems that are synchronized in time but have different sampling rates. Before combining the two signals for model training, necessary interpolation and other operations are required to align the data and time labels and unify the time dimension for subsequent model training.

[0067] (3) Standardization: Neural signals and behavioral signals usually have different dimensions. After normalization using the zscore method, signals of different dimensions are transformed to the same scale, reducing dimensionality deviation during model training.

[0068] (4) Downsampling: The sampling rate of neural signals (e.g., 1kHz) is usually higher than the sampling rate of behavioral signals (e.g., 20Hz). Neural signals can usually be downsampled to the sampling rate of behavioral signals, which has a certain smoothing effect on the data and can also reduce the amount of computation.

[0069] 2. Training and test set division: The neural data and behavioral data collected during the clinical follow-up debugging phase are divided into training and test sets in a ratio (for example, 8:2). The training set is used to determine the model structure and model parameters, and the test set is used to preliminarily confirm the model effect.

[0070] 3. Model hyperparameter debugging: The dimensions of the latent state and the dimension of the behavior-related latent state are two hyperparameters that need to be debugged. They can be obtained using the grid search method based on the training set data.

[0071] 4. Model Training: Model parameter identification utilizes a prioritized subspace identification method with a two-stage identification mechanism, prioritizing the extraction of latent states related to behavior. The training process involves only matrix operations, primarily computing linear least squares solutions for orthogonal projections and singular value decomposition. During the identification process, future measurable data, such as future behavior and future neural activity, are projected onto past neural activity, limiting the extraction of only latent states present in neural activity. Normal behavior (interference) is not reflected in the neural activity of brain regions corresponding to disease symptoms, and therefore, the behavioral signals decoded from this neural activity do not contain normal behavioral information.

[0072] 5. Behavior decoding: After model training is completed, a recursive Kalman filter is constructed based on the identified model, and the input neural signal is used to decode, predict and output the behavior.

[0073] In some embodiments of the present invention, the training process of the linear discrete state space model includes:

[0074] Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals;

[0075] The construction process of the behavior decoding model includes:

[0076] A recursive Kalman filter is constructed based on the solution result of the linear discrete state space model, and the recursive Kalman filter is used as the behavior decoding model.

[0077] It should be noted that the preprocessing process of historical neural signals and historical behavioral signals can refer to the data preprocessing process described above. When solving the linear discrete state-space model, the preprocessed historical neural signals and historical behavioral signals can be used as known quantities to solve the linear discrete state-space model, thereby determining multiple parameter matrices of the linear discrete state-space model. Subsequently, a recursive Kalman filter can be constructed based on the multiple parameter matrices of the linear discrete state-space model, and the recursive Kalman filter can be used as the behavioral decoding model. By constructing a behavioral decoding model, neural activity can be decoded into behavioral activity while eliminating the interference of normal behavior.

[0078] In some embodiments of the present invention, the linear discrete state space model is expressed as:

[0079]

[0080]

[0081] in, Represents the time index, Represents neural signals, Indicates behavioral signals, Indicates status signal, and Represents Independent zero-mean white noise, Indicates interference behavior signal, represents the cross-correlation operation, for The autocovariance matrix of for The autocovariance matrix of for and The covariance matrix between represents the output matrix of the state signal, represents the output matrix of the neural signal, Output matrix representing the behavioral signal.

[0082] In some embodiments of the present invention, solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals includes:

[0083] The neural signal and behavioral signal at the target moment are projected onto the neural signal at the previous moment of the target moment, and a linear least squares solution of the orthogonal projection is performed to determine the linear discrete state space model. 、 、 、 、 、 .

[0084] It should be noted that when the pre-processed historical neural signals and historical behavioral signals are used as known quantities to solve the linear discrete state space model, the neural signals and behavioral signals at the target moment can be projected onto the neural signals at the previous moment of the target moment, and a linear least squares solution of orthogonal projection can be performed. At the same time, a singular value decomposition can be performed to determine the linear discrete state space model. 、 、 、 、 、 The target moment can be any sampling moment of the preprocessed historical neural signals and historical behavioral signals.

[0085] 6. Decoding effect evaluation: Pearson linear correlation coefficient is usually used to evaluate decoding effect.

[0086] In some embodiments of the present invention, in order to ensure the reliability of the behavior decoding model, the method further includes:

[0087] After completing the construction of the behavior decoding model, taking the historical neural signal as the input of the behavior decoding model, and calculating the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model;

[0088] The decoding effect of the behavior decoding model is evaluated based on the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model.

[0089] It should be noted that the value range of the Pearson linear correlation coefficient is -1 to 1, where -1 indicates a complete negative correlation, 1 indicates a complete positive correlation, and 0 indicates no correlation. When evaluating the decoding effect of the behavior decoding model, it is possible to determine whether the behavior decoding model can achieve the expected decoding effect based on actual needs. For example, a correlation coefficient threshold can be set to determine whether the behavior decoding model can achieve the expected decoding effect. When the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model is within the correlation coefficient threshold, it can be considered that the behavior decoding model can achieve the expected decoding effect. Otherwise, the parameters of the behavior decoding model can be adjusted until it can achieve the expected decoding effect.

[0090] In some embodiments of the present invention, Figure 4 Figure 2 shows the process of adaptive DBS closed-loop regulation using decoded disease symptom markers.

[0091] The adaptive mode can be started by timing or manually. The exit method is to exit after the preset condition is met. The preset condition can be that the number of normal collections reaches a certain value, for example, 3 times.

[0092] The parameters that need to be initialized during parameter initialization are: abnormal thresholds of symptom markers (amplitude and duration thresholds), initial effective stimulation parameters, and stimulation parameter safety boundary values.

[0093] The present invention does not impose any restrictions on specific setting values such as the adaptive timing start time interval, the acquisition and recording refractory period, and the neural signal acquisition duration, and these values can be adjusted and set according to actual conditions.

[0094] In some embodiments of the present invention, adjusting DBS parameters based on the symptom marker values includes:

[0095] Based on the symptom marker value, determining a stimulation parameter value corresponding to the symptom marker value in a stimulation parameter table, wherein the stimulation parameter table is determined based on clinical data, and the stimulation parameters include stimulation amplitude, stimulation frequency, and stimulation bandwidth;

[0096] The DBS parameters are adjusted according to the stimulation parameter values corresponding to the symptom marker values.

[0097] It should be noted that a relatively simple method for adjusting stimulation parameters is the table lookup method, in which different stimulation parameter groups are selected according to the amplitude of the decoded symptom markers. This method has a small amount of computational complexity. The stimulation amplitude parameters can be adjusted in combination with the symptom amplitude and the bilateral threshold mechanism. The adjustment of the stimulation amplitude can be determined according to the standard adaptation mechanism of aDBS. The stimulation frequency and pulse width can be combined with control algorithms such as Proportional Integral Derivative (PID) to fine-tune according to changes in the symptom marker monitoring variables within a certain safety margin around the initial value.

[0098] The method provided by the present invention uses clean disease symptom markers obtained through neural activity decoding for adaptive DBS closed-loop regulation. Compared with the existing adaptive DBS closed-loop regulation technology based on neural markers, it avoids the difficulty of selecting and obtaining the best neural markers; compared with the existing adaptive DBS closed-loop regulation technology based on symptom markers, it not only utilizes neural activity information but also can effectively avoid the interference of patients' daily behaviors on symptom markers, while not increasing the complexity of the overall system during the daily use stage of the product.

[0099] Compared with other mathematical modeling methods based on deep learning, the mathematical modeling process of the method provided by the present invention only involves matrix algebraic operations and singular value decomposition, and the computational complexity is relatively low; the model is small and has good real-time performance in practical applications; a priority subspace identification method with a two-stage identification mechanism is adopted to preferentially extract behavior-related latent states, effectively reducing the influence of latent states in neural activities that are not related to the behavior of interest, so that the extracted symptom markers do not contain interference from normal behavior, eliminating the overhead of the artifact interference removal step.

[0100] The present invention uses clean symptom markers as control quantities for adaptive DBS closed-loop regulation. The relevant parameters of the markers, such as amplitude abnormality threshold, have more intuitive physical meanings, and clinicians can give more accurate recommended values.

[0101] In order to better implement the adaptive DBS parameter adjustment method in the embodiment of the present invention, based on the adaptive DBS parameter adjustment method, correspondingly, Figure 5 As shown, an embodiment of the present invention further provides an adaptive DBS parameter adjustment device, and the adaptive DBS parameter adjustment device 500 includes:

[0102] A decoding module 501 is configured to use the real-time neural signal of the target patient as the input of a behavior decoding model, and use the output of the behavior decoding model as the symptom marker value of the target patient;

[0103] An adjustment module 502 is configured to adaptively adjust the DBS of the target patient based on the symptom marker values of the target patient;

[0104] The behavior decoding model is trained based on the historical neural signals and historical behavior signals of the target patient collected during the clinical stage.

[0105] The adaptive DBS parameter adjustment device 500 provided in the above embodiment can implement the technical solution described in the above embodiment of the adaptive DBS parameter adjustment method. The specific implementation principles of the above modules or units can be found in the corresponding content of the above embodiment of the adaptive DBS parameter adjustment method, which will not be repeated here.

[0106] like Figure 6 As shown, the present invention also provides a DBS parameter adjustment device 600. The DBS parameter adjustment device 600 includes a processor 601, a memory 602 and a display 603. Figure 6 Only some components of the DBS parameter adjustment device 600 are shown, but it should be understood that it is not required to implement all of the shown components, and more or fewer components may be implemented instead.

[0107] In some embodiments, the processor 601 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 602 , such as the magnetic resonance image optimization method of the present invention.

[0108] In some embodiments, processor 601 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, processor 601 may be local or remote. In some embodiments, processor 601 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, multiple clouds, or any combination thereof.

[0109] In some embodiments, the memory 602 may be an internal storage unit of the DBS parameter adjustment device 600, such as a hard disk or memory of the DBS parameter adjustment device 600. In other embodiments, the memory 602 may also be an external storage device of the DBS parameter adjustment device 600, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a flash card, etc., equipped on the DBS parameter adjustment device 600.

[0110] Furthermore, the memory 602 may include both an internal storage unit of the DBS parameter adjustment device 600 and an external storage device. The memory 602 is used to store application software for installing the DBS parameter adjustment device 600 and various data.

[0111] In some embodiments, the display 603 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an organic light-emitting diode (OLED) touchscreen. The display 603 is used to display information on the DBS parameter adjustment device 600 and to display a visual user interface. Components 601-603 of the DBS parameter adjustment device 600 communicate with each other via a system bus.

[0112] In one embodiment, when the processor 601 executes the adaptive DBS parameter adjustment program in the memory 602, the following steps may be implemented:

[0113] Decoding real-time neural signals into real-time behavioral signals based on a behavioral decoding model;

[0114] adjusting DBS parameters based on the real-time behavioral signal;

[0115] Among them, the behavioral decoding model is constructed based on the training results of the linear discrete state space model, and the linear discrete state space model is obtained by constructing a state equation based on neural signals, behavioral signals and state signals, and the state signal is a potential state that simultaneously drives the neural signals and behavioral signals.

[0116] It should be understood that, when the processor 601 executes the adaptive DBS parameter adjustment program in the memory 602 , in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0117] Furthermore, the embodiments of the present invention do not specifically limit the type of DBS parameter adjustment device 600. The DBS parameter adjustment device 600 may be a portable electronic device such as a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, or laptop computer. Exemplary portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in other embodiments of the present invention, the DBS parameter adjustment device 600 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0118] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the adaptive DBS parameter adjustment method provided in the above-mentioned method embodiments can be implemented.

[0119] Those skilled in the art will appreciate that all or part of the process flow of the above-described method embodiment can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0120] The above describes in detail the adaptive DBS parameter adjustment method and apparatus provided by the present invention. Specific examples are used herein to illustrate the principles and implementations of the present invention. The description of the above embodiments is intended only to facilitate understanding of the method and core concepts of the present invention. Furthermore, those skilled in the art will appreciate that variations in the specific implementations and scope of application are possible based on the concepts of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. An adaptive DBS parameter adjustment device, characterized in that: include: A decoding module, configured to decode the real-time neural signal into a real-time behavioral signal based on a behavioral decoding model; an adjustment module, configured to adjust DBS parameters based on the real-time behavior signal; The behavior decoding model is constructed based on the training results of a linear discrete state space model, which is obtained by constructing a state equation based on neural signals, behavioral signals, and state signals. The state signal is the potential state that drives both neural signals and behavioral signals. The training process of the linear discrete state space model includes: Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals; The construction process of the behavior decoding model includes: Constructing a recursive Kalman filter based on the solution result of the linear discrete state space model, and using the recursive Kalman filter as the behavior decoding model; The expression of the linear discrete state space model is: in, Represents the time index, Represents neural signals, Indicates behavioral signals, Indicates status signal, and Represents Independent zero-mean white noise, Indicates interference behavior signal, represents the cross-correlation operation, for The autocovariance matrix of for The autocovariance matrix of for and The covariance matrix between represents the output matrix of the state signal, represents the output matrix of the neural signal, The output matrix representing the behavioral signal; Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals includes: The neural signal and behavioral signal at the target moment are projected onto the neural signal at the previous moment of the target moment, and a linear least squares solution of the orthogonal projection is performed to determine the linear discrete state space model. 、 、 、 、 、 .

2. A DBS parameter adjustment device, characterized in that: comprising a memory and a processor, wherein, The memory is used to store programs; The processor is coupled to the memory and configured to execute the program stored in the memory to implement an adaptive DBS parameter adjustment method, the adaptive DBS parameter adjustment method comprising: Decoding the real-time neural signal into a real-time behavioral signal based on a behavioral decoding model; and adjusting DBS parameters based on the real-time behavioral signal. The behavior decoding model is constructed based on the training results of a linear discrete state space model, which is obtained by constructing a state equation based on neural signals, behavioral signals, and state signals. The state signal is the potential state that drives both neural signals and behavioral signals. The training process of the linear discrete state space model includes: Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals; The construction process of the behavior decoding model includes: Constructing a recursive Kalman filter based on the solution result of the linear discrete state space model, and using the recursive Kalman filter as the behavior decoding model; The expression of the linear discrete state space model is: in, Represents the time index, Represents neural signals, Indicates behavioral signals, Indicates status signal, and Represents Independent zero-mean white noise, Indicates interference behavior signal, represents the cross-correlation operation, for The autocovariance matrix of for The autocovariance matrix of for and The covariance matrix between represents the output matrix of the state signal, represents the output matrix of the neural signal, The output matrix representing the behavioral signal; Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals includes: The neural signal and behavioral signal at the target moment are projected onto the neural signal at the previous moment of the target moment, and a linear least squares solution of the orthogonal projection is performed to determine the linear discrete state space model. 、 、 、 、 、 .

3. The DBS parameter adjustment device according to claim 2, characterized in that: The adaptive DBS parameter adjustment method further includes: After completing the construction of the behavior decoding model, taking the historical neural signal as the input of the behavior decoding model, and calculating the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model; The decoding effect of the behavior decoding model is evaluated based on the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model.

4. The DBS parameter adjustment device according to claim 2, characterized in that: The adjusting of DBS parameters based on the real-time behavior signal includes: determining a symptom marker value based on the real-time behavioral signal; When the symptom marker value is not within a preset threshold range, the DBS parameter is adjusted based on the symptom marker value.

5. The DBS parameter adjustment device according to claim 4, characterized in that: The adjusting of DBS parameters based on the symptom marker values includes: Based on the symptom marker value, determining a stimulation parameter value corresponding to the symptom marker value in a stimulation parameter table, wherein the stimulation parameter table is determined based on clinical data, and the stimulation parameters include stimulation amplitude, stimulation frequency, and stimulation bandwidth; The DBS parameters are adjusted according to the stimulation parameter values corresponding to the symptom marker values.

6. A computer-readable storage medium, characterized in that Used to store a computer-readable program or instruction, wherein when the program or instruction is executed by a processor, an adaptive DBS parameter adjustment method can be implemented, wherein the adaptive DBS parameter adjustment method includes: Decoding the real-time neural signal into a real-time behavioral signal based on a behavioral decoding model; and adjusting DBS parameters based on the real-time behavioral signal. The behavior decoding model is constructed based on the training results of a linear discrete state space model, which is obtained by constructing a state equation based on neural signals, behavioral signals, and state signals. The state signal is the potential state that drives both neural signals and behavioral signals. The training process of the linear discrete state space model includes: Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals; The construction process of the behavior decoding model includes: Constructing a recursive Kalman filter based on the solution result of the linear discrete state space model, and using the recursive Kalman filter as the behavior decoding model; The expression of the linear discrete state space model is: in, Represents the time index, Represents neural signals, Indicates behavioral signals, Indicates status signal, and Represents Independent zero-mean white noise, Indicates interference behavior signal, represents the cross-correlation operation, for The autocovariance matrix of for The autocovariance matrix of for and The covariance matrix between represents the output matrix of the state signal, represents the output matrix of the neural signal, The output matrix representing the behavioral signal; Solving the linear discrete state space model based on the preprocessed historical neural signals and historical behavioral signals includes: The neural signal and behavioral signal at the target moment are projected onto the neural signal at the previous moment of the target moment, and a linear least squares solution of the orthogonal projection is performed to determine the linear discrete state space model. 、 、 、 、 、 .

7. The computer-readable storage medium according to claim 6, wherein: The adaptive DBS parameter adjustment method further includes: After completing the construction of the behavior decoding model, taking the historical neural signal as the input of the behavior decoding model, and calculating the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model; The decoding effect of the behavior decoding model is evaluated based on the Pearson linear correlation coefficient between the historical behavior signal and the output of the behavior decoding model.

8. The computer-readable storage medium according to claim 6, wherein: The adjusting of DBS parameters based on the real-time behavior signal includes: determining a symptom marker value based on the real-time behavioral signal; When the symptom marker value is not within a preset threshold range, the DBS parameter is adjusted based on the symptom marker value.

9. The computer-readable storage medium according to claim 8, wherein: The adjusting of DBS parameters based on the symptom marker values includes: Based on the symptom marker value, determining a stimulation parameter value corresponding to the symptom marker value in a stimulation parameter table, wherein the stimulation parameter table is determined based on clinical data, and the stimulation parameters include stimulation amplitude, stimulation frequency, and stimulation bandwidth; The DBS parameters are adjusted according to the stimulation parameter values corresponding to the symptom marker values.

Images