Tourette syndrome treatment system based on multi-band median nerve regulation

Abstract

The invention belongs to the technical field of intelligent medical treatment, and particularly relates to a Tourette syndrome treatment system based on multi-band median nerve regulation. The system comprises an electrical stimulator system, a software control system and a face twitch recognition module. The electrical stimulator system comprises a main control module and a nerve stimulation chip; the software control system comprises TS intervention treatment upper computer software and mobile terminal control software; the face twitch recognition module comprises a face calibration point information extraction sub-module, a calibration point data processing and feature extraction sub-module, a machine learning sub-module and a transfer learning classification sub-module; the portable design is adopted, the nerve stimulation chip is controlled by the main control chip to output adjustable pulse current, and the median nerve is stimulated in combination with a frequency band cross sequence. Clinical tests show that the twitch symptom of a patient under 20Hz stimulation is averagely improved by 37%, the endurance of equipment exceeds 8 hours, home non-invasive treatment is supported, the treatment convenience and safety are remarkably improved, and an effective solution is provided for TS patients with insufficient curative effects of traditional medicines.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent medicine, and particularly relates to a treatment system for Tourette syndrome. Background Art

[0002] Tourette syndrome (TS) is a neurodevelopmental disorder characterized by sudden, repetitive, involuntary, and non-rhythmic movements or vocalizations [1][2]. Common motor tics include eye blinking, involuntary eye movement, mouth opening, head twisting, shoulder shrugging, or making inappropriate gestures. Simple vocal tics are manifested as unconscious sounds emitted from the larynx or nasal cavity, such as throat clearing, sniffing, coughing, or repeating words or phrases heard [3][4][5]. These behaviors often have a social impact and increase the psychological burden of patients. In addition, TS is usually accompanied by other mental and behavioral disorders such as obsessive-compulsive disorder, attention deficit hyperactivity disorder, autism, anxiety disorder, and depression [6]. Approximately 0.3-0.9% of school-age children (4-18 years old) [7] and 0.002-0.08% of adults [8] are diagnosed with Tourette syndrome.

[0003] Currently, the treatment for TS mainly includes drug therapy, behavioral therapy, invasive brain stimulation, and non-invasive brain stimulation therapy. Drugs such as risperidone and clonidine are used to inhibit tic symptoms by affecting the abnormal activities of the striatum and basal ganglia [9]. They are easy to use, but may have significant side effects and drug resistance. Behavioral therapies such as habit reversal therapy and comprehensive behavioral intervention for tics have fewer side effects, but there are fewer experts providing this treatment. For patients with severe symptoms or ineffective drug treatment, invasive brain stimulation such as deep brain stimulation (DBS) becomes an alternative therapy. Currently, more than 300 patients with Tourette syndrome worldwide have received deep brain stimulation

[10] , but this treatment method requires surgery and has the risk of postoperative complications.

[0004] In recent years, the application of non-invasive brain stimulation technology in the treatment of TS has gradually received more attention. Research has shown that the oscillations of the human brain, especially in the α (8-12 Hz), β (13-20 Hz), and γ (20-100 Hz) frequency bands, are closely related to the processes of movement execution and termination. During the processes of movement execution and termination, enhancements or attenuations in amplitude can be observed in the α, β, and γ frequency bands. Among them, the amplitudes of the α and β rhythms decrease before movement execution and resynchronize after termination, while the γ rhythm remains enhanced during execution and termination

[11] . Moreover, the changes in the α and β frequency bands are closely related to the tic symptoms of TS patients

[12] . Transcranial magnetic stimulation in the α frequency band

[13] and transcranial alternating current stimulation in the β frequency band

[14] have both been proven to be able to regulate the local activity of the motor cortex by entraining cortical oscillations, thereby suppressing tics. However, this treatment method is difficult to precisely locate and is costly, and the device is relatively large in size and cannot be used as a daily home treatment means. In addition, peripheral nerve electrical stimulation, as a new treatment method for TS, has received increasing attention. Morera Maiquez et al.

[15] proposed that median nerve stimulation (MNS) at 12 Hz can induce the synchronization of contralateral electroencephalogram activity in the primary sensorimotor cortex, and they also demonstrated in a subsequent study using magnetoencephalography that 10 Hz and 20 Hz MNS have similar effects on cortical activity

[16] . In a randomized double-blind controlled trial, 19 TS patients received 10 Hz MNS intervention, and the results showed that the tic count and severity decreased significantly during the stimulation. In addition, the therapeutic effect persisted after the MNS was turned off. Subsequently, Morera Maiquez et al. conducted a double-blind controlled trial to evaluate the effect of home neuromodulation treatment for tics in 132 patients. The results showed that after 4 weeks of 10 Hz MNS intervention for 10 minutes per day, the tic severity in the active stimulation group decreased by 35%, while the decreases in the sham stimulation and control groups were only about 10%

[17] . Ann M. Iverson et al. demonstrated through a home double-blind experiment on 27 patients that MNS can improve the tic frequency and intensity

[18] . These studies indicate that home rhythmic MNS therapy using portable wearable devices may develop into an efficient and convenient treatment method for TS. It not only improves the treatment accessibility for patients but also provides a non-drug and sustainable treatment option for TS patients, and is expected to play an important role in future home and community medical care. However, the electrical stimulation treatments in these studies mainly focus on the α frequency band, the frequency selection for median nerve stimulation in the treatment of TS is relatively single, and it does not take into account the individual differences of patients for personalized adjustment. In addition, although median nerve electrical stimulation has initially shown a remission effect on TS symptoms in clinical studies, the existing devices still have the following problems in design:

[0005] Large volume and poor portability: Most traditional electrical stimulator devices are bulky. For example, the Digitimer DS7A stimulator (225mm×100mm×255mm) used by Morera of the University of Nottingham

[17]

[19] and the TENS7000 stimulator (198mm×223.52mm×50.8mm) used by Iverson of the University of Washington

[18]

[20] ; this not only increases the burden on patients to carry and use, but also limits the application of the device in daily life scenarios.

[0006] Single output mode, difficult to meet personalized sequence requirements: Existing commercial devices mostly adopt a single-frequency or amplitude stimulation mode, which is difficult to flexibly adjust according to the individual needs of patients, restricting its adaptability to diverse pathological conditions. And the stimulation devices used in the above-mentioned laboratories require certain professional backgrounds and programming abilities if they are to output specific sequences;

[0007] Complex operation and poor linkage: Most devices require the assistance of professionals to operate, and it is difficult for ordinary patients or family members to complete the settings and use independently. Moreover, it is very difficult to synchronize with other devices, making it difficult to achieve closed-loop control.

[0008] The above problems directly affect the popularization and use effect of median nerve electrical stimulation technology in the treatment of TS. Therefore, developing a portable and flexibly adjustable median nerve stimulator is of great significance for exploring better stimulation parameters and improving the treatment effect of TS.

[0009] Moreover, the treatment devices used are large in volume or have poor controllability, and are not suitable for home treatment. Summary of the Invention

[0010] The purpose of the present invention is to provide a Tourette syndrome treatment system based on multi-band median nerve regulation that is suitable for personalized home treatment, easy to use, and has excellent curative effects.

[0011] The Tourette syndrome treatment system based on multi-band median nerve regulation provided by the present invention conducts home treatment after finding the best treatment parameters through multi-band stimulation sequence testing and facial tic recognition and assessment; it specifically includes three parts: an electrical stimulator system, a software control system, and a facial tic recognition module; among them:

[0012] (1) The main control module, which is used to receive control instructions and generate stimulation parameters;

[0013] The electrical stimulator system is shown as follows; including: Figure 1 as shown;

[0014] (1) The main control module, which is used to receive control instructions and generate stimulation parameters;

[0015] (2) The nerve stimulation module (a nerve stimulation chip), communicates with the main control module through a serial port, and is used to generate a bidirectional pulsed square wave signal with adjustable frequency, amplitude, and pulse width; that is, it can support a variety of stimulation parameters, including the configuration of frequency, amplitude, and pulse width, with high control precision and a maximum error of no more than 2%.

[0016] (3) The power supply module, which includes a 500 mA lithium battery and a charging management circuit, supports the system to work continuously for 8 hours;

[0017] (4) The electrode interface module, connects non-woven gel electrodes through a circular headphone jack and outputs stimulation signals;

[0018] (5) The status indication module, displays the system status through a programmable LED;

[0019] (6) A customized 3D printed shell, with an overall size of 67.5 mm × 48.7 mm × 23.6 mm, which is convenient for daily use;

[0020] The host computer on the computer side or the software on the mobile phone side communicates with the main control module through the Bluetooth module and sends stimulation parameters and control instructions; among them:

[0021] The main control module is responsible for processing system control, Bluetooth communication, and instruction interaction with the host computer, and specifically completes the following functions:

[0022] (1) Communicates with the host computer or mobile phone APP through Bluetooth, receives stimulation parameters and control instructions, and executes corresponding control operations;

[0023] (2) Exchanges data with the nerve stimulation chip through a serial port, accurately controls the generation of electrical stimulation signals, and realizes the parsing and real-time control of user-set parameters such as frequency, amplitude, and pulse width;

[0024] (3) Communicates with the programmable LED lamp through the I 2 C protocol, and displays the system status in real time, facilitating users to view the operation of the device;

[0025] (4) Realizes time synchronization with external devices through the GPIO port, and supports a two-way trigger mode. It can be pre-configured to receive a high-level trigger signal from an external device at the start of the experiment to start the system operation, or it can be configured to send a high-level signal to an external device at the start of the experiment.

[0026] Furthermore, the main control module supports a TYPE-C to TTL debugging interface, and the device can connect two non-woven gel electrodes through a circular electrode connection socket to realize the output of electrical stimulation.

[0027] To ensure the beauty and portability of the stimulator, a customized 3D printed shell is equipped to ensure the stability of the device, such as Figure 2As shown. The overall size of the device is 67.5mm×48.7mm×23.6mm, which is convenient for carrying. The physical connection diagram is as Figure 3 shown.

[0028] The output parameters of the described electrical stimulator system satisfy:

[0029] (1) Pulse frequency range: 1 - 500Hz, minimum adjustment step size 0.5Hz, error ≤ 2%;

[0030] (2) Current amplitude range: 0.5 - 20mA, minimum adjustment step size 0.1mA, error ≤ 2%;

[0031] (3) Pulse width range: 10 - 2000μs, minimum adjustment step size 100μs, error ≤ 3%;

[0032] (4) Pulse energy range: 1.25×10 -8 J to 1×10 -4 J;

[0033] (5) DC component ≤ 0.01mA.

[0034] (2) Software control system. It includes:

[0035] The described software control system includes: the upper computer software for TS intervention treatment, the mobile control software; among which:

[0036] (1) The upper computer software for TS intervention treatment;

[0037] That is, the upper computer software system of the PC side, the stimulator human - machine interaction interface (GUI), which includes a Bluetooth connection module, a parameter setting area, and a stimulation sequence control area, as Figure 4 shown.

[0038] The Bluetooth module supports Bluetooth 5.2 dual - mode communication, can quickly scan nearby Bluetooth devices and display all connectable electrical stimulator devices. Users can select the target electrical stimulator from the device list. After establishing a connection, the software will monitor the connection status in real - time to ensure a stable connection. When the connection is disconnected, the system will pop up a prompt, allowing users to reconnect or select other devices.

[0039] The parameter setting area is used to set the output parameters of the electrical stimulator, including frequency, pulse width, amplitude, etc. Users can adjust these parameters through simple input or selection boxes. After the parameter values are adjusted, the software sends them to the electrical stimulator via Bluetooth to control its output in real - time.

[0040] Stimulation sequence control area. The user can edit and manage the stimulation sequence in the stimulation sequence control area. By setting the start and end times of the stimulation, the frequency and intensity of the stimulation, etc., the user can design a complete stimulation sequence. After finding the threshold of median nerve stimulation, the parameters can be saved and edited and adjusted on the interface. During the execution of the stimulation sequence, the status of the stimulation, the progress of the experiment, and the remaining stimulation time can be displayed in real time, which is used to monitor the experimental process and ensure the smooth progress of the experiment. The interaction system between the stimulator and the host computer software is as Figure 6 shown.

[0041] (2) Mobile control software

[0042] That is, the mobile APP provides preset treatment modes and manual modes, and is packaged as an APK file through HBuilderX;

[0043] The mobile control software supports direct installation and use on mobile phones. It can quickly scan nearby Bluetooth devices and display connectable electrical stimulation devices. The target electrical stimulator can be selected from the device list. After establishing a connection, enter the parameter setting interface, as Figure 6 shown. The amplitude of the stimulation current can be set, and there are two modes: manual switch and start treatment according to the preset mode, both of which can manually interrupt the stimulation. The software will monitor the connection status in real time to ensure a stable connection. When the connection is disconnected, the system will pop up a prompt to allow the user to reconnect or select other devices. The software supports two treatment modes: manual mode and preset treatment mode. In the manual mode, the user can flexibly control the start and stop of the stimulation. Different from the host computer on the computer side, the main application scenario of the mobile side is for patients to use at home. To ensure that the stimulation mode and stimulation duration are the same each time, the preset treatment mode is set, and the stimulation is automatically performed according to the treatment plan preset by the software. Other interaction operations with the electrical stimulator system are the same as Figure 5 consistent, and the user can also interrupt the stimulation at any time. The software ensures a stable connection by monitoring the connection status in real time. When the connection is disconnected, the system will pop up a prompt to allow the user to reconnect or select other devices. For the convenience of patients using at home, the software is packaged as an APK file, and the user can directly install and use it on mobile devices such as mobile phones or tablets.

[0044] (3) Facial tic recognition module

[0045] The efficacy of median nerve intervention is compared through certain evaluation methods. The present invention uses facial tic recognition to evaluate the efficacy under median nerve intervention with different parameters. Specifically, it includes three sub-modules: facial landmark information extraction, landmark data processing and feature extraction, and machine learning and transfer learning classification; among them:

[0046] (1) Facial landmark information extraction sub-module;

[0047] First, the recorded facial twitch video of the patient is segmented into 1s video windows, and then MedianPipeFaceMesh

[21] is selected to extract the landmark points of the facial model. Since the number of its landmark points is too large, and only some of these landmark points contribute significantly to classification, therefore, the present invention selects 222 key landmark points therefrom, specifically including 80 for the eyes, 22 for the eyebrows, 49 for the nose, 35 for the mouth, and 36 for the facial contour, to construct the basic features of the classification model. This includes 170 points used by Brügge

[22] , as well as points that are commonly considered important by doctors and researchers, and can extract facial landmark points in real time under 60 frames of video.

[0048] (2) Landmark point data processing and feature extraction sub-module;

[0049] To process these landmark points, this method organizes the data into a four-dimensional data block (G, L, T, P), where G represents the number of window partitions, L represents the number of landmark points in each window, T represents the number of frames in each window, and P represents the spatial coordinates of the landmark points. Since the original landmarks detected by MediaPipe FaceMesh may have noise, the generated coordinate trajectories may show unnatural jitter. To minimize the impact of encoding noise, this paper uses a digital filter to perform low-pass filtering on the 222 landmark point trajectory sequences. Finally, the trajectory features of the landmark points are extracted from the filtered data. For each data window, the j-th landmark point of the i-th window is extracted to form a 30×3 trajectory matrix. Based on the trajectory matrix, four features of the maximum Euclidean distance, coordinate variance, average value and maximum value of absolute spatial drift are calculated for each set of landmark point coordinate sets.

[51] :

[0050] The maximum Euclidean distance D max :

[0051] The Euclidean distance is the shortest straight-line distance between two points in space. For the coordinates of a landmark point in two consecutive frames being p1: (x1, y1, z1) and p2: (x2, y2, z2) respectively, the Euclidean distance of this landmark point in consecutive frames is:

[0052]

[0053] D max is to calculate the maximum Euclidean distance of the coordinates of all landmark points between two frames:

[0054]

[0055] In the formula, p t,i and p t+1,iThey are the coordinates of the same calibration point in two consecutive frames t and t+1 respectively. Here, i represents the i-th calibration point, and the same applies hereinafter.

[0056] Coordinate variance Var:

[0057] Calculate the variance of the coordinates of each group of calibration points over all frames to evaluate the position stability of the calibration points:

[0058]

[0059] Here, x i represents the x coordinate of the i-th calibration point, is the average value of these coordinates. It needs to be calculated separately for each dimension (x, y, z) and then averaged. L is the number of calibration points.

[0060] Average value S of absolute spatial drift mean :

[0061] Calculate the average value of the spatial drift of each calibration point between two consecutive frames:

[0062]

[0063] Maximum value S of absolute spatial drift max :

[0064] Maximum value of the absolute spatial drift of all calibration points between two consecutive frames:

[0065]

[0066] Finally, these calculation results are integrated into a feature data block of size (G, F) for model training, where F is the number of features, which is L×4.

[0067] (3) Machine learning and transfer learning classification sub-module; Use a machine learning classifier for classification

[0068] The present invention supports classification using machine learning methods such as support vector machines, random forests, linear discriminant analysis, classification and regression trees, etc. Among them, the support vector machine (Support Vector Machine, SVM)

[23] has the best classification effect, so the SVM method is used for classification subsequently. SVM is a supervised learning algorithm for binary classification. Its basic principle is to find an optimal hyperplane that can maximize the margin between training samples to achieve effective discrimination of different classes. In an n-dimensional feature space, data points are mapped on an n-dimensional coordinate plane. The goal of SVM is to find an n-1-dimensional hyperplane such that the margin between samples of the two classes is as large as possible. The mathematical expression of this hyperplane can be expressed as:

[0069] f(x) = wT x + b,

[0070] where w is the normal vector of the hyperplane and b is the intercept. The distance from any point in space to this hyperplane can be expressed as |w T x + b|. The decision rule for classification can be formalized as:

[0071]

[0072] In practical applications, some data features are linearly inseparable in the original feature space. To handle this situation, SVM introduces the concept of kernel functions, which allows the algorithm to map the data into a higher-dimensional Hilbert space. In this new feature space, data points that were originally difficult to separate can be effectively linearly separated by a hyperplane. This method extends the application scope of SVM, enabling it to handle more complex classification problems. In the present invention, the classification results of linear kernel functions, quadratic kernel functions, and Gaussian kernel functions are tested. Among them, the linear kernel function performs the best, so the linear kernel function is used for classification. The model is trained with the manually labeled facial tic data of 20 individuals, and the recognition accuracy of tics in the main parts of the face can reach over 95%. It is used to evaluate the number of tics per minute of TS patients (how many 1s windows are tics), and then to judge the improvement of tics by median nerve regulation with different parameters.

[0073] Based on the above three parts, the present invention can output a multi-frequency sequence of stimulating pulse currents to stimulate the median nerve in the upper computer software for TS intervention treatment. The set of frequencies that is most effective in improving the patient's tics is evaluated by the facial tic recognition method as the home treatment parameters, and then the treatment parameters are preset on the mobile control software for the convenience of daily use by patients and their families.

[0074] Features of the present invention:

[0075] (1) Multi-band electrical stimulation: The present invention proposes a treatment plan of multi-band electrical stimulation, which can optimize the treatment effect of patients with Tourette syndrome (TS) through a multi-frequency stimulation sequence. Different from the traditional single-frequency electrical stimulation, the present invention can flexibly adjust according to the actual needs of patients to obtain the best treatment parameters. Clinical tests show that under 20Hz stimulation, the tic symptoms of patients are on average improved by 37% (p < 0.01), the device battery life exceeds 8 hours, supports home non-invasive treatment, significantly improves the treatment convenience and safety, and provides an effective solution for TS patients with insufficient efficacy of traditional drugs.

[0076] (2) Portable device design: The present invention designs a small and portable electrical stimulation device with dimensions of 67.5mm × 48.7mm × 23.6mm. The device is designed to be compact and easy to carry, allowing patients to perform autonomous treatment at home, greatly reducing the inconvenience and dependence during the treatment process, and solving the defects of large size and poor portability of traditional devices.

[0077] (3) Personalized adjustment: The system provides flexible parameter settings and supports the custom adjustment of multi-band electrical stimulation (frequency, amplitude, pulse width, etc.), providing personalized treatment plans for patients. In particular, the system supports control by a host computer and a mobile device, facilitating timely adjustment by patients and doctors according to the treatment progress.

[0078] (4) Automatic efficacy evaluation: The present invention combines facial tic recognition technology. By analyzing the key landmark data on the patient's face, it monitors the treatment effect, can provide effective efficacy evaluation for patients and doctors, and obtain the optimal treatment frequency.

[0079] (5) Non-invasive treatment: Compared with invasive treatment methods such as deep brain stimulation, the median nerve electrical stimulation adopted in the present invention is a non-invasive treatment method, avoiding surgical risks, having high safety, and being suitable for long-term home treatment.

[0080] (6) Support for home treatment: Through the portable device and mobile device control, patients can perform treatment at home, avoiding the trouble of frequent medical visits, and greatly improving the treatment convenience and sustainability for patients. Brief Description of the Drawings

[0081] Figure 1 It is a schematic structural diagram of the tic disorder treatment system of the present invention.

[0082] Figure 2 It is an illustration of the 3D printed housing.

[0083] Figure 3 Physical connection diagram of the electrical stimulator.

[0084] Figure 4 Software interface of the host computer on the PC side.

[0085] Figure 5 Control flow chart of the host computer - stimulator.

[0086] Figure 6 Connection interface and treatment interface on the mobile phone side.

[0087] Figure 7 Schematic diagram of the implementation plan for multi-band median nerve treatment.

[0088] Figure 8 Schematic diagram of the personalized home treatment plan.

[0089] Figure 9 Comparison chart of the treatment effects of electrical stimulation at the same frequency.

[0090] Figure 10 Home treatment results of subjects S3 and S5 (the dotted line is the linear fitting result). Specific implementation manner

[0091] Based on the above-mentioned Tourette's syndrome treatment system, the specific operations for multi-band electrical stimulation diagnosis and treatment are as follows:

[0092] (1) Setting of multi-band electrical stimulation parameters

[0093] The bipolar electrode is applied above the median nerve of the right wrist, with the anode close to the wrist side, the edge tangent to the palm crease line, and the cathode 5 cm away from the center of the anode. A non-woven hydrogel electrode with a diameter of 2 cm is used for electrical stimulation. The current is a pulsed signal with a pulse width of 0.1 ms. The reference frequencies are sequentially selected as 10 Hz (α band), 20 Hz (β band), and 50 Hz (γ band). The current intensity is set to the level that can just cause the thumb of the subject to flex slightly or tremble slightly.

[0094] (2) Design of multi-band electrical stimulation scheme

[0095] The on-site electrical stimulation treatment scheme is as Figure 7 shown. Before the start of the experiment, the electrical stimulation starts from 1 mA and gradually increases until the current magnitude that can cause the thumb of the subject to flex slightly or tremble slightly is found as the current threshold. During the experiment, different parameters and stimulation sequences are set in the upper computer software on the PC according to different experimental purposes.

[0096] (3) Recording of electrical stimulation data

[0097] During the on-site stimulation interval, the subjects were evaluated using scales, specifically including the Yale Global Tic Severity Scale (YGTSS), Premonitory Urge to Tics Scale (PUTS), Brown Attention-Deficit Disorder Symptom Checklist (BATS), Yale-Brown Obsessive Compulsive Scale (Y-BOCS), and Montreal Cognitive Assessment Scale (MOCA) to determine the basic medical history, mental state, and cognitive level of the subjects. A dual-camera synchronous recording method was used to capture the all-round behavior of the subjects. The main camera was set to shoot from the front view, adjusted to the horizontal line of the subjects' eyes, ensuring that the face was centered in the frame and the upper body (from the head to the waist) was fully visible. The auxiliary camera was located on the left side of the subject, and the angle was set to clearly capture the arm movements and neck activities of the subject. The perspectives of the two cameras were complementary, covering the key behavior areas of the subject. To ensure the precise synchronization of video data, an electronic clock with the display of Coordinated Universal Time (UTC) was placed at the edge of the camera's field of view, ensuring that the clock was always visible throughout the recording process for subsequent synchronization of video data. During the recording process, "Tom and Jerry" was played on a tablet on the table in front of the subject to make the subject in a passive and relaxed state, and the subject was informed not to actively suppress their tics. After the recording, a model of facial tics trained by doctors to mark the results was used to identify the tics of the patients.

[0098] (4) According to the identified changes in tic frequency, the best stimulation parameters for improving tics in the stimulation sequence were set on the mobile control software as the parameters for daily home use for daily home treatment.

[0099] Example:

[0100] A total of 6 subjects fully participated in the on-site electrical stimulation experiment. They all had a disease course of more than 7 years (10.2 ± 4.0), and all suffered from moderate or above tic disorders (YGTSS = 50.2 ± 15.1, PUTS = 23.2 ± 7.8, BATS = 45.5 ± 8.5), without obvious obsessive-compulsive behaviors (Y-BOCS = 10 ± 6.2), and had sufficient cognitive ability (MoCA = 29.5 ± 1.05) to understand the experimental content. The main facial symptoms included eye blinking, squinting, swallowing, nose sniffing, throat clearing, etc. Overall, for these 6 subjects, at a stimulation frequency of 10 Hz, the average number of tics decreased by 29% (p < 0.05); at 20 Hz stimulation, the average improvement of tic symptoms was the most significant, reaching 37% (p < 0.01); while at 50 Hz stimulation, the average number of tics decreased by 24% (p < 0.05). Although the effect was still obvious, compared with 20 Hz stimulation, the improvement amplitude decreased slightly, and the significance was not high (p = 0.047), as Figure 9 shown. These results suggest that different frequencies of electrical stimulation may have different effects on improving tic symptoms, especially at a frequency of 20 Hz, the effect is particularly significant.

[0101] Subjects S2, S3, S5, and S6 and their families subjectively reported a reduction in tic impulses or alleviation of tic symptoms after electrical stimulation treatment. Among them, the subjective reports of S2, S5, and S6 indicated that 20 Hz stimulation could significantly relieve their tic symptoms, while S3 felt that 10 Hz electrical stimulation treatment had a better effect. This was consistent with the statistical results. These 4 subjects had a relatively large reduction in tic frequency after receiving electrical stimulation. S1 and S4 subjectively reported that electrical stimulation treatment did not significantly relieve their symptoms after receiving different electrical stimulation treatments. All subjects gave positive evaluations of the portability and ease of operation of the electrical stimulator and expressed their willingness to use it in daily life. Except for S3 and S6 indicating a very high somatic comfort during electrical stimulation, the other 4 subjects said that muscle contraction during electrical stimulation might affect daily operations such as typing and holding a pen.

[0102] After the on-site experiments of subjects S3 and S5, according to the changes in the improvement of tic counts and the subjective reports of the subjects, S3 chose 20 Hz and S5 chose 50 Hz as the home electrical stimulation treatment frequencies. After 5 consecutive days of home experiments, twice a day, the average number of tics per minute in the pre-stimulation and stimulation phases counted every day was averaged and linearly fitted, as Figure 10 shown by the dotted line. The average number of tics per minute of the two subjects showed an overall downward trend. The average number of tics per minute of patient S2 during stimulation on the 5th day decreased by 43% compared with the first day (p < 0.05), while the number of tics before stimulation in S5 changed little within 5 days, but changed significantly during stimulation, decreasing by 76% (p < 0.05). Both subjects said that the electrical stimulator was very convenient for daily use.

[0103] The electrical stimulation in this study is small in size (67.5 mm × 48.7 mm × 23.6 mm), easy to carry, can be controlled by a computer or mobile phone, is simple to operate, and can conduct experiments and treatments according to preset stimulation modes. And it has achieved treatment effects in the multi-band median nerve treatment implementation plan for six patients and the personalized home treatment implementation for two patients, providing strong support for the clinical application of treatment methods for tic disorders, providing new treatment options for TS patients, especially for those patients with poor effects of traditional drug treatments.

[0104] References

[0105] [1] American Psychiatric Association D, American Psychiatric Association D S. Diagnostic and statistical manual of mental disorders: DSM-5[M]. Washington, DC: American psychiatric association, 2013.

[0106] [2] Tourette Syndrome Classification Study Group. Definitions and classification of tic disorders[J]. Archives of neurology, 1993, 50: 1013-1016.

[0107] [3] Shapiro A K, Shapiro E S, Young J G, et al. Gilles de la Tourette syndrome[M]. Raven Press, Publishers, 1988.

[0108] [4] H.S. Singer, “Treatment of Tics and Tourette Syndrome,” Curr Treat Options Neurol, vol. 12, no. 6, pp. 539-561, Nov. 2010.

[0109] [5] C. Nilles, A. Hartmann, E. Roze, D. Martino, and T. Pringsheim, Chapter 23 - Tourette syndrome and other tic disorders of childhood, in Handbook of Clinical Neurology, vol. 196, D.S. Younger, Ed., in Motor System Disorders, Part II: Spinal Cord, Neurodegenerative, and Cerebral Disorders and Treatment, vol. 196., Elsevier, 2023, pp. 457–474.

[0110] [6]Lebowitz E R,Motlagh M G,Katsovich L,et al.Tourette syndrome inyouth with and withoutobsessive compulsive disorder and attention deficithyperactivity disorder[J].European child&adolescent psychiatry,2012,21:451-457.

[0111] [7]K.A.Johnson,Y.Worbe,K.D.Foote,C.R.Butson,A.Gunduz,and M.S.Okun,Tourettesyndrome:clinical features,pathophysiology,and treatment,The LancetNeurology,vol.22,no.2,pp.147–158,Feb.2023.

[0112] [8]Levine J L S,Szejko N,Bloch M H.Meta-analysis:adulthood prevalenceof Tourettesyndrome[J].Progress in Neuro-Psychopharmacology and BiologicalPsychiatry,2019,95:109675.

[0113] [9]Hollis C,Pennant M,Cuenca J,et al.Clinical effectiveness andpatient perspectives ofdifferent treatment strategies for tics in childrenand adolescents with Tourette syndrome:asystematic review and qualitativeanalysis[J].2016.

[0114]

[10] K.A. Johnson, Y. Worbe, K.D. Foote, C.R. Butson, A. Gunduz, and M.S. Okun, Tourette syndrome: clinical features, pathophysiology, and treatment, The Lancet Neurology, vol. 22, no. 2, pp. 147–158, Feb. 2023.

[0115]

[11] S. Armstrong, M.V. Sale, and R. Cunnington, Neural Oscillations and the Initiation of Voluntary Movement, Front. Psychol., vol. 9, Dec. 2018, doi:10.3389 / fpsyg.2018.02509.

[0116]

[12] Armstrong S, Sale M V, Cunnington R. Neural oscillations and the initiation of voluntary movement[J]. Frontiers in psychology, 2018, 9:2509.

[0117]

[13] Thut, G.; Veniero, D.; Romei, V.; Miniussi, C.; Schyns, P.; Gross, J. Rhythmic TMS causes local entrainment of natural oscillatory signatures. Curr. Biol. 2011, 21, 1176 - 1185

[0118]

[14] Pogosyan A, Gaynor L D, Eusebio A, et al. Boosting cortical activity at beta - band frequencies slows movement in humans[J]. Current biology, 2009, 19(19):1637 - 1641.

[0119]

[15] B.Morera Maiquez et al.,Entraining Movement-Related BrainOscillations to Suppress Ticsin Tourette Syndrome,Current Biology,vol.30,no.12,pp.2334-2342.e3,Jun.2020.

[0120]

[16] Houlgreave,M.S.;Morera Maiquez,B.;Brookes,M.J.;Jackson,S.R.Theoscillatory effectsof rhythmic median nerve stimulation.Neuroimage 2022,251,118990.

[0121]

[17] B.M.Maiquez et al.,A double-blind,sham-controlled,trial of home-administered rhythmic10-Hz median nerve stimulation for the reduction oftics,and suppression of the urge-to-tic,inindividuals with Tourette syndromeand chronic tic disorder,Journal of Neuropsychology,vol.17,no.3,pp.540–563,2023.

[0122]

[18] A.M.Iverson et al.,Median Nerve Stimulation for Treatment ofTics:Randomized,Controlled,Crossover Trial,Journal of Clinical Medicine,vol.12,no.7,Art.no.7,Jan.2023.

[0123]

[19] Digitimer-DS7A.DS7A & DS7AH High Voltage Constant Current Stimulators[EB / OL].https: / / www.digitimer.com / product / human - neurophysiology / peripheral - stimulators / ds7a - ds7ah - hv - current - stimulator / .

[0124]

[20] TENS7000.TENS 7000 2nd Edition Digital TENS Unit with Accessories[EB / OL].https: / / tens7000.com / products / tens - 7000 - 2nd - edition - digital - tens - unit - with - accessories.

[0125]

[21] Thaman B,Cao T,Caporusso N.Face mask detection using mediapipe facemesh[C] / / 2022 45th Jubilee International Convention on Information,Communication and Electronic Technology(MIPRO).IEEE,2022:378 - 382.

[0126]

[22] N S,Sallandt G M,Schappert R,et al.Automated motor tic detection:A machine learning approach[J].Movement Disorders,2023,38(7):1327 - 1335.

[0127]

[23] Pisner D A,Schnyer D M.Support vector machine[M] / / Machine learning.Academic Press,2020:101 - 121。

Claims

1. A Tourette syndrome treatment system based on multi-band median nerve regulation, characterized in that: Through multi-band stimulation sequence testing and facial tic recognition evaluation, the optimal treatment parameters are found before home treatment; specifically, it includes three parts: electrical stimulator system, software control system and facial tic recognition module; among which: (i) The electrical stimulator system comprises: (1) A main control module, used to receive control instructions and generate stimulation parameters; (2) A neural stimulation module is a neural stimulation chip that communicates with the main control module through a serial port and is used to generate a bidirectional pulse square wave signal with adjustable frequency, amplitude and pulse width; that is, it supports multiple stimulation parameters, including configuration of frequency, amplitude and pulse width; (3) Power supply module, including 500mA lithium battery and charging management circuit, supporting the system to work continuously for 8 hours; (4) an electrode interface module, which is connected to the non-woven gel electrode through a circular headphone jack to output stimulation signals; (5) Status indicator module, which displays the system status through programmable LEDs; (6) Customized 3D printed housing for easy daily use; The computer host or mobile phone software communicates with the main control module through the Bluetooth module to send stimulation parameters and control instructions; (ii) The software control system includes: host computer software for TS intervention treatment and mobile terminal control software; wherein: (1) Host computer software for TS intervention treatment; That is, the PC-side host computer software system, the stimulator human-computer interaction interface (GUI), including the Bluetooth connection module, parameter setting area and stimulation sequence control area; (2) Mobile terminal control software; That is, the mobile APP provides preset treatment modes and manual modes, which are packaged into APK files through HBuilderX; (III) The facial tic recognition module is used to evaluate the efficacy of median nerve intervention with different parameters; specifically, it includes three submodules: facial landmark information extraction, landmark data processing and feature extraction, and machine learning and transfer learning classification; wherein: (1) Facial landmark information extraction submodule; First, the recorded video of the patient's facial twitch is divided into 1s video windows, and then MedianPipe FaceMesh is selected to extract the facial model calibration points. 222 key calibration points are selected from them, including 80 points of the eyes, 22 points of the eyebrows, 49 points of the nose, 35 points of the mouth, and 36 points of the facial contour to build the basic features of the classification model. These include 170 points used by Brügge and points that doctors and researchers agree are more important. Facial calibration points are extracted in real time at 60 frames of video; (2) calibration point data processing and feature extraction submodule; The data is organized into a four-dimensional data block (G, L, T, P), where G represents the number of windows, L represents the number of calibration points in each window, T represents the number of frames in each window, and P represents the spatial coordinates of the calibration points; a digital filter is used to low-pass filter the trajectory sequence of 222 calibration points; the trajectory features of the calibration points are extracted from the filtered data; for each data window, the jth calibration point of the i-th window is extracted to form a 30×3 trajectory matrix; based on the trajectory matrix, the maximum Euclidean distance, coordinate variance, average value and maximum value of absolute spatial drift are calculated for each set of calibration point coordinates: (3) Machine learning and transfer learning classification submodule: Use machine learning classifiers to classify and determine the improvement of tics by median nerve regulation with different parameters.

2. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 1, characterized in that: In the electrical stimulator system, the main control module is responsible for processing system control, Bluetooth communication and command interaction with the host computer, and specifically completes the following functions: (1) Communicate with the host computer or mobile phone APP via Bluetooth, receive stimulation parameters and control instructions, and perform corresponding control operations; (2) Exchange data with the neural stimulation chip through the serial port to accurately control the generation of electrical stimulation signals and realize the analysis and real-time control of user-set parameters such as frequency, amplitude, and pulse width; (3) Through I 2 C protocol communicates with programmable LED lights to display system status in real time, making it easy for users to check the operation of the equipment; (4) Time synchronization with external devices is achieved through the GPIO port, and a bidirectional trigger mode is supported. The system can be configured in advance to receive a high-level trigger signal from an external device to start system operation at the beginning of the experiment, or it can be configured to send a high-level signal to an external device when the experiment starts.

3. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 2, characterized in that: In the electrical stimulator system, the output parameters satisfy: (1) Pulse frequency range: 1-500Hz, minimum adjustment step 0.5Hz, error ≤2%; (2) Current amplitude range: 0.5-20mA, minimum adjustment step 0.1mA, error ≤2%; (3) Pulse width range: 10-2000μs, minimum adjustment step 100μs, error ≤3%; (4) Pulse energy range: 1.25×10 -8 J to 1×10 -4 J; (5) DC component ≤ 0.01mA.

4. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 1, characterized in that: In the host computer software module of the TS intervention treatment: The Bluetooth module supports Bluetooth 5.2 dual-mode communication, which can quickly scan nearby Bluetooth devices and display all connectable electrostimulator devices; the user selects the target electrostimulator from the device list, and after the connection is established, the software monitors the connection status in real time to ensure a stable connection; when the connection is disconnected, the system pops up a prompt to allow the user to reconnect or select another device; The parameter setting area is used to set the output parameters of the electrostimulator, including frequency, pulse width, and amplitude; users adjust these parameters by simple input or selection boxes; after the parameter value is adjusted, the software sends it to the electrostimulator via Bluetooth to control its output in real time; Stimulation sequence control area, where users can edit and manage stimulation sequences; design a complete stimulation sequence by setting the start and end time, frequency and intensity of stimulation; after finding the threshold of median nerve stimulation, save the parameters, and edit and adjust them on the interface; During the execution of the stimulation sequence, the stimulation status, test progress and remaining stimulation time are displayed in real time to monitor the experimental process and ensure the smooth progress of the experiment.

5. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 4, characterized in that: The mobile terminal control software: It supports direct installation and use on mobile phones. It can quickly scan nearby Bluetooth devices and display the electrical stimulation devices that can be connected. Select the target electrical stimulator from the device list. After the connection is established, enter the parameter setting interface; set the amplitude of the stimulation current, and have two modes: manual switch and start treatment according to the preset mode. Both modes can manually interrupt the stimulation; the software monitors the connection status in real time to ensure a stable connection; when the connection is disconnected, the system pops up a prompt to allow the user to reconnect or select another device; The software supports two treatment modes: manual mode and preset treatment mode. In manual mode, users can flexibly control the start and stop of stimulation. In preset treatment mode, stimulation is performed automatically. Users can also interrupt stimulation at any time. The connection status is monitored in real time to ensure a stable connection. When the connection is disconnected, the system pops up a prompt to allow the user to reconnect or select another device.

6. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 1, characterized in that: In the calibration point data processing and feature extraction submodule, the formula for calculating the four features of maximum Euclidean distance, coordinate variance, average value and maximum value of absolute spatial drift is: Maximum Euclidean distance D max : For the coordinates of the calibration point in two consecutive frames, p1:(x1,y1,z1) and p2:(x2,y2,z2), the Euclidean distance of this calibration point in the consecutive frames is: The maximum Euclidean distance D between the coordinates of all calibration points in two frames max : In the formula, p t,i and p t+1,i are the coordinates of the same calibration point in two consecutive frames t and t+1, i represents the i-th calibration point, the same below; Coordinate variance Var: Here, x i represents the x-coordinate of the ith calibration point, is the average value of these coordinates, which needs to be calculated for each dimension (x, y, z) and then averaged. L is the number of calibration points. The average value of the absolute spatial drift S mean : That is, the average value of the spatial drift of each calibration point between two consecutive frames. The formula is: The maximum value of absolute spatial drift S max : That is, the maximum value of the absolute spatial drift of all calibration points between two consecutive frames: Finally, these calculation results are integrated into a feature data block of size (G, F) for model training, where F is the number of features, that is, L×4.

7. The Tourette syndrome treatment system based on multi-band electrical stimulation according to claim 6, characterized in that: In the machine learning and transfer learning classification submodule, support vector machine (SVM) is used for classification; specifically, an optimal hyperplane is found to maximize the interval between training samples to achieve effective distinction between different categories; Assuming that in an n-dimensional feature space, the data points are mapped on an n-dimensional coordinate plane, the goal of SVM is to find an n-1-dimensional hyperplane so that the interval between samples of two categories is as large as possible; the mathematical expression of this hyperplane is expressed as: f(x)=w T x+b, Among them, w is the normal vector of the hyperplane, b is the intercept; the distance from any point in space to this hyperplane is expressed as |w T x+b|; The decision rule for classification is formalized as: Introducing kernel functions in SVM allows the algorithm to map data to a higher-dimensional Hilbert space; in this new feature space, data points that were originally difficult to separate can be effectively linearly separated by the hyperplane; specifically, the kernel function introduced is the linear kernel function.

Images