AI-POWERED ADAPTIVE AND LOAD-BALANCED DEEP BRAIN STIMULATION SYSTEM
Patent Information
- Application Number
- TR202613132
- Authority / Receiving Office
- TR · TR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-08-04
- Publication Date
- 2026-08-21
Smart Images
Figure 00000020_0000
Abstract
Description
1 TARIFF AI-POWERED ADAPTIVE AND LOAD-BALANCED DEEP BRAIN STIMULATION SYSTEM Technological Field: The invention is based on electrophysiological principles for the management of neurodegenerative diseases. AI-powered adaptive and load-balanced systems that process biological data layers together. It is related to the deep brain stimulation system. 10 State of the Art: Current techniques for deep brain stimulation systems are used in the treatment of various neurological diseases. It is used to apply electrical stimulation to target brain regions. This 15 These systems generally target a specific brain region in advance via electrodes. an electrical signal with specified frequency, amplitude, and pulse width values is provided. However, stimulation is involved in a significant portion of current practices. parameters are automatically determined according to the patient's current neurophysiological state. It does not change; it operates based on fixed values determined by the physician. This 20 the condition in which the course of the disease and neural activity vary over time This results in a stimulation approach that is unable to provide adequate adaptation in these situations. One of the main problems encountered in current systems is that stimulation is often It is implemented using open-loop logic. In open-loop systems, the device, target 25 By actively evaluating the feedback received from the brain region, I stimulate myself. It does not rearrange the parameters. Therefore, the patient's sleep, wakefulness, drug effect, taking into account symptom severity, disease stage, or momentary changes in neural activity The same stimulation profile can be maintained without interruption. This mode of operation is sometimes necessary. This can lead to insufficient stimulation, and sometimes to unnecessary or excessive stimulation. 30 2 Another problem with the current state of the art is some of what is called a closed loop. even systems with a limited number of electrical signals or simple threshold values It is the ability to withstand the effects. In such systems, the power increase is usually in a specific frequency band. in a specific signal amplitude or a predefined single biosignal stimulus decision This is taken as the basis. However, neural damage occurring in neurodegenerative diseases is 5 It is too complex to be explained by just a single electrical parameter. The brain synchronization between networks, phase relationships, spectral distribution, and signal irregularity. When multiple indicators such as these are not evaluated together, the decision to stimulate the patient may not be based on the patient's needs. It becomes difficult to represent the true neurophysiological state. Another significant deficiency in terms of neurodegenerative diseases is the current stimulation. The problem is that these systems often fail to take into account the biological and molecular burden of disease. In Alzheimer's and similar diseases, the clinical picture is not solely based on electrical activity. It is not limited to the disorder itself, but also includes protein accumulation, neuroinflammation, and genetic predisposition. Many biological factors, such as neurotransmitter changes and disease stage, also influence the process. However, known systems often only detect electrophysiological signals. because it makes decisions based on the patient's biological background and electrical activity. A comprehensive assessment cannot be made between them. Current systems also have limited patient-specific adaptation capabilities. For the same disease diagnosis, 20... Even in individuals with other conditions, the response of the target brain region depends on the stage of the disease, tissue impedance, Drug use, genetic predisposition, and biomarker profiles may vary. Fixed or Systems that operate with a narrow control logic do not adequately account for individual differences. because it could not take into account the same level of efficacy and safety in every patient. It may not be able to provide this. This creates the need for personalized stimulation. 25 Another problem encountered with current technology is the energy associated with unnecessary stimulation. It is consumption. In systems that operate continuously with fixed parameters, the symptom is low. Even during periods when the need for stimulation is reduced, the device continues to consume energy. This situation particularly leads to shorter battery life in implantable devices, charging or battery 30 the need for change increases and consequently the patient's operational burden 3 This can cause it to rise. Therefore, stimulation should only be done when needed. Administering it under the right conditions and at the correct dosage is a technically important requirement. In addition, the electrical stimulation applied to the target tissue must be safely controlled. This is a problem that remains significant in the current state of the technique. During stimulation 5 Insufficient balance of the electrical charge delivered to the tissue, at the electrode-tissue interface. This can lead to unwanted stress accumulation. This situation affects tissue pH with long-term use. Changes in value, electrode corrosion, tissue irritation, or risk of cellular damage. This can lead to complications. Therefore, the stimulation current should be considered not only in terms of parameters, but also... It also needs to be checked in terms of the net load transferred to the tissue. 10 Changes in tissue impedance are also an important technique for existing systems. This creates a limitation. The characteristics of the biological tissue surrounding the electrode change over time. Because it can vary, stimulation applied at the same voltage level at different times This may correspond to different current values. This situation means that the actual stimulation dose is 15. This reduces its predictability. Therefore, the current applied to the target tissue... as independently and controllably as possible from impedance changes It needs to be produced. “Fornical Closed-Loop Stimulation for 20” published by Senova and colleagues. The study titled "Alzheimer's Disease" examines memory processes in Alzheimer's disease. in terms of closed-loop stimulation strategies of the associated fornix region It has been stated that it needs to be evaluated. The study in question concerns open-loop deep brain research. stimulation applications to improve memory impairments in Alzheimer's disease that it may be limited in its effectiveness and that stimulation depends on the patient's neural state 25 This highlights the importance of adaptation. However, the study shows that electrophysiological where data is processed along with biochemical, molecular, or genetic data layers, It does not recommend an AI-powered and load-balanced device. “Multisensory gamma stimulation 30” published by Murdock and colleagues. In the study titled "promotes glymphatic clearance of amyloid," a 40 Hz gamma frequency was used. The effect of multisensory stimulation on the glyphatic clearing process in an Alzheimer's model and 4 Its relationship with amyloid removal has been investigated. In the study, it was related to gamma rhythm. sensory stimulation affects cerebrospinal fluid movements and interstitial fluid flow. It has been shown that it can have an effect. However, the study in question was about implantable In deep brain stimulation systems, electrophysiological data are obtained from the target brain region. Stimulation 5 by evaluating the signals together with biomarker data. a system for determining parameters in a closed-loop manner It does not explain. “Optimal anti-amyloid-beta therapy for” published by Hao and colleagues In the study titled "Alzheimer's disease via a personalized mathematical model," 10 Personalized mathematical anti-amyloid beta therapy in Alzheimer's disease Optimization using model and biomarker data is addressed in this study. clinical data such as cognitive data, cerebrospinal fluid biomarkers, and imaging data By utilizing these parameters, treatment planning can be made individualized. This shows. However, the study shows that a 15 that produces electrical stimulation closed-loop deep stimulation using a neuromodulation device or electrophysiological feedback It does not reveal a brain stimulation system. US20230233858A1 and “Method and device to enhance waste clearance in the In the patent application titled "brain," 20 methods are described to increase the removal of waste from the brain. A stimulation-based device and method are described in the application. Stimulation waveform and temporal pattern, cerebral pulsatility and associated waste products. It is stated that the cleaning processes may be affected. However, this document... Neurophysiological findings derived from electrophysiological signals related to Alzheimer's disease attributes with biochemical, molecular, genetic or behavioral data layers 25 combined to allow artificial intelligence to select the stimulation parameter and active load. It does not explain how balancing is applied to the target brain region. “PET-validated EEG-machine learning” published by Kim and colleagues. algorithm predicts brain amyloid pathology in pre-dementia Alzheimer's disease” 30 In the study titled [Study Title], pre- EEG data and machine learning algorithm were used. The aim was to predict amyloid pathology in the Alzheimer's stage of dementia. In the study, EEG-based features were correlated with PET-confirmed amyloid pathology. and how machine learning models can be associated with Alzheimer's screening It has been shown that it can be used in this regard. However, the study is based on prediction. It offers a diagnostic / classification approach, using the obtained EEG or biomarkers. Using the data, stimulation parameters can be determined in real-time or closed-loop. 5 It does not describe a deep brain stimulation system that determines this in this way. In conclusion, the current technology includes fixed-parameter excitation, limited closed-loop control, Decision-making based on a single data source, without considering the biological patient profile. Lack of individual adaptation, energy inefficiency, load balance risk, impedance 10 Technical problems such as dose uncertainty and processing delay resulting from variations Therefore, electrophysiological signals are integrated with the biological data layer. by evaluating them together, dynamically determining the stimulation parameters, and More advanced deep brain stimulation providing safe / load-balanced stimulation. The system is needed. 15 Description of the invention: The primary goal of the invention is to replicate the classical deep brain imaging techniques used in neurodegenerative diseases. 20 resulting from the fixed and unidirectional stimulation structure of stimulation systems The goal is to address the deficiencies. In current systems, stimulation often provides immediate relief to the patient. without considering neurological state, biological load or pathological changes This practice leads to unnecessary stimulation, energy loss, side effects, and in the long term... This leads to a decrease in treatment effectiveness. The invention uses electrophysiological signals. by evaluating biological, molecular, genetic and behavioral data together 25 adapting the stimulation to the patient's current neurobiological status It provides. One of the most important advantages of the invention is that the stimulation decision is made by only one person. It does not rely on electrical signals or simple threshold values. The system targets the brain 30 when extracting neurophysiological features from electrophysiological signals received from the region, It also takes into account biomarker and patient profile data. In this way... 6 Not only the electrical manifestation of the disease, but also its biological and molecular background. It participates in the evaluation. This results in a more accurate, personalized, and situation-specific assessment. Stimulation control is achieved. Because the invention offers an adaptive operating structure instead of fixed-parameter continuous excitation, 5 stimulation only under the required conditions and with appropriate parameters This allows for its implementation. This reduces unnecessary current application, targeting the goal. It reduces the likelihood of stimulation of tissues outside the affected area and can cause speech disorders and muscle problems. It helps reduce side effects such as muscle spasms, loss of balance, or similar issues. In addition, there is the habituation effect that can result from the brain being constantly exposed to the same stimulus. 10 It can be reduced. The use of an artificial intelligence inference module enables linear and simple control of the system. According to their approaches, this enables them to make more flexible decisions. Brain signals In most cases, it exhibits complex and variable characteristics, so simply increasing the amplitude or 15 Making decisions based on a single frequency band may be insufficient. In the invention... The AI-powered evaluation structure used combines multiple data layers. by considering the pathological condition in a more holistic way and accordingly It defines the stimulation parameter set. Another advantage of the invention is that the stimulation parameters undergo safety checks. It is implemented by passing through. The system uses a specified frequency, amplitude, pulse width, and It can check similar values in terms of tissue safety. Thanks to this structure... Exceeding limits such as permitted current, voltage, total load, or tissue impedance This ensures safer application of stimulation by preventing complications. 25 The active load balancing unit provides a significant technical safety advantage to the invention. It provides benefits. Positive and negative energies are transferred to the target tissue during stimulation. Balancing the charges prevents charge accumulation at the electrode tip. Thus, changes in tissue pH, electrode corrosion, and long-term tissue damage can occur. Risks are reduced. This feature is especially important for long-term or repeated stimulation. It increases the reliability of the system in its applications. 7 Using a voltage-controlled current pump allows for changes in tissue impedance, despite the variations in current. This allows for a more stable stimulation current to be applied to the target area. Impedance changes that may occur over time at the electrode-tissue interface, in classical structures This may cause an undesirable change in the stimulation dose. In this invention, however, the stimulation is 5. more controlled production of the current, predictability of the applied dose, and It increases its repeatability. The invention's ability to work with embedded processing architecture reduces delays in the decision-making process. It reduces. The reception, processing, and stimulation of the electrophysiological signal 10 Parameter determination and stimulation application within the same system. Since this can be achieved, dependence on external processing units is reduced. This situation especially in monitoring rapidly changing neural activity and closed-loop control. This provides a significant advantage in the event. The modular nature of the invention allows it to be adapted to different disease types and different target brain regions. This ensures its adaptability. The biomarkers used are electrophysiological. Features, AI model, and stimulation parameters vary depending on the application area. It can be modified. Therefore, the system is not limited to just one specific disease, When properly structured, different neurodegenerative or neuromodulation 20 It can also be used in situations where it is necessary. In conclusion, the invention combines electrophysiological feedback with a biological data layer. an AI-powered decision-making mechanism that determines stimulation parameters and Holistic deep brain 25 that provides safe stimulation thanks to active load balancing. It offers a stimulation system. In these respects, the invention is superior to existing fixed-parameter and single-stimulation systems. More personalized, more secure, and more energy-efficient compared to data-driven systems. It provides more efficient and technically more advanced stimulation control. Explanation of the Figures: 30 8 The invention will be described by referring to the attached figures, so that the features of the invention can be explained. It will be understood and appreciated more clearly, but the purpose of this invention is this obvious It is not about limiting it with regulations. On the contrary, the invention is defined by the accompanying claims. all alternatives, modifications, and options that could be included within the defined area The aim is to cover their equivalences. The details shown are only for 5 of the present invention. It is shown to illustrate the preferred arrangements and both the methods shaping, as well as the most useful and conceptual features of the invention's rules and principles It should be understood that they are presented to provide an easily understandable definition. In these drawings; Figure 1 is a block diagram showing the operation of the system described in the invention. 10 Illustrations that will help understand this invention are shown in the attached image. They are numbered and their names are given below. References Explanation: 15 10. Deep brain stimulation system 11. Biopotential electrode 12. Analog preprocessing and analog-to-digital converter unit 13. Central Processing and Control Unit 20 14. Artificial intelligence inference module 15. Biological data input interface 16. Dynamic stimulation signal generator 17. Voltage-controlled current pump 18. Active load balancing unit 25 19. Wireless communication module Detailed Description of the Invention: The terminology used here is intended solely to describe specific applications. 30 and does not limit the scope of the invention. Used here The term "and / or" refers to any of the items listed as related. 9 It includes one and all combinations thereof. Also, the singular "one" used here, "one" The terms "number" and "specified" are used in their plural forms unless the context explicitly indicates otherwise. It is designed to include singular forms such as those mentioned above. Furthermore, the terms used in this specification... The terms "includes" and / or "contains" refer to the specified features, steps, processes, It indicates the presence of elements and / or components, but one or more other 5 features, steps, processes, elements, components and / or groups thereof It will be understood that this does not exclude its existence or addition. Unless otherwise noted, all terms used herein (including technical and scientific terms), 10 in the sense that a person with general knowledge in the field to which this invention belongs would usually understand it. They have the same meaning. Furthermore, as defined in commonly used dictionaries... The terms will have a meaning consistent with the context of the relevant field and this explanation. it should be interpreted as idealized unless otherwise explicitly defined here. or it will be understood that it will not be interpreted in an overly formal sense. The description of the invention will reveal a series of techniques and steps involved. These are: Each of them provides benefits individually, and at the same time, one or more of them can be used together. In some cases, all of the other techniques described may be used in combination. Accordingly, To ensure clarity, each step in the disclosure of the invention should be presented as accurately as possible. By avoiding unnecessary repetition of all combinations, 20 Together, the specification and claims, such combinations fully constitute an invention and claims. It should be read with the understanding that it falls within its scope. The invention is an electrophysiological study developed for use in neurodegenerative diseases. evaluating the data together with biological and molecular biomarker data, 25 Patient or subject-specific stimulation parameters from the multilayered data obtained. determining and applying the determined parameters to the brain tissue in a load-balanced manner adaptive closed-loop deep brain stimulation system and how this system works It is related. The system developed within the scope of this invention is a classic fixed-parameter deep brain stimulation. unlike other systems, it only uses a predetermined fixed frequency and amplitude. It does not provide constant stimulation with its values. Instead, it uses values taken from the target brain region. It processes electrophysiological signals and interprets these signals to identify disease or pathological activity. It extracts attributes representing the condition, as well as biomarker data entered into the system. It takes this into account and evaluates all this data through an artificial intelligence inference module. It dynamically determines the stimulation parameters to be applied. Thus, the stimulation is 5 It can be adapted to the patient's or subject's immediate or current neurobiological status. becomes. In a preferred configuration of the invention, the system utilizes adaptive closed-loop deep brain technology. stimulation system (10), bio-potential electrodes (11), analog preprocessing and analog-10 Digital converter unit (12), central processing and control unit (13), artificial intelligence inference module (14), biological data input interface (15), dynamic stimulation signal generator (16), voltage controlled current pump (17), active load balancing unit (18) and wireless It includes the communication module (19). Adaptive closed-loop deep brain stimulation system (10), perception, data processing, decision It combines the functions of giving, generating, and receiving stimulation signals. It refers to the main system that performs this function. The system, specifically the hippocampus, entorhinal Electrophysiological studies of target brain regions such as the cortex, CA1, CA3, or similar areas 20 for the purpose of processing signals and determining stimulation parameters accordingly It is being structured. However, the system is limited only to Alzheimer's disease. not, but by selecting the appropriate target region and biomarker, other neurodegenerative or It can also be applied in disease groups requiring neuromodulation. It is configurable. Bio-potential electrodes (11) detect electrophysiological signals from the target brain region. to receive and / or transmit the specified stimulation signal to the target tissue These electrodes include micro-wire electrodes and deep brain stimulation electrodes. multi-contact electrode or similar electrode usable in the relevant technical field These structures can be selected. The signals detected by the electrodes are the local field potential, 30 ECoG or similar neural activity signals may be present. In the preferred approach. The electrodes studied the relationship between theta and gamma bands, phase-amplitude coupling, and spectral power. 11 changes or other signals representing pathological activity in the target brain region It is positioned to detect its components. Analog preprocessing and analog-to-digital converter unit (12), from bio-potential electrodes (11) makes the incoming low-amplitude raw neural signals processable. This unit; 5 signal amplification, analog filtering, noise reduction, sampling, and digital data. It performs the conversion processes. During preprocessing, unwanted noise is removed. components, motion artifacts, mains noise, or stimulation-induced disturbances The components can be reduced. Created by the analog-to-digital converter unit. The digital signal is transmitted to the central processing and control unit (13). 10 Central processing and control unit (13) manages the flow of data within the system, signal processing coordinating the processes, running the AI inference module (14) and stimulation It is the processing unit that controls the signal generator (16) and the active load balancing unit (18). This unit is a microcontroller, microprocessor, digital signal processor, FPGA, ASIC, or 15 It can be configured in the form of similar embedded processing hardware. In the preferred configuration... central processing and control unit, low-latency decision-making, and implantable or Embedded processor to provide energy consumption suitable for portable device architecture. It is created with its architecture. The AI inference module (14) runs on the central processing and control unit (13). or a software or hardware-supported decision module that communicates with this unit. This The module includes features extracted from electrophysiological data and a biological data input interface. It evaluates the biomarker data transferred to the system via (15) together. Artificial Intelligence inference module (14), XGBoost, decision tree based model, deep learning model, 25 A regression model can be structured as a probabilistic model, or a combination of both. The module's task is to apply the acquired multimodal data to the target brain region. frequency, amplitude, pulse width, waveform, stimulation duration and / or This involves defining parameters such as the stimulation pattern. Biological data input interface (15), transfer of non-electrophysiological data to the system. It refers to the data entry structure used for this. Through this interface, data is entered into the system and provided to the patient. 12 or biochemical, molecular, genetic or behavioral data belonging to the subject may be transferred. This data includes, for example, amyloid beta level, phosphorylated tau level, total tau level, Acetylcholine levels, inflammation-related biomarkers, APOE, PSEN1 or PSEN2. related genetic / molecular information, mRNA expression levels, or cognitive These could be test data representing performance. This data is directly processed by the system. 5 This could include real-time sensor data that is measured, as well as laboratory analysis, clinical monitoring, obtained from an external measuring device or database and fed into the system periodically. This could also include transmitted data. Thus, real-time electrophysiological feedback, This is supported by current or periodic biological data. The biological data input interface (15) can operate via wired or wireless communication. This interface can be, for example, SPI, I2C, UART, USB, Bluetooth, Wi-Fi, or a medical device. central processing and control with a similar communication protocol suitable for communication It can transfer data to the unit (13). The transferred biological data is transferred to the artificial intelligence inference module. (14) can be normalized before use with patient-based reference values 15 comparable to or represented within a common state vector with electrophysiological data. It can be done. In the electrophysiological signal processing stage, the central processing and control unit (13) is analog. 20 on the digital signals coming from the preprocessing and analog-to-digital converter unit (12) band-pass filtering, time-frequency analysis, spectral power calculation, phase-amplitude coupling determination, entropy calculation, noise / artifact removal, or similar processes. This can be achieved. In the preferred configuration, coupling between the theta and gamma bands, It is used to determine the network synchronization status in the target brain region. Signal processing, Fourier transform, fractional Fourier transform, wavelet transform or 25 This can be accomplished using a similar time-frequency analysis method. The system inputs biological data using features derived from electrophysiological signals. By combining the biomarker data obtained from the interface (15), the disease or pathological It creates a state vector representing the current status of the activity. This status is 30 vector, for example theta band strength, gamma band strength, theta-gamma phase-amplitude coupling value, signal entropy, amyloid beta level, tau level, genetic risk information and 13 It can include at least one of the behavioral performance metrics. State vector, artificial intelligence It is given as input to the inference module (14). The AI inference module (14) assesses the disease severity by evaluating the state vector. This score can form an indicator of pathological activity or a stimulation need score. These 5 The score is used to determine the stimulation parameters to be applied. Thus The system is based not only on a single threshold value, but also on multiple biological and The decision is made based on a combined evaluation of the electrophysiological indicator. This structure, This helps reduce false positive stimulation decisions and only stimulate what is needed. It contributes to the application of stimulation under these conditions. 10 In a preferred configuration, the AI inference module (14) uses specific stimulation. the relationship between parameters and previous biological and electrophysiological responses It generates candidate stimulation parameters using. Candidate parameters include frequency, amplitude, pulse width, duty cycle, excitation duration, waveform, and electrode contact 15 It may include values such as selection. The module selects the targeted parameter from among these candidate parameters. predicted to provide electrophysiological regulation and not exceeding safety limits It selects the parameter set. For security purposes in the system, 20 suggested by the artificial intelligence inference module (14) stimulation parameters undergo a control phase before being directly applied. It can be passed through. This control phase, the highest permitted current, the highest permitted voltage, pulse width limit, total load limit, load density limit, tissue impedance It can take into account parameters such as the limit and uncertainty value. In the preferred configuration. Gaussian Process Regression or a similar probabilistic 25 the safety model, the reliability of the proposed stimulation parameters or It assesses the uncertainty. High uncertainty or exceeding safety limits. Possible parameter sets are not applied, or a safe subset of parameters is used. It is transformed. Dynamic stimulation signal generator (16), central processing and control unit (13) This unit converts the stimulation commands determined by the device into an analog signal. 14 It may include a digital-to-analog converter or similar signal generation circuit. Dynamic stimulation signal generator (16), specified frequency, amplitude, pulse width and waveform It generates an analog control signal that conforms to the form and uses this signal to create voltage-controlled current. transmits to its pump (17). The voltage-controlled current pump (17) receives the signal from the dynamic stimulation signal generator (16). the analog control signal should be as free as possible from changes in tissue impedance. It converts to an independent constant current output. For this purpose, a Howland current pump or A similar constant current source structure can be used. This will improve the electrode-tissue interface. Uncontrolled changes in stimulation dose due to impedance variations 10 This is prevented. The current pump directs the stimulation current to be applied to the target brain region. It produces within safety limits. The active load balancing unit (18) distributes the electrical charge delivered to the tissue during stimulation. It is a safety structure that monitors and maintains the balance between the positive and negative phases. This 15 The unit measures the net charge transferred to the tissue after each applied stimulation pulse. It can calculate and, if necessary, apply reverse phase equalization current or equalization pulse. It can produce charge buildup at the electrode tip, leading to changes in tissue pH. The risk of electrode corrosion and long-term tissue damage is reduced. Active load balancing unit. (18) an algorithm controlled by the central processing and control unit (13), a 20 It can be configured as an analog / digital equalization circuit or a combination of both. Wireless communication module (19) for transferring data from the system to an external device, system updating settings, updating the artificial intelligence model, or patient / subject It is used for monitoring data. This module is available on tablets, computers, and clinical devices. 25 It can communicate with a programming device or data acquisition system. Wireless communication. module (19) is dependent on the external device for the basic real-time cycle of the stimulation decision. It can be structured in a way that will not lead to a situation where key decision-making and security are compromised. While the monitoring is carried out within the system, external devices are mainly used for monitoring, recording, It is used for calibration and update functions. 30 The invention's working method generally involves the following steps: from the target brain region electrophysiological signal reception, amplification and filtering of the received signal, signal processing conversion of digital data, extraction of neurophysiological features from digital signals Extraction of biomarkers or behavioral data via a biological data input interface. the collection of electrophysiological and biological data within a common state vector 5 combining stimulation parameters via an AI inference module determination of safety and load balancing criteria of the determined parameters Monitoring in terms of parameters found to be safe, and converting them to an analog stimulation signal. conversion, application to the target tissue via a constant current source, and implementation. The cycle is then repeated by receiving new electrophysiological feedback. 10 In one use case of the invention, the system can be used with an Alzheimer's model created or Regarding the hippocampal region from a subject being followed for neurodegenerative processes. It receives electrophysiological signals. Analog preprocessing and analog-to-digital converter. The central processing unit (12) converts these signals into processable digital data. 15 control unit (13), theta-gamma coupling, spectral power distribution of the signals in question. and extracts features representing signal irregularities. It also extracts biological features. data entry interface (15) for the subject’s amyloid beta, tau or genetic risk associated with Current or periodic data are transferred to the system. Artificial intelligence inference module (14), By evaluating this data together, we determine the subject's level of pathological activity and assign a score of 20 It selects the appropriate stimulation parameters. The selected stimulation parameters are dynamic after passing through a safety check. The stimulation signal is transferred to the dynamic stimulation signal generator (16). (16) produces an analog control signal in accordance with the specified command. Voltage controlled 25 The current pump (17) converts this signal into a constant current output independent of tissue impedance. It converts and transmits to the target brain region. Active load balancing unit (18), It monitors the net load delivered to the tissue during stimulation and performs phase reversal if necessary. By applying the signal, it ensures that the net load remains within safe limits. After application, bio-potential electrodes (11) produce new electrophysiological signals 30 The system detects and adjusts the stimulation parameters according to the new feedback received. It is reorganizing. 16 An important aspect of the invention is that biological data must be measured in real-time. It does not require anything. The system primarily functions in real-time closed loop. While conducting the analysis via electrophysiological signals, biological and molecular data are collected periodically. 5 Current patient profile data from clinical, laboratory or external measurement sources. They can be integrated into the system. In this way, biomarker data can influence the stimulation decision. It is used as a supporting contextual data layer. Thus, the system only not the instantaneous state of electrical activity, but the biological risk of the patient or subject and It offers a more holistic control system that also takes pathological burden into account. In alternative configurations of the invention, the artificial intelligence inference module (14) central processing and It can run embedded on the control unit (13), as well as system security an external programming device, clinical server, or cloud-based system that will not interfere with the process. It can be updated with the education system. However, stimulation is the preferred approach. The decision-making and security check are carried out within the system 15 in order to reduce delay. Fractional Fourier transform is also used for electrophysiological signal processing. instead, wavelet transform, short-time Fourier transform, or in the related technical field Another known time-frequency analysis method could be used. In another alternative configuration of the invention, the system would only target Alzheimer's disease in 20 not with specific biomarkers, Parkinson's disease, epilepsy, multiple sclerosis, dementia different biomarkers associated with types or other conditions requiring neuromodulation and It can be configured to work with electrophysiological indicators. In this case, the goal is... brain region, electrode placement, biomarkers used, and artificial intelligence model The dataset used for training can be modified according to the relevant disease. However, the basic structure of the system is 25 combined evaluation of electrophysiological feedback and biological data layers and Accordingly, it is based on the principle of applying load-balanced adaptive stimulation.
Claims
17 REQUESTS 1- The invention is based on neurobiological data relating to neurodegenerative diseases, and involves adaptive mechanisms. an adaptive closed-loop deep brain structured to provide stimulation stimulation system (10) and its feature is; 5 Detecting electrophysiological signals from the target brain region and the target brain at least one biopotential configured to transmit a stimulation signal to the region electrode (11), Raw electrophysiological data obtained from the bio-potential electrode (11) in question amplifying, filtering, and converting the signal to a digital signal 10 a structured analog preprocessor and analog-to-digital converter unit (12), to process the digital signal in question and manage the data flow within the system a central processing and control unit structured to (13), a working in relation to the central processing and control unit (13) artificial intelligence inference module (14), 15 non-electrophysiological biological, molecular, genetic or behavioral a structured biological data input interface for transferring data to the system (15), Stimulation determined by the central processing and control unit (13) 20 configured to generate analog stimulation control signals according to its commands a dynamic stimulation signal generator (16), according to the analog stimulation control signal, the target brain region a voltage-controlled system configured to generate the stimulation current to be applied current pump (17), and the net electrical charge transferred to the target tissue during stimulation is 25 It includes an active load balancing unit (18) configured for balancing. 2- Adaptive closed-loop deep brain stimulation system mentioned in Claim 1 (10) and its characteristic feature is; hippocampus, entorhinal cortex, CA1 region, CA3 region or these. electrophysiological signal from a target brain region selected from combinations 30 with its structured bio-potential electrode (11) for sensing It is the characterization of the situation. 18 3- Adaptive closed-loop deep brain stimulation system mentioned in Claim 1 (10) Its characteristic feature is; amplification of the electrophysiological signal, band-pass filtering, at least one of the following processes: noise reduction, artifact reduction, and digital sampling. It has an analog preprocessing and analog-to-digital converter unit (12) which performs 5 It is the characterization of the situation. 4- The adaptive closed-loop deep brain mentioned in any of the previous requests. The stimulation system (10) has the characteristic of theta band power from the electrophysiological signal, gamma band power, theta-gamma phase-amplitude coupling, spectral power density, signal 10 extracting neurophysiological features selected from entropy or combinations thereof It is characterized by having a central processing and control unit (13). 5- The adaptive closed-loop deep brain mentioned in any of the previous requests. The stimulation system (10) has the following characteristics: amyloid beta level, phosphorylated tau level, total 15 tau level, acetylcholine level, inflammation-related biomarker data, APOE genotype data, PSEN1 data, PSEN2 data, mRNA expression data, behavioral test data or having a biological data input interface (15) that receives combinations of these It is characterized by... 6- The adaptive closed-loop deep brain mentioned in any of the previous requests. The stimulation system (10) is characterized by its artificial intelligence inference module (14), XGBoost model, decision tree-based model, deep learning model, regression model, probabilistic characterized by containing a model or a model selected from combinations thereof. It is done. 25 7- The invention relates to the operation of an adaptive closed-loop deep brain stimulation system. It is a related method, and its characteristic is; via bio-potential electrode (11) from the target brain region electrophysiological signal acquisition, 30 19 Analog preprocessing and analog-to-digital converter of the acquired electrophysiological signal amplification, filtering and conversion of the digital signal through unit (12) transformation, neurophysiological from digital signal via central processing and control unit (13) Feature extraction, 5 non-electrophysiological biological data input interface (15) via collecting molecular, genetic or behavioral data, from the biological data input interface with the neurophysiological attributes in question (15) Combining the acquired data into a common state vector, the situation vector in question is inferred by the AI inference module (14) 10 By evaluating and determining a set of stimulation parameters, Central processing and control unit of the determined set of stimulation parameters (13) audited by [the relevant authority] in terms of safety and load balancing criteria, Dynamic stimulation signal according to the stimulation parameter set deemed safe. Generating an analog stimulation control signal by the generator (16), 15 the voltage-controlled current pump of the analog stimulation control signal in question (17) conversion of the stimulation current, stimulation current through bio-potential electrode (11) target brain application to the region, and target 20 via active load balancing unit (18) during stimulation. Balancing the net electrical charge transferred to the tissue, procedure steps It includes. 30