ICU compatible transcranial ultrasound neuromodulation system

Abstract

The application discloses an ICU compatible transcranial ultrasound neuromodulation system, a front planning and basic guarantee layer provides a target benchmark and energy basis; an initial verification layer completes transducer spatial positioning verification, acoustic transmission path quality verification and comprehensive safety initial state verification; a closed-loop control layer synchronously judges all preconditions from the front planning and basic guarantee layer and the initial verification layer, receives real-time physiological feedback from patient monitoring equipment, generates an intervention permission command or a rejection command; an execution and real-time safety layer executes ultrasound stimulation or false stimulation under the intervention permission command, and performs real-time safety monitoring; a feedback and data management layer collects post-stimulation multi-modal feedback, performs closed-loop parameter adaptive correction, provides operator interaction and override, and archives whole-process data, improves transcranial ultrasound delivery repeatability, reduces risks brought by unstable coupling or unsafe physiological states, and realizes adaptive neuromodulation under closed-loop safety constraints.

Description

Technical Field

[0001] This invention belongs to the field of closed-loop control technology, specifically relating to an ICU-compatible transcranial ultrasound neuromodulation system. It is suitable for critically ill patients experiencing sleep fragmentation, sleep-wake rhythm disorders, abnormal awakenings, sedation-related sleep structure disruption, ventilator-related sleep interruption, or other ICU sleep disorders. It can be used for clinical research, sham-stimulation control studies, neurophysiological studies, and sleep state-dependent neuromodulation programs. Background Technology

[0002] Sleep disruption is a common and clinically significant problem in the intensive care unit (ICU). ICU patients are exposed to mechanical ventilation, alarm noise, nursing interventions, pain, sedatives, analgesics, light interference, risk of delirium, hemodynamic instability, and respiratory instability. These factors can lead to severe sleep fragmentation and circadian rhythm alterations. Poor sleep in ICU patients can further interact with delirium, inflammation, anxiety, ventilator intolerance, and delayed recovery.

[0003] Existing methods for improving sleep in the ICU mainly include environmental modifications, noise and light reduction, earplugs, eye masks, medication adjustments, sedative drug selection, melatonin, early mobilization, pain management, and optimization of nursing workflows. These methods are clinically useful, but they are generally not brain state-specific, cannot adapt to real-time sleep physiology, and are not designed to deliver targeted neuromodulation when poor sleep conditions are detected.

[0004] Non-invasive brain stimulation techniques have been investigated for sleep regulation and neuropsychiatric applications. Transcranial ultrasound neuromodulation is attractive because its acoustic energy can be directed through the skull to cortical or deeper neural targets, and its spatial specificity is higher than many surface electrical or magnetic stimulation methods. However, conventional transcranial ultrasound research systems are often not optimized for ICU environments. Laboratory transducers, either handheld by the operator or fixed by a simple mechanical frame, are not suitable for prolonged ICU monitoring, mechanically ventilated patients, repetitive sleep-dependent stimulation, or real-time safety gating control.

[0005] Currently, transcranial ultrasound (TCE) devices face several practical limitations in ICU translation. First, maintaining stable acoustic coupling over extended periods is difficult due to head movements, sweating, hair growth, dressing changes, nursing procedures, and bed position variations. Second, actual transcranial exposure is affected by skull thickness, skull curvature, target location, transducer location, and coupling status. Third, vulnerable ICU patients may experience sudden changes in oxygenation, blood pressure, ventilation, rhythm, arousal, neurological status, or sedation levels; therefore, the device must be able to suppress or terminate stimulation under unsafe conditions. Fourth, sham stimulation and blinding are crucial for clinical trials, but many devices do not provide standardized sham patterns that match appearance, sound, vibration, and operator workflow. Fifth, clinical translation requires medical-grade documentation, audit trails, parameter logging, and software version rollback.

[0006] Therefore, there is a need for a bedside transcranial ultrasound neuromodulation device that integrates stable positioning, acoustic coupling monitoring, patient-specific targeting, safety monitoring, closed-loop command input, adaptive parameter control, sham stimulation, manual overdrive, and safety event recording. Such a device should be compatible with the ICU environment and should collaborate with sleep interruption detection systems and safety-gated closed-loop controllers. Summary of the Invention

[0007] To address the aforementioned technical problems, the present invention provides an ICU-compatible transcranial ultrasound neuromodulation system, comprising the following steps:

[0008] The pre-planning and basic support layer includes a patient-specific targeting and dose planning unit and a power and equipment status management unit to provide target baselines and energy bases; The initial verification layer includes the initial verification sub-functions of the bedside positioning and fixing unit, the acoustic coupling unit, and the safety monitoring unit, which are used to complete the transducer spatial positioning verification, acoustic transmission path quality verification, and comprehensive safety initial state verification, respectively. The closed-loop control layer includes a closed-loop control interface for synchronously deciding on all preconditions from the pre-planning and basic support layer and the initial verification layer, and receiving real-time physiological feedback from the patient monitoring device to generate intervention permission or rejection commands. The execution and real-time safety layer includes real-time monitoring sub-functions of ultrasound output / stimulation program, therapeutic transcranial ultrasound equipment, transcranial ultrasound transducer, sham stimulation / blinding unit, and safety monitoring unit, as well as alarm / emergency stop / fail-safe unit, for performing ultrasound stimulation or sham stimulation under the intervention permission command and performing real-time safety monitoring; The feedback and data management layer includes a post-intervention feedback unit, a clinician interface, and a security database / audit trail. It is used to collect multimodal feedback after stimulation, perform closed-loop parameter adaptive correction, provide operator interaction and control, and archive data throughout the entire process.

[0009] Preferably, the patient-specific targeting and dose planning unit is configured to: determine the ultrasound target coordinates, acoustic path, incident angle, skull transmission compensation and exposure parameters based on images, anatomical landmarks or physiological responses, and output them to the closed-loop control interface and ultrasound output / stimulation program.

[0010] Preferably, the power supply and device status management unit is configured to: collect power supply voltage and current consumption data of each submodule in real time, perform power quality analysis and ripple detection, and generate a global power health status word containing three levels of identifiers: "normal / degraded / failed"; at the same time, it aggregates the operating status of each device to form a system status topology diagram, and periodically sends the status word to the closed-loop control interface and the safety monitoring unit.

[0011] Preferably, the bedside positioning and fixing unit is configured to: maintain a stable spatial relationship between the transcranial ultrasound transducer and the patient's head, and detect the displacement of the transducer relative to the head; its real-time pose data is obtained through multi-sensor fusion, and the offset is compared with a preset threshold to generate a position deviation flag; this flag serves as one of the gating conditions, and an enable signal is output to the transducer drive circuit only when all safety gate conditions are met.

[0012] Preferably, the acoustic coupling unit is configured between the transducer and the scalp. It acquires multimodal signals through sensors in a disposable hydrogel coupling pad, calculates the comprehensive coupling quality score and compares it with an acceptable threshold to generate a binary coupling state verification signal. During ultrasonic delivery, it continuously monitors the dynamic changes in coupling quality and generates a coupling failure indicator when the quality decreases or changes abruptly.

[0013] Preferably, the safety monitoring unit performs initial safety data aggregation and initial vital sign status acquisition at the initial verification layer to generate an initial safety clearance comprehensive flag; it performs hardware-level interrupt path independent of the closed-loop control interface for executing and real-time safety layer data processing; it performs multi-level threshold comparison on multi-source safety signals, classifying anomalies into warning level, suppression level and shutdown level, and for shutdown level events, it directly triggers the shutdown execution circuit through an independent hardware interrupt signal line.

[0014] Preferably, the closed-loop control interface is configured as follows: with prerequisite condition synchronous verification as a premise, it receives multimodal physiological signals from the patient monitoring device, calculates the sleep instability index and confidence score, and generates an intervention permission command when the index exceeds the threshold, the confidence score is sufficient, and the duration is predetermined; the command includes an operation mode identifier and a set of planned parameters, and is simultaneously sent to the ultrasound output / stimulation program, the sham stimulation / blinding unit, and the safety database; after the command is sent, a local safety check is performed again, and only after passing the check can the system enter the ready-to-launch state.

[0015] Preferably, the ultrasound output / stimulation program is configured to: parse the operation mode identifier in the intervention permission command; if it is "active stimulation", then generate a digital waveform sample sequence using a direct digital frequency synthesis algorithm according to the plan parameters; if it is "sham stimulation", then route the drive signal to the virtual load and synchronously generate a timing control signal and sensory simulation signal that are exactly the same as those for active stimulation.

[0016] Preferably, the therapeutic transcranial ultrasound device is configured to: receive digital waveform sample sequences, perform digital-to-analog conversion, power amplification and channel allocation, apply phase delay to the multi-element phased array transducer to achieve electronic focus deflection; and collect voltage, current and reverse power data of the output channel in real time, and send them to the safety monitoring unit through an independent data link.

[0017] Preferably, the sham stimulation / blinding unit is configured as follows: in sham stimulation mode, the driving signal is connected to the virtual load without generating transcranial ultrasound; indicator lights, cooling fan noise and tactile vibration signals matching active stimulation are generated simultaneously to maintain the same operating procedure as active stimulation; sham stimulation events are encrypted and stored in the same recording format as active stimulation, and the operator interface only displays the unified status information of the "stimulation event".

[0018] This invention provides an ICU-compatible transcranial ultrasound neuromodulation system. The system comprises a pre-planning and basic support layer, including a patient-specific targeting and dosage planning unit and a power and equipment status management unit, providing target references and energy bases; an initial verification layer, including bedside positioning and fixation units, acoustic coupling units, and safety monitoring units, for performing transducer spatial positioning verification, acoustic transmission path quality verification, and comprehensive safety initial state verification, respectively; and a closed-loop control layer, including a closed-loop control interface, for synchronously adjudicating all pre-conditions from the pre-planning and basic support layer and the initial verification layer, receiving real-time physiological feedback from patient monitoring equipment, and generating intervention permission or rejection commands; and execution... The system includes a real-time safety layer, comprising real-time monitoring sub-functions for ultrasound output / stimulation programs, therapeutic transcranial ultrasound equipment, transcranial ultrasound transducers, sham stimulation / blinding units, and safety monitoring units, as well as alarm / emergency stop / fail-safe units. These sub-functions are used to execute ultrasound stimulation or sham stimulation under the intervention permission command and to perform real-time safety monitoring. The feedback and data management layer includes a post-intervention feedback unit, a clinician operation interface, and a safety database / audit trail. This layer is used to collect multimodal feedback after stimulation, perform closed-loop parameter adaptive correction, provide operator interaction and control, and archive full-process data. This improves the repeatability of transcranial ultrasound delivery, reduces the risks associated with unstable coupling or unsafe physiological states, and achieves adaptive neuromodulation under closed-loop safety constraints. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the architecture of an ICU-compatible transcranial ultrasound neuromodulation system; Figure Labels 1-ICU patients and bedside environment; 2-Bedside positioning and fixation unit; 3-Transcranial ultrasound transducer; 4-Acoustic coupling unit; 5-Patient-specific targeting and dose planning unit; 6-Safety monitoring unit; 7-Closed-loop control interface; 8-Ultrasound output / stimulation program; 9-Therapeutic transcranial ultrasound device; 10-Sham stimulus / blinding unit; 11-Post-intervention feedback unit; 12-Clinical physician operation; 13-Security Database / Audit Trail; 14-Power and Equipment Status Management Unit; 15 - Alarm / Emergency Stop / Fail-Safe Unit. Detailed Implementation

[0020] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0021] This invention provides an ICU-compatible transcranial ultrasound neuromodulation system, applied in ICU patients and bedside environment 1. The system includes: First layer: Pre-planning and basic support layer The first layer includes a patient-specific targeting and dose planning unit 5 and a power and equipment status management unit 14. Both operate in a parallel initialization mode, providing target references and energy bases for all subsequent operations of the system.

[0022] The patient-specific targeting and dose planning unit 5 is configured to determine the ultrasound target coordinates, acoustic path, incident angle, skull transmission compensation, and exposure parameters based on imaging, anatomical landmarks, or physiological responses. The data processing procedure is as follows: Step S101: Data Acquisition and Registration. The unit acquires patient head image data through its data interface, including one or more of preoperative CT, MRI, or ultrasound imaging. If the image data is available, three-dimensional reconstruction and skull boundary extraction are performed, and skull thickness distribution and density field are calculated; if the image data is unavailable, the unit switches to conservative target template mode, uses head surface landmarks and skull landmarks to estimate target position, and automatically generates output intensity upper limit reduction markers and duty cycle reduction markers.

[0023] Step S102: Target Coordinate Calculation. The unit registers the image coordinate system or anatomical landmark coordinate system with the transducer spatial coordinate system provided by the bedside positioning and fixation unit 2, and determines the three-dimensional coordinates of one or more stimulation targets through reference markers or neuronavigation coordinate transformation. The stimulation targets include one or more of the following regions: prefrontal cortex, medial prefrontal cortex, dorsolateral prefrontal cortex, thalamus, hypothalamus, hippocampus, brainstem arousal network, default mode network nodes, salience network nodes, or individualized sleep-wake network regions.

[0024] Step S103: Acoustic Path Planning and Dosage Calculation. Based on target depth, skull thickness estimation, and acoustic simulation algorithms, the unit calculates the attenuation coefficient, focal point offset, and relative exposure distribution for transcranial transmission. On this basis, it calculates the required carrier frequency, sound intensity, pulse duration, duty cycle, and pulse repetition frequency, among other planned exposure parameters, and generates a set of planned ultrasound transmission parameters. When image data is unavailable, the unit automatically reduces the upper limit of output intensity or limits the duty cycle, requiring clinicians to confirm via the clinician's interface.

[0025] Step S104: Data Output and Storage. The unit outputs a planning dataset containing target coordinates, acoustic path, incident angle, skull attenuation compensation coefficient, and planned exposure parameters to the closed-loop control interface 7 as a baseline database for stimulation decisions; simultaneously, it sends the skull attenuation compensation coefficient to the ultrasound output / stimulation program for waveform pre-compensation calculation. The unit also adds a timestamp to the planning dataset and sends it to the security database / audit trail 13 for storage.

[0026] The power supply and equipment status management unit 14 is configured to manage the power supply and operating status of all sub-modules of the system, and its data processing procedure is as follows: Step S201: Multi-channel power parameter acquisition. The unit acquires the power supply voltage and current consumption data of the transcranial ultrasound transducer, acoustic coupling unit, ultrasound output / stimulation program, therapeutic transcranial ultrasound equipment, and clinician operation interface in real time through an analog-to-digital converter at a predetermined sampling rate. If the system includes a battery power supply module, the remaining battery capacity and charge / discharge cycle count are also acquired simultaneously.

[0027] Step S202: Power Quality Analysis and Status Word Generation. The unit performs digital filtering and ripple analysis on the collected voltage and current data, comparing it with preset voltage tolerance ranges and current limits. If a voltage transient drop, overcurrent, or excessive ripple is detected, a power abnormality flag for the corresponding channel is generated. The unit integrates the status of each channel to generate a global power health status word, which includes three levels of identification: "Normal / Degraded / Failure".

[0028] Step S203: Device Status Aggregation and Topology Construction. The unit receives the operating temperature and impedance status of the transcranial ultrasound transducer, the sensor self-test status of the acoustic coupling unit, the cooling fan status and power amplifier temperature of the therapeutic transcranial ultrasound device, and summarizes them to form an overall system status topology diagram.

[0029] Step S204: Status Flag Transmission and Anomaly Warning. The unit periodically sends the global power health status word to the closed-loop control interface and the safety monitoring unit. If the global power health status word is not "normal," a power warning flag is sent to the alarm / emergency stop / fail-safe unit 15 via an independent signal line; simultaneously, power and status data are periodically uploaded to the safety database / audit trail 13 for log maintenance and equipment life assessment. The system can only proceed to the subsequent ultrasonic transmission process if the global power health status word is in a "normal" state.

[0030] The second layer includes the initial verification sub-functions of the bedside positioning and fixation unit 2, the acoustic coupling unit 4, and the safety monitoring unit 6. The three operate in parallel to complete the transducer spatial positioning verification, acoustic transmission path quality verification, and comprehensive safety initial state verification, respectively, providing the necessary conditions for the stimulation permission judgment of the closed-loop control interface 7.

[0031] The bedside positioning and fixation unit 2 is configured to maintain a stable spatial relationship between the transcranial ultrasound transducer and the head of the ICU patient during ICU monitoring, and to detect the displacement of the transcranial ultrasound transducer relative to the patient's head. The bedside positioning and fixation unit 2 includes a headband, helmet, cap, pillow support, bed-mounted robotic arm, ultrasound driver, power amplifier, microcontroller or embedded computer, safety interface board, emergency stop circuit, data acquisition interface and network interface, articulated robotic arm, adhesive positioning pad, skull landmark frame, or any combination thereof. The data processing procedure is as follows: Step S301: Initial Position Calibration. The unit acquires the three-dimensional initial coordinates and attitude angles of the transcranial ultrasound transducer relative to the patient's head coordinate system through manual input by the operator or automatic registration. The initial coordinates are determined by readings from the skull landmark frame, adhesive positioning pad, or joint encoder of the bed-mounted robotic arm.

[0032] Step S302: Real-time pose data acquisition. The unit continuously acquires the spatial pose signal of the transducer at a sampling rate of not less than 100Hz through displacement sensors built into the fixed mechanism. The sensors include any combination of contact pressure sensors, inertial measurement units, optical motion sensors, magnetic position sensors, strain sensors, or encoders. The microcontroller or embedded computer performs data fusion and Kalman filtering on the acquired multi-sensor signals, removes measurement noise, and generates the current pose estimate.

[0033] Step S303: Offset Calculation and Threshold Comparison. The unit performs a difference calculation between the current pose estimate and the initial coordinates recorded in step S301 to calculate the three-dimensional translational offset and rotational offset angle of the transducer relative to the initial reference point. The offset is compared with preset position tolerance thresholds and angle tolerance thresholds. If either offset exceeds the corresponding threshold, a position deviation flag is generated and the local status register is updated.

[0034] Step S304: Position Data Distribution and Gating Output. The unit sends the calibrated actual transducer pose data to the safety monitoring unit 6 and the closed-loop control interface 7 in real time through the data acquisition interface and network interface, serving as a contact safety criterion and space gating condition. Simultaneously, the unit receives sleep state and sleep instability commands from the closed-loop control interface 7, and only outputs an enable signal to the transducer drive circuit when all local safety gates (including the position deviation flag being "not out of tolerance" and other preset mechanical safety conditions) are met.

[0035] Step S305: Actuator adjustment command generation. For implementations equipped with an active articulated robotic arm or adjustable support, the unit generates an adjustment drive signal based on the offset calculation results, and maintains the predetermined spatial relationship between the transducer and the patient's head through closed-loop servo control.

[0036] The acoustic coupling unit 4 is configured between the transcranial ultrasound transducer and the patient's scalp. It collects data through sensors in a disposable hydrogel coupling pad, calculates the coupling quality fraction based on the collected data, verifies the acoustic coupling and device integrity, and provides controllable transcranial acoustic transmission and coupling status monitoring.

[0037] The acoustic coupling unit 4 further includes a water-filled membrane, a degassed water cavity, a disposable coupling box, a soft silicone acoustic window, a conformal coupling pad, or a sterile disposable interface. The disposable hydrogel coupling pad includes an acoustic window aligned with the aperture of the transcranial ultrasound transducer. The disposable hydrogel coupling pad contains any combination of sensors, including a contact pressure sensor, an inertial measurement unit, an optical motion sensor, a magnetic position sensor, a strain sensor, or an encoder. The acoustic coupling unit includes at least one of a pressure sensor, a temperature sensor, a bubble detection element, an impedance measurement element, an optical coupling indicator, or a liquid level sensor. The data processing procedure is as follows: Step S401: Multimodal sensor signal acquisition. The unit begins data acquisition before the ultrasonic transducer is activated, acquiring the following analog signals through a sensor array configured inside the disposable hydrogel coupling pad and along the acoustic path: pressure distribution signal output from the contact pressure sensor, acoustic impedance signal output from the impedance measurement element, scalp interface temperature signal output from the temperature sensor, bubble presence signal output from the bubble detection element, and coupling pad deformation signal output from the strain sensor. All signals are converted from analog to digital to generate a digitized multi-channel sensor dataset.

[0038] Step S402: Signal preprocessing and feature extraction. The unit performs digital filtering (including low-pass filtering to remove high-frequency noise and notch filtering to suppress power frequency interference), outlier removal, and mean smoothing on the multi-channel sensor dataset. Based on this, the following feature parameters are extracted: average contact pressure and uniformity index, acoustic impedance matching coefficient, interface temperature, bubble count, and strain consistency index.

[0039] Step S403: Calculation of coupling quality score. The unit inputs the feature parameters extracted in step S402 into a predetermined weighted fusion algorithm to calculate the comprehensive coupling quality score. The score is a continuous value between 0 and 100, which comprehensively reflects pressure uniformity, acoustic impedance matching, coupling layer integrity, and temperature safety margin.

[0040] Step S404: Coupling Acceptability Judgment. The unit compares the calculated coupling quality score with a preset coupling acceptability threshold (preferably ≥85 points) to generate a binary coupling state verification signal, i.e., "coupling acceptable" or "coupling unacceptable". If any single feature parameter (such as temperature, bubble content) exceeds the independent safety boundary, it is directly judged as "coupling unacceptable" regardless of the overall score.

[0041] Step S405: Status Signal Distribution and Continuous Monitoring. The unit sends the coupling status verification signal to the closed-loop control interface in real time as an acoustic transmission gating condition. Simultaneously, it sends the coupling quality score and various original characteristic parameters to the safety monitoring unit and the clinician's operating interface for display. During ultrasound delivery, the unit continuously monitors the dynamic changes in coupling quality at a refresh rate of no less than 10Hz. If the coupling quality score drops below the threshold or a sudden change occurs, a coupling failure flag is immediately generated and sent to the safety monitoring unit 6 and the alarm / emergency stop / fail-safe unit 15, respectively.

[0042] Safety monitoring unit 6 is configured to perform the following data processing steps during the initial verification phase of the second layer, based on coupling status, transcranial acoustic transmission, acoustic output status, scalp temperature, contact pressure, patient movement, cardiopulmonary stability, neurological safety, and clinician-defined commands to inhibit or terminate ultrasound output: Step S501-A: Initial safety data aggregation. The unit synchronously receives the transducer position deviation flag and actual pose data output by the bedside positioning and fixation unit 2, the coupling status verification signal and coupling quality score output by the acoustic coupling unit, the global power health status word output by the power supply and equipment status management unit, and the reference impedance value of the transcranial ultrasound transducer in standby mode.

[0043] Step S502-A: Initial acquisition of vital signs. The unit acquires the patient's current heart rate, blood pressure, blood oxygen saturation, respiratory rate, and EEG background activity data through the monitoring equipment interface, and compares them with the preset acceptable physiological range for the ICU to generate baseline vital sign status indicators.

[0044] Step S503-A: Initial Safety Clearance Synthesis Flag Generation. The unit performs a logical AND operation on all status flags obtained in steps S501-A and S502-A to generate an initial safety clearance synthesis flag. This flag is only output as "cleared" when the position deviation flag is "not out of tolerance," the coupling status verification signal is "acceptable," the global power health status word is "normal," the transducer reference impedance is within the normal range, and all vital sign baseline status flags are "within range." This synthesis flag is sent to the closed-loop control interface via an independent data link as a prerequisite for allowing subsequent stimulus decisions.

[0045] Step S504-A: Handling of the initial abnormal state. If the initial safety clearance integrated flag is "not permitted", the safety monitoring unit generates the corresponding abnormal cause code (including one or more of the following: location abnormality, coupling abnormality, power supply abnormality, transducer abnormality, or vital sign abnormality), sends it to the clinician operation interface 12 for display, and records it to the safety database / audit trail 13.

[0046] The third layer contains a closed-loop control interface, serving as the system's globally unique decision-making center. This layer synchronously adjudicates all preconditions from the first and second layers, receives real-time physiological feedback from patient monitoring equipment, generates a final intervention permission or denial command, and determines whether the system enters active stimulation mode or sham stimulation mode.

[0047] The closed-loop control interface 7 is configured to receive commands from the safety monitoring unit 6, determine whether the patient is in a candidate sleep interruption state, and only provide an intervention permission command when the sleep instability index exceeds a threshold and the confidence score is sufficient. Subsequently, a local safety check is performed. If all checks pass, the ultrasound frequency, intensity, pulse duration, duty cycle, pulse repetition frequency, ultrasound emission duration, stimulation interval, and target selection are adjusted to prepare the transcranial ultrasound transducer. If any check fails, the stimulation command is rejected, the reason is stored, and the clinician is notified of operation 12. The data processing procedure of the closed-loop control interface 7 is as follows: Step S601: Precondition synchronization verification. The closed-loop control interface 7 polls the following input data at a predetermined period (preferably 1 second): The planning dataset from patient-specific targeting and dosing planning unit 5 confirmed that the target coordinates and planning parameters were not outdated and that the confidence score was higher than the predetermined confidence threshold. Real-time transducer pose data and position deviation flags from bedside positioning and fixing unit 2 confirm that the positioning validity flag is "valid"; The coupling status verification signal and real-time coupling quality score from acoustic coupling unit 4 confirmed that the coupling was "acceptable". The initial safety clearance indicator from safety monitoring unit 6 confirms the "permitted" status. The global power health status word from the power and device status management unit 14 is confirmed as "normal".

[0048] If any of the above conditions are not met, the closed-loop control interface 7 generates a rejection command, carrying the category code of the unmet condition, and sends it to the clinician operation interface 12 to display the reason for rejection and record it in the security database / audit trail 13; the process terminates and does not proceed to subsequent steps.

[0049] Step S602: Determining Sleep Instability. If all preconditions in step S601 are met, the closed-loop control interface 7 initiates the sleep instability index calculation process. The unit receives multimodal physiological signals from the patient monitoring device, including EEG power spectrum data (especially the ratio of delta wave to theta wave power), time-domain and frequency-domain indicators of heart rate variability, respiratory pattern parameters, and body movement signals. The unit inputs the above physiological characteristics into a pre-deployed sleep state classification model to calculate the sleep instability index (a continuous value from 0 to 100) and the corresponding confidence score. The sleep instability index is compared with a predetermined threshold (preferably ≥65), and it is confirmed that the confidence score is higher than the predetermined confidence threshold (preferably ≥0.8).

[0050] Step S603: Intervention permission command generation. If the sleep instability index exceeds the threshold and the confidence score is sufficient in step S602, the closed-loop control interface 7 further determines whether the state continues for a predetermined duration (preferably 3 consecutive sampling cycles, each cycle being 1 minute). If the duration condition is met, an intervention permission command is generated, which includes the following data fields: Operating mode identifier: Assign the value "active stimulation" or "sham stimulation" according to the experimental protocol or clinical settings; Planning parameter set: ultrasound frequency, intensity, pulse duration, duty cycle, pulse repetition frequency, emission interval, and target coordinates extracted from the planning dataset of patient-specific targeting and dose planning unit 5 or adjusted according to preset adaptive rules.

[0051] If step S602 fails, the closed-loop control interface 7 records the "stimulation threshold not reached" status to the safety database and reassesses it in the next cycle, without generating an intervention permission command during this period.

[0052] Step S604: Mode Arbitration and Parameter Issuance. Closed-loop control interface 7 simultaneously sends an intervention permission command containing the operating mode identifier and the planned parameter set to: Ultrasound output / stimulation program 8: used for digital waveform generation; False stimulus / blinding unit 10: used for pattern matching and cue synchronization; Security Database / Audit Trail 13: Used to record decision-making basis and timestamps.

[0053] Step S605: Local Safety Check and Final Readiness. Simultaneously with the intervention permission command being issued to the ultrasound output / stimulation program, the closed-loop control interface 7 performs another local safety check, including coupling verification, temperature check, transducer position check, equipment integrity check, patient motion check, and patient vital signs check. If all checks pass, the closed-loop control interface 7 generates a final readiness signal, putting the transcranial ultrasound transducer into a ready-to-fire state; if any check fails, the closed-loop control interface 7 immediately withdraws the intervention permission command, generates a rejection stimulation command, stores the reason, and notifies the clinician to proceed.

[0054] The fourth layer includes an ultrasound output / stimulation program 8, a therapeutic transcranial ultrasound device 9, a transcranial ultrasound transducer 3, a sham stimulation / blinding unit 10, a safety monitoring unit 6 (real-time monitoring sub-function), and an alarm / emergency stop / fail-safe unit 15, forming an execution main chain of "digital waveform generation → analog power amplification → transducer acoustic emission". At the same time, the safety monitoring unit runs in parallel with a hardware-level interrupt path independent of the decision layer, realizing the selective execution of active stimulation and sham stimulation and real-time safety assurance.

[0055] Ultrasonic output / stimulation program 8 is configured to receive stimulation commands from the closed-loop control interface, generate corresponding digital drive waveforms, and perform output mode routing control. The data processing procedure is as follows: Step S701: Command Parsing and Pattern Recognition. The program receives the intervention permission command from the closed-loop control interface 7 and parses the operation mode identifier within it. If the operation mode identifier is "active stimulus," the program will execute the path of step S702-A; if it is "sham stimulus," the program will execute the path of step S702-B. The two paths are mutually exclusive, and only one path is active at a time.

[0056] Step S702-A (Active Stimulation Path): Digital Waveform Parameter Calculation and Generation. The program extracts the carrier frequency, pulse repetition frequency, number of cycles per pulse train, duty cycle, ultrasound emission duration, rise time, fall time, and emission interval parameters from the intervention permission command. Based on these parameters, the program uses a direct digital frequency synthesis algorithm to generate a digital waveform sample sequence. The waveform sample sequence contains pulse trains arranged at predetermined time intervals, each pulse train containing a sine wave period of the carrier frequency, and there are silent intervals with no output between pulse trains. The ultrasound emission duration is adjustable from 0.1 seconds to 600 seconds, preferably from 1 second to 300 seconds.

[0057] Step S702-B (Sham Stimulation Path): The drive signal is routed to the virtual load. If the operating mode identifier is "Sham Stimulation", the program does not output a valid drive waveform to the therapeutic transcranial ultrasound device, but instead routes the drive signal to the internal virtual load of the sham stimulation / blinding unit; at the same time, the program generates timing control signals (including emission start time, duration, and end time) that are exactly the same as those in active stimulation mode, to ensure that the transducer positioning status, warning lights, audible prompts, and operator perception are consistent with those in active stimulation mode in sham stimulation mode.

[0058] Step S703: Control signal output and synchronization. The program sends the generated digital waveform sample sequence (active path) or virtual load control sequence (dummy path) to the therapeutic transcranial ultrasound device according to precise timing requirements, and at the same time sends the synchronization clock signal to the dummy stimulation / blinding unit 10 to ensure the consistency of operation prompts with the expected transmission timing.

[0059] Step S704: Output Parameter Recording and Feedback. During transmission, the program records the actual generated waveform parameters (including actual carrier frequency, actual output duration, and actual number of pulses) and sends them to the security database / audit trail. After transmission, the program packages the total transmission duration and the actual delivered parameter set and sends it to the post-intervention feedback unit.

[0060] The therapeutic transcranial ultrasound device 9 is configured to receive the digital waveform generated by the ultrasound output / stimulation program 8, perform digital-to-analog conversion, power amplification and channel allocation, and drive the transcranial ultrasound transducer to output ultrasound energy. The data processing procedure is as follows: Step S801: Device Self-Test and Calibration Confirmation. Upon power-up and before each drive signal reception, the device performs an initialization self-test, reading internally stored calibration coefficients, power amplifier temperature, cooling fan status, and output channel impedance matching status to generate a device health status code. If the health status code indicates an abnormality in any submodule, the device sends a device fault flag to the safety monitoring unit and the alarm / emergency stop / fail-safe unit, and refuses to receive drive signals.

[0061] Step S802: Digital Waveform Reception and Analog-to-Digital Conversion. The device receives a sequence of digital waveform samples from the ultrasound output / stimulation program 8, converts it into an analog voltage waveform signal through a built-in high-speed digital-to-analog converter, and forms the power amplifier input signal after pre-amplification and filtering.

[0062] Step S803: Power Amplification and Channel Allocation. Based on the target coordinates and spatial allocation instructions in the intervention permission command, the device routes the amplified high-voltage electrical signal to the corresponding array element channel. For multi-element phased array transducers, the device further applies a time delay to each channel signal according to preset phase delay parameters to achieve electronic focus deflection. The device's built-in current-limiting and temperature-limiting circuits monitor the output current and power amplifier temperature in real time; if they exceed the safety limit, they automatically clamp the circuit.

[0063] Step S804: Real-time output status feedback. The device collects voltage, current, and reverse power data of each output channel in real time at a sampling rate of not less than 1kHz, calculates the actual output power and load impedance, and simultaneously sends the above data to the safety monitoring unit and the transcranial ultrasound transducer via an independent data link. The device also sends the accumulated power-on time and temperature data to the power supply and device status management unit 14.

[0064] The transcranial ultrasound transducer 3 is configured to deliver low-intensity pulsed transcranial ultrasound to one or more sleep-related brain regions of the patient based on the displacement of the transcranial ultrasound transducer relative to the patient's head. The transcranial ultrasound transducer includes a single-element transducer, a multi-element phased array transducer, a ring array transducer, a linear array transducer, a matrix array transducer, a phased array transducer, or any combination thereof. The transcranial ultrasound transducer is configured to operate at a carrier frequency in the range of 0.1 MHz to 5.0 MHz, preferably 0.2 MHz to 3.0 MHz, more preferably 0.25 MHz to 1.5 MHz for transcranial neuromodulation in ICU use. The transcranial ultrasound transducer is only activated when the sleep instability index exceeds a predetermined or adaptive threshold for a predetermined duration and all safety gating status indications allow it. The data processing procedure is as follows: Step S901: Drive signal reception and electroacoustic conversion. The transducer receives the high-voltage electric drive signal from the therapeutic transcranial ultrasound device, converts the electrical energy into mechanical vibration energy through the piezoelectric array element, generates ultrasound waves with the same frequency as the drive signal, and radiates them to the target area of ​​the patient's scalp and skull through the acoustic coupling unit.

[0065] Step S902: Element Status Self-Sensing. The transducer's built-in temperature, voltage, and current sensors collect the real-time operating temperature, RMS value of the driving voltage, and RMS value of the operating current of each element at a sampling rate of no less than 100Hz. For multi-element transducers, the above parameters are collected independently for each element channel.

[0066] Step S903: Element status data reporting. The transducer simultaneously sends the element status parameters collected in step S902 to: Safety monitoring unit: used for dynamic safety gating and overheat protection judgment; Closed-loop control interface: used for recording and calibrating actual delivery dose and transducer operating status.

[0067] Step S904: Cumulative Usage Time Recording. The transducer's internal control chip cumulatively records the total duration of ultrasonic emission and the number of times each element is powered on, for lifespan prediction and periodic calibration reminders. This data is sent to the security database / audit trail.

[0068] The sham stimulation / blinding unit 10 is configured to reproduce the appearance, device operation, patient contact status, acoustic cues, vibrational cues, visual cues, timing, or operator workflow of active stimulation while delivering ineffective therapeutic ultrasound energy or exposure levels below a predetermined neuromodulation threshold, generating a non-therapeutic control condition with matching appearance, sound, vibration, and operational workflow. The sham stimulation unit can route drive signals to a virtual load or deliver subthreshold acoustic output. Matching low-level vibrational or acoustic cues can be generated via auxiliary components to improve blinding. The device records sham stimulation events using the same event pattern as active stimulation events.

[0069] The data processing procedure for the sham stimulus / blinding unit is as follows: Step S1001: Mode Reception and State Switching. The unit receives the intervention permission command issued by the closed-loop control interface and parses the operation mode identifier within it. When the identifier is "sham stimulus", the unit switches to sham stimulus operating mode; when the identifier is "active stimulus", the unit switches to active monitoring mode (at this time, only the state is recorded, and the output path is not intervened).

[0070] Step S1002 (Sham Stimulation Mode): Virtual Load Driving and Signal Absorption. The unit connects the driving signal from the ultrasound output / stimulation program to the internal hydrophone load or resistive virtual load, absorbing all electrical signal energy and not generating transcranial ultrasound waves.

[0071] Step S1003 (Sham Stimulation Mode): Sensory Simulation Signal Generation. To ensure the blinding effect, the unit synchronously generates the following simulation signals that perfectly match the active stimulation: drive status indicator control signal (color and flashing pattern consistent with the active stimulation), internal cooling fan speed control signal (generating the same auditory cues as the active stimulation), and optional tactile vibration signal (simulating the working state through low-level vibration of the transducer housing). The timing and parameters of all simulation signals are consistent with the active stimulation standard template recorded in step S1004.

[0072] Step S1004 (Sham Stimulation Mode): Maintaining consistency in the operational process. Throughout the sham stimulation process, the unit maintains the same transducer positioning, coupling pad contact timing, warning light sequence, and operator workflow as during active stimulation, ensuring that the blinded operator cannot distinguish between sham stimulation and active stimulation events through sight, hearing, or touch.

[0073] Step S1005: Event Logging and Encrypted Storage. The unit generates event logs for sham stimulus events using the exact same logging format as active stimulus events. These logs include the event type (marked as "sham stimulus" but hidden from the external interface; the real pattern identifier is only stored in the encrypted audit log), timestamp, simulation parameters, and duration. This record is sent to the secure database / audit trail via an encrypted channel. The operator interface only displays the unified status information of the "stimulus event," achieving blinding.

[0074] During the real-time execution phase of the fourth layer, the safety monitoring unit 6 operates continuously at a sampling rate higher than the transmission frequency. Its data processing has a hardware-level interrupt path independent of the closed-loop control interface, and its priority is higher than all software decisions.

[0075] The safety monitoring unit 6 performs the following data processing steps during the real-time monitoring phase: Step S501-B: Parallel acquisition of multi-source security signals. The security monitoring unit 6 simultaneously acquires the following input data at a sampling rate of not less than 1 kHz: Acoustic coupling unit 4: Real-time coupling of mass fraction, scalp interface temperature, and bubble detection status; Bedside positioning and fixing unit 2: Real-time position and orientation offset and position deviation indicator of the transducer; Transcranial ultrasound transducer 3: operating temperature, effective value of driving current, and impedance value of each element; Therapeutic transcranial ultrasound equipment 9: power amplifier temperature and output power; Patient monitoring equipment: heart rate, blood oxygen saturation, respiratory rate, EEG signal, blood pressure waveform; Clinical physician operation interface 12: Manual pause or over-control command.

[0076] Step S502-B: Multi-level threshold comparison and anomaly detection. The unit inputs the parameters collected in step S501-B into multiple independent comparators in parallel, comparing them with the corresponding upper and lower safety limits. The safety thresholds include, but are not limited to: upper limit of scalp temperature (preferably 41℃), lower and upper limits of contact pressure (preferably 5-50kPa), upper limit of transducer impedance change rate (preferably ±20% relative to the reference value), patient heart rate range, lower limit of blood oxygen saturation (preferably ≤88%), EEG epileptiform activity detection markers, and ventilator asynchrony event markers.

[0077] Step S503-B: Security Incident Classification Determination. Based on the comparison results of step S502-B, the unit classifies the detected anomalies into the following three levels: Warning level: When parameters are close to the safety threshold but have not yet exceeded it (e.g., the temperature reaches 38°C), a warning sign is generated and sent to the clinician's operating interface, without affecting ultrasound output; Suppression level: When parameters exceed the safety threshold but do not immediately endanger safety (e.g., coupling quality score drops to 75-85 points), a suppression command is generated and sent to the closed-loop control interface to forcibly interrupt the current intervention permission state and temporarily prevent hardware shutdown. Shutdown level: When parameters exceed emergency safety boundaries (such as temperature exceeding 41°C, detection of epileptiform EEG activity, arrhythmia, or equipment overheating), an unconditional shutdown command is generated.

[0078] Step S504-B: Hardware-level interrupt and shutdown trigger. For the shutdown-level event determined in step S503-B, the safety monitoring unit sends it directly to the shutdown execution circuit via an independent hardware interrupt signal line without going through the software logic of the closed-loop control interface. This signal has the highest priority and can cover all software control commands. At the same time, the unit records the event type, detection time, response time, and all current relevant parameter values ​​to a non-volatile memory and sends it to the alarm / emergency stop / fail-safe unit to trigger an audible and visual alarm.

[0079] Step S505-B: State Maintenance and Recovery Determination. After triggering suppression or shutdown, the safety monitoring unit continuously monitors relevant parameters. Only after the abnormal state is eliminated and a predetermined safety waiting time (preferably 30 seconds) has elapsed can it send a "safe recovery" flag to the closed-loop control interface, allowing the system to reassess the stimulation conditions.

[0080] The alarm / emergency stop / fail-safe unit 15 is configured as the final safety protection level of the system, and its data processing has the highest priority and runs independently of the main control software.

[0081] The data processing procedure for the alarm / emergency stop / fail-safe unit is as follows: Step S1401: Multi-channel interrupt signal monitoring. The unit continuously monitors the following independent signal channels: Hardware interrupt signal for emergency stop switch (physical button or foot switch); Shutdown command signal of the safety monitoring unit; Power warning signs and power failure detection signals of the power supply and equipment status management unit; Software watchdog timer overflow signal (if the main control software fails to refresh the watchdog within the predetermined period, it is determined to be a software failure). Voltage monitoring signal of hardware watchdog circuit.

[0082] Step S1402: Shutdown execution and physical isolation. When any of the above channels is triggered, the unit executes the following sequence of actions within a predetermined maximum response time (preferably ≤10ms): The high-voltage power supply circuit of the transcranial ultrasound transducer and the therapeutic transcranial ultrasound equipment is cut off by a redundant shutdown switch. The output current is clamped to zero by a current-limiting circuit; The power amplifier output terminal is shorted to ground by the fail-safe power-off circuit to release the residual charge. Activate the audible and visual alarm devices (including bedside alarm lights, buzzers, and remote nurse station notifications).

[0083] Step S1403: Fault Status Recording and Latching. The unit records the fault event type, trigger channel, response time, and system status snapshot to non-volatile memory. This record cannot be erased by software and can only be reset by maintenance personnel via a physical button. Simultaneously, the unit sends the aforementioned fault event information to the security database / audit trail via a separate communication interface.

[0084] Step S1404: Graded alarm output. For conditions that do not require shutdown but require alerting (such as warning-level events), the unit issues graded alarm signals, including visual cues (yellow warning light) and audible cues (intermittent buzzer), without performing hardware shutdown.

[0085] The fifth layer includes a post-intervention feedback unit 11, a clinician operation interface 12, and a security database / audit trail 13, forming a closed-loop processing link of "post-stimulation multimodal feedback acquisition → closed-loop parameter adaptive correction → operator interaction and control → full-process data archiving", realizing the system's self-optimization and traceability.

[0086] The post-intervention feedback unit 11 is configured to receive EEG, ECG, PPG, SpO2, respiration, ventilator, exercise, environmental, or medication-related feedback after ultrasound delivery, and transmit the feedback to the closed-loop controller for subsequent parameter adjustment. The data processing procedure is as follows: Step S1101: Feedback data acquisition initiated. The unit automatically starts upon completion of ultrasound delivery (or sham stimulation), or after a delay set by the clinician (preferably 1 minute after stimulation). The unit acquires the following multimodal feedback signals from monitoring equipment, EEG machine, ventilator, and bedside sensors via the system data interface: EEG signal: Multi-lead EEG power spectrum data, especially the absolute and relative power of each frequency band of delta waves (0.5-4Hz), theta waves (4-8Hz), alpha waves (8-13Hz), and beta waves (13-30Hz); ECG signals: heart rate, heart rate variability (time-domain indicators such as SDNN and RMSSD, and frequency-domain indicators such as LF / HF). PPG signal and SpO2: peripheral blood oxygen saturation and pulse wave morphology; Respiratory signals: respiratory rate, tidal volume, and ventilator support parameters (if applicable); Motion signal: Activity intensity index calculated from data from triaxial accelerometers and gyroscopes; Environmental and medication records: ward noise level, light intensity, and the types and dosages of sedatives / analgesics administered during the same period.

[0087] Step S1102: Signal Preprocessing and Feature Extraction. The unit performs the following preprocessing on the raw signal obtained in step S1101: baseline drift correction, power frequency notch filtering (50Hz / 60Hz), motion artifact detection and removal (marking but not deleting periods with excessive acceleration), and segmented windowing. Based on this, key indicators after intervention are extracted, including: the rate of change of EEG power in each frequency band relative to the pre-stimulation baseline, the δ / θ power ratio, various indicators of heart rate variability, blood oxygenation trend slope, and respiratory stability index.

[0088] Step S1103: Intervention Response Feature Vector Generation. The unit integrates the key indicators extracted in step S1102 into an intervention response feature vector. This vector contains quantified response values ​​for multiple time windows after stimulation (preferably 0-5 minutes, 5-15 minutes, and 15-30 minutes). The vector output format is a standardized message matching the predefined data structure of the closed-loop control interface.

[0089] Step S1104: Feedback Data Transmission and Storage. The unit transmits the intervention response feature vector to the closed-loop control interface via a real-time data link. This is used by the closed-loop controller to compare the pre-stimulation baseline with the post-stimulation response, calculate the response difference, and determine the adjustment direction. Simultaneously, the complete raw feedback data and the processed feature vector are timestamped and sent to the safety database / audit trail for efficacy evaluation and safety review.

[0090] Step S1105: Adaptive Parameter Correction. After receiving feedback data, the closed-loop control interface performs parameter correction calculations: if the feedback shows increased slow-wave EEG activity in the target area and no adverse safety events, the stimulation intensity is maintained or gradually increased; if the feedback shows no significant EEG response but good safety conditions, the target coordinates are adjusted or the pulse train duration is increased; if the feedback shows any negative safety trend, the current intensity is maintained or the output is reduced to ensure that a complete negative feedback regulation loop is formed after intervention monitoring.

[0091] The clinician's operating interface 12 is configured to display the transcranial ultrasound transducer position, coupling mass fraction, target coordinates, planned exposure parameters, delivered exposure parameters, safety gate status, intervention permission status, and manual stop or over-control settings. The data processing procedure is as follows: Step S1201: System status data aggregation and display. The interface reads real-time status data from the following units at a refresh rate of not less than 1Hz: Bedside positioning and fixing unit 2: Graphical display of transducer spatial position and offset value; Acoustic coupling unit 4: Coupling quality score progress bar and various sub-indicators; Patient-Specific Targeting and Dosage Planning Unit 5: Overlay display of target coordinates on a 3D brain model and list of planned exposure parameters; Safety monitoring unit 6: Safety door status indicator lights (each individual check passed / failed) and real-time safety event log; Closed-loop control interface 7: Intervention permission status indicator, current sleep instability index curve and confidence score; Ultrasound output / stimulation program 8 and therapeutic transcranial ultrasound equipment 9: delivered exposure parameters (real-time cumulative acoustic dose, duration of emission, and duration of remaining emission).

[0092] The interface displays the above data in a partitioned panel of the graphical user interface after formatting and rendering, and uses color coding (green for normal, yellow for warning, and red for alarm) to assist in quick identification.

[0093] Step S1202: Receiving and Verifying Clinician Input. The interface receives operation commands input by clinicians via touchscreen, mouse, or keyboard, including: manual target selection or adjustment, modification of stimulus parameter values ​​(frequency, intensity, duration, etc.), triggering of stimulus start / stop buttons, and input of safety over-limit commands (requiring dual authorization). Before executing each command, the interface verifies the validity of the input value, checking whether it is within the system's preset safety limits. If it exceeds these limits, the input is rejected and a prompt box is displayed.

[0094] Step S1203: Command Conversion and Distribution. The interface converts the operator's commands that have passed the validity verification into internal system control instructions and sends them to the following according to the predetermined communication protocol: closed-loop control interface (parameter adjustment and start / stop commands), safety monitoring unit (ultrasound control authorization commands), and ultrasound output / stimulation program (instant stop commands with high priority).

[0095] Step S1204: Operation Behavior Recording. The interface records each interaction behavior of the operator, including the operator's identity identifier (login ID), operation timestamp, operation type (such as "parameter modification", "manual stop", "super control authorization"), values ​​before and after modification, and operation result (success / failure and reason code). All operation logs are sent to the security database / audit trail in real time in a structured format.

[0096] The safety database / audit trail 13 is configured to store stimulation parameters, safety status information, coupling status, intervention events, feedback responses, and operator behavior. The recording unit stores at least one of the following in de-identified form: patient identifier, transducer type, transducer location, target, stimulation parameters, acoustic dose estimation, coupling quality, scalp temperature, safety gate status, controller commands, delivered ultrasound emission logs, post-intervention feedback, operator behavior, alarms, and model or software version. Each intervention event is timestamped. The logs include pre-intervention sleep instability, intervention permission status, safety gate status, selected target, selected parameters, actual delivery output, coupling score, transducer location, temperature, device status, clinician actions, alarms, and post-intervention feedback. The database can export structured reports for physician review, research analysis, quality control, and regulatory documentation. The data processing procedure is as follows: Step S1301: Data Reception and Event Queue Management. The database continuously receives data records from all system units in an event-driven manner, employing asynchronous non-blocking write queue management to ensure that high-speed data flow does not lead to write conflicts or data loss. Each arriving data packet is assigned an incrementing sequence number and added to the write queue.

[0097] Step S1302: De-identification. Before storage, the database performs de-identification on each data record, replacing patient-identifying information (name, hospital number, ID number, etc.) with an internally generated unique research identifier. The mapping table for this identifier is stored separately in encrypted form and is only accessible to authorized personnel. The de-identified data is used for research analysis and quality control.

[0098] Step S1303: Add timestamps and associate them with events. The database adds a high-precision timestamp (accurate to milliseconds) to each record and associates the record with the corresponding global intervention event ID based on the event type. Planning data, positioning data, coupling data, safety data, output parameters, feedback data, and operation logs belonging to the same intervention event are all indexed and associated through the event ID, supporting complete event chain tracing.

[0099] Step S1304: Structured Storage and Index Construction. The database parses and categorizes each field according to a predetermined data model, constructing a multi-level index structure including time indexes, event type indexes, patient identifier indexes, and target location indexes to support efficient queries. The database also maintains an incremental backup mechanism, recording transaction logs for each write operation to ensure data integrity.

[0100] Step S1305: Data Query and Report Generation. The database allows operators to perform the following searches through the query module of the clinician's interface: querying the list of intervention events by time period, querying the complete treatment history by patient identifier, querying alarm records by safety event type, and querying dose distribution statistics by target location. The database can generate structured reports based on predefined templates, including: single intervention event reports (containing all relevant parameters and feedback), daily / weekly summary reports, and safety event audit reports, for physician review, research analysis, quality control, and regulatory documentation.

[0101] Step S1306: Clinical outcome association analysis. The database also provides an interface for association with clinical outcome data, supporting the association query and statistical analysis of intervention event data with clinical endpoint indicators such as delirium assessment results, total dose of sedative drugs, number of days without ventilator use, or sleep quality scores, but does not limit the scope of this invention to any specific clinical endpoint.

[0102] The beneficial effects are as follows: 1. The device provides an ICU-compatible physical platform for transcranial ultrasound neuromodulation, solving practical problems related to prolonged bedside use, airway equipment, monitoring leads, and nursing workflow.

[0103] 2. The device improves the repeatability of transcranial ultrasound delivery by integrating positioning, coupling status monitoring, targeting, and acoustic dose planning.

[0104] 3. The device improves safety by requiring that coupling, temperature, patient status, device status, and clinician-defined conditions be met before ultrasound output can be activated.

[0105] 4. The device achieves closed-loop neural regulation by receiving sleep instability and security gating commands and returning the intervention feedback to the controller.

[0106] 5. The device supports high-quality clinical research through sham stimulation units and detailed audit logs.

[0107] 6. The device can be extended to other neuromodulation targets and ICU applications by adjusting the target, parameters and safety gating criteria.

Claims

1. An ICU-compatible transcranial ultrasound neuromodulation system, characterized in that, The system includes: The pre-planning and basic support layer includes a patient-specific targeting and dose planning unit and a power and equipment status management unit to provide target baselines and energy bases; The initial verification layer includes the initial verification sub-functions of the bedside positioning and fixing unit, the acoustic coupling unit, and the safety monitoring unit, which are used to complete the transducer spatial positioning verification, acoustic transmission path quality verification, and comprehensive safety initial state verification, respectively. The closed-loop control layer includes a closed-loop control interface for synchronously deciding on all preconditions from the pre-planning and basic support layer and the initial verification layer, and receiving real-time physiological feedback from the patient monitoring device to generate intervention permission or rejection commands. The execution and real-time safety layer includes real-time monitoring sub-functions of ultrasound output / stimulation program, therapeutic transcranial ultrasound equipment, transcranial ultrasound transducer, sham stimulation / blinding unit, and safety monitoring unit, as well as alarm / emergency stop / fail-safe unit, for performing ultrasound stimulation or sham stimulation under the intervention permission command and performing real-time safety monitoring; The feedback and data management layer includes a post-intervention feedback unit, a clinician interface, and a security database / audit trail. It is used to collect multimodal feedback after stimulation, perform closed-loop parameter adaptive correction, provide operator interaction and control, and archive data throughout the entire process.

2. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The patient-specific targeting and dose planning unit is configured to determine the ultrasound target coordinates, acoustic path, incident angle, skull transmission compensation, and exposure parameters based on images, anatomical landmarks, or physiological responses, and output them to the closed-loop control interface and ultrasound output / stimulation program.

3. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The power supply and device status management unit is configured to: collect power supply voltage and current consumption data of each submodule in real time, perform power quality analysis and ripple detection, and generate a global power health status word containing three levels of identifiers: "normal / degraded / failed"; at the same time, it aggregates the operating status of each device to form a system status topology diagram, and periodically sends the status word to the closed-loop control interface and the safety monitoring unit.

4. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The bedside positioning and fixation unit is configured to maintain a stable spatial relationship between the transcranial ultrasound transducer and the patient's head, and detect the displacement of the transducer relative to the head. Its real-time pose data is obtained through multi-sensor fusion, and the offset is compared with a preset threshold to generate a position deviation flag. This flag serves as one of the gating conditions, and an enable signal is only output to the transducer drive circuit when all safety gate conditions are met.

5. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The acoustic coupling unit is configured between the transducer and the scalp. It collects multimodal signals through sensors in a disposable hydrogel coupling pad, calculates the comprehensive coupling quality score, and compares it with an acceptable threshold to generate a binary coupling state verification signal. During ultrasonic delivery, the dynamic changes in coupling quality are continuously monitored, and a coupling failure indicator is generated when the quality deteriorates or changes abruptly.

6. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The safety monitoring unit performs initial safety data aggregation and initial vital sign status acquisition at the initial verification layer, generating an initial safety clearance comprehensive flag; it performs hardware-level interrupt path independent of the closed-loop control interface for executing and real-time safety layer data processing; it performs multi-level threshold comparison on multi-source safety signals, classifying anomalies into warning level, suppression level, and shutdown level, and directly triggers the shutdown execution circuit through an independent hardware interrupt signal line for shutdown level events.

7. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The closed-loop control interface is configured as follows: with prior condition synchronous verification as a prerequisite, it receives multimodal physiological signals from the patient monitoring device, calculates the sleep instability index and confidence score, and generates an intervention permission command when the index exceeds the threshold, the confidence score is sufficient, and the duration is predetermined. This command includes an operation mode identifier and a set of planned parameters, and is simultaneously sent to the ultrasound output / stimulation program, the sham stimulation / blinding unit, and the safety database. After the command is issued, a local safety check is performed again, and only after passing the check can the system enter the ready-to-launch state.

8. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The ultrasound output / stimulation program is configured as follows: if the operation mode identifier in the intervention permission command is parsed, and if it is "active stimulation", a digital waveform sample sequence is generated using a direct digital frequency synthesis algorithm according to the plan parameters; if it is "sham stimulation", the drive signal is routed to the virtual load and a timing control signal and sensory simulation signal that are exactly the same as those of active stimulation are generated synchronously.

9. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The therapeutic transcranial ultrasound device is configured to: receive digital waveform sample sequences, perform digital-to-analog conversion, power amplification and channel allocation, apply phase delay to multi-element phased array transducers to achieve electronic focus deflection; and collect voltage, current and reverse power data of the output channels in real time and send them to the safety monitoring unit through an independent data link.

10. The ICU-compatible transcranial ultrasound neuromodulation system as described in claim 1, characterized in that, The sham stimulation / blinding unit is configured as follows: in sham stimulation mode, the drive signal is connected to the virtual load without generating transcranial ultrasound; indicator lights, cooling fan noise and tactile vibration signals matching active stimulation are generated simultaneously to maintain the same operating procedure as active stimulation; sham stimulation events are encrypted and stored in the same recording format as active stimulation, and the operator interface only displays the unified status information of the "stimulation event".

Images