Description:Brief Description of the Drawings

Generally, the present disclosure relates to health monitoring systems. Particularly, the present disclosure relates to a system for proactive seizure prediction and alerting.
Background
The background description includes information that may be useful in understanding the present invention. It is not an admission that any of the information provided herein is prior art or relevant to the presently claimed invention, or that any publication specifically or implicitly referenced is prior art.
In the realm of medical diagnostics and patient monitoring, a significant challenge remains in the early detection and management of seizure disorders. Seizures, unpredictable in nature, necessitate continuous monitoring to mitigate risks and administer timely intervention. Traditional methods of seizure detection largely depend on the monitoring of physiological signals, primarily through hospital-based systems that require significant resources and specialized equipment.
Furthermore, the advent of wearable technology has seen the development of portable monitoring systems. These systems, equipped with sensors for the acquisition of physiological data, mark a significant advancement in patient-centric healthcare. Among the data acquired, the Electroencephalogram (EEG) is paramount for its direct correlation with brain activity, often utilized alongside other signals such as the Electrocardiogram (ECG), Electromyography (EMG), and Electrooculography (EOG) to enhance diagnostic accuracy.
However, the processing and analysis of such data present inherent challenges. The extraction of meaningful features from raw physiological signals requires sophisticated algorithms capable of identifying potential indicators of seizure activity. Traditional systems often fall short in this regard, lacking the advanced computational techniques necessary for effective feature isolation.
Privacy concerns further complicate the widespread adoption of continuous monitoring systems. The sensitive nature of physiological data mandates stringent measures to ensure patient privacy. Existing systems, while effective to a degree, often lack the necessary mechanisms to provide differential privacy without compromising the integrity of the data.
Moreover, the classification and prediction of seizure events remain a complex task. The variability in seizure manifestations necessitates a highly nuanced approach to detection. Previous systems have employed various methods, but many fail to achieve a balance between sensitivity and specificity, leading to a high rate of false positives or negatives.
Additionally, the confidence in seizure predictions is a crucial factor. Systems that lack a reliable method for evaluating prediction confidence may lead to unnecessary alarm and anxiety among patients and caregivers. The assessment of prediction confidence, thus, remains a key challenge for existing seizure detection systems.
Given the limitations of current methodologies and technologies, there exists an urgent need for solutions that overcome the challenges associated with conventional systems and techniques for the proactive prediction and alerting of seizure events. Such solutions should offer advanced feature extraction, ensure privacy preservation, accurately classify and predict seizure events, and provide reliable confidence evaluation mechanisms to facilitate timely and appropriate response.
Summary
The following presents a simplified summary of various aspects of this disclosure in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements nor delineate the scope of such aspects. Its purpose is to present some concepts of this disclosure in a simplified form as a prelude to the more detailed description that is presented later.
The following paragraphs provide additional support for the claims of the subject application.
In an aspect, the present disclosure provides a system for proactive seizure prediction and alerting. The system comprises an array of non-invasive sensors configured to acquire continuous physiological data, including at least an Electroencephalogram (EEG), an Electrocardiogram (ECG), an Electromyography (EMG), and an Electrooculography (EOG); and a central monitoring unit. The central monitoring unit includes a feature extraction unit operative to process the acquired physiological data to isolate features indicative of potential seizure activity; a privacy unit comprising a noise injector operative to apply a calculated amount of random noise to the extracted features to provide differential privacy while preserving patterns indicative of seizures; a seizure classification and prediction unit implementing a DeepConvEEGgNet technique, configured with multiple Conv2D layers for feature detection, batch normalization layers for input normalization, maxPooling layers for spatial size reduction of the representation, and a dense neural network with global average pooling and dropout rate mechanisms for classification and prediction of seizure events; a confidence evaluation mechanism configured to assess the prediction confidence using a threshold mechanism; and a real-time alert unit operative to communicate alerts to healthcare providers or caregivers when the prediction exceeds the confidence threshold.
The EEG and EOG data are collected through a wearable helmet comprising a headgear lining, a pair of flexible arms, a plurality of EEG electrode arrays embedded within the helmet lining, EOG sensor placements near the eyes, embedded within flexible arms that gently press against the temples to detect eye movement without causing discomfort, a central processing unit that amplifies and filters the EEG and EOG signals before transmission, a wireless communication module for transmitting processed data to the central monitoring unit, adjustable straps, and cushioning for secure placement and patient comfort.
The EMG data and ECG are collected by a wristband comprising a flexible, adjustable strap that comfortably fits around the patient’s wrist, embedded ECG electrodes on the inner surface of the wristband that make contact with the skin for cardiac activity monitoring, EMG sensors located on the strap to align with the muscles of interest on the forearm for electromyography, a miniaturized signal processing unit within the wristband to condition and digitize the ECG and EMG signals, a rechargeable battery with wireless charging capability for long-term continuous operation, and an integrated wireless transmitter for relaying the collected ECG and EMG data to the central monitoring unit.
The feature extraction unit utilizes machine learning techniques to enhance feature isolation from the physiological data. The headgear lining comprises a reservoir to store a conductive gel, and discharge mechanisms to apply the conductive gel on the surface of the sensor that come in contact with the skull. The noise injector is further configured to adjust the level of noise based on a privacy-utility trade-off technique. The DeepConvEEGgNet technique includes Conv2D layers with ReLU activation functions. The MaxPooling layers are configured to follow each batch normalization layer to enhance feature invariance to scale and orientation changes. The confidence evaluation mechanism utilizes a dynamic threshold that is adaptive based on historical prediction accuracy. The system further comprises an intervention unit equipped to automatically administer medication upon the triggering of an alert.

Field of the Invention

The features and advantages of the present disclosure would be more clearly understood from the following description taken in conjunction with the accompanying drawings in which:
FIG. 1 illustrates a block diagram of a system (100) for proactive seizure prediction and alerting, in accordance with the embodiments of the present disclosure.
FIG. 2 illustrates a seizure prediction system that utilizes multiple biosignals for proactive seizure detection, in accordance with the embodiments of the present disclosure.
FIG. 3 illustrates architecture of the secure DeepConvEEGNet algorithm, in accordance with the embodiments of the present disclosure.

Detailed Description
In the following detailed description of the invention, reference is made to the accompanying drawings that form a part hereof, and in which is shown, by way of illustration, specific embodiments in which the invention may be practiced. In the drawings, like numerals describe substantially similar components throughout the several views. These embodiments are described in sufficient detail to claim those skilled in the art to practice the invention. Other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the scope of the present invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is defined only by the appended claims and equivalents thereof.
The use of the terms “a” and “an” and “the” and “at least one” and similar referents in the context of describing the invention (especially in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The use of the term “at least one” followed by a list of one or more items (for example, “at least one of A and B”) is to be construed to mean one item selected from the listed items (A or B) or any combination of two or more of the listed items (A and B), unless otherwise indicated herein or clearly contradicted by context. The terms “comprising,” “having,” “including,” and “containing” are to be construed as open-ended terms (i.e., meaning “including, but not limited to,”) unless otherwise noted. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., “such as”) provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.
Pursuant to the "Detailed Description" section herein, whenever an element is explicitly associated with a specific numeral for the first time, such association shall be deemed consistent and applicable throughout the entirety of the "Detailed Description" section, unless otherwise expressly stated or contradicted by the context.
The term "system" as used throughout the present disclosure relates to an assembly configured to predict seizure events proactively and to alert healthcare providers or caregivers upon the prediction of such events. The system encompasses an array of non-invasive sensors and a central monitoring unit, which together facilitate the acquisition, processing, and analysis of physiological data to predict seizure events.
The term "array of non-invasive sensors" as used throughout the present disclosure relates to a collection of devices configured to acquire continuous physiological data without penetrating the skin or body of the patient. These sensors include, but are not limited to, sensors for recording Electroencephalogram (EEG), Electrocardiogram (ECG), Electromyography (EMG), and Electrooculography (EOG) signals. The acquired data is essential for the detection of potential seizure activity.
The term "central monitoring unit" as used throughout the present disclosure relates to the primary processing and analysis component of the system. The central monitoring unit comprises several sub-units including a feature extraction unit, a privacy unit, a seizure classification and prediction unit, and a real-time alert unit. Each of these sub-units plays a crucial role in the system's ability to predict seizures and protect patient privacy.
The term "feature extraction unit" as used throughout the present disclosure relates to a component of the central monitoring unit responsible for processing the acquired physiological data to isolate features indicative of potential seizure activity. This unit employs advanced algorithms to analyze the data and identify features that are characteristic of seizures.
The term "privacy unit" as used throughout the present disclosure relates to a component designed to ensure the privacy of the patient's data. It includes a noise injector operative to apply a calculated amount of random noise to the extracted features. This process provides differential privacy while preserving the patterns indicative of seizures, thereby balancing the need for privacy with the need for accurate seizure prediction.
The term "seizure classification and prediction unit" as used throughout the present disclosure relates to a component implementing the DeepConvEEGgNet technique. It is equipped with multiple Conv2D layers for feature detection, batch normalization layers for input normalization, maxPooling layers for spatial size reduction of the representation, and a dense neural network with global average pooling and dropout rate mechanisms. This unit classifies and predicts seizure events based on the processed data.
The term "confidence evaluation mechanism" as used throughout the present disclosure relates to a component configured to assess the prediction confidence using a threshold mechanism. It evaluates the reliability of the seizure predictions made by the seizure classification and prediction unit, ensuring that alerts are generated only when the predictions exceed a certain confidence threshold.
The term "real-time alert unit" as used throughout the present disclosure relates to a component operative to communicate alerts to healthcare providers or caregivers. The unit is activated when the prediction confidence exceeds the predetermined threshold, facilitating timely intervention in the event of a predicted seizure.
FIG. 1 illustrates a block diagram of a system (100) for proactive seizure prediction and alerting, in accordance with the embodiments of the present disclosure. The system (100) includes an array of non-invasive sensors (102) and a central monitoring unit (104). The array of non-invasive sensors (102) is configured to acquire continuous physiological data from a subject. This data encompasses a spectrum of biological signals, including but not limited to an Electroencephalogram (EEG), an Electrocardiogram (ECG), an Electromyography (EMG), and an Electrooculography (EOG). The central monitoring unit (104) is responsible for processing and analyzing the physiological data collected by the array of non-invasive sensors (102). Within the central monitoring unit (104), a feature extraction unit is operative to isolate features from the acquired physiological data, indicative of potential seizure activity. A privacy unit is incorporated to provide differential privacy by injecting calculated random noise into the extracted features, thereby preserving patterns indicative of seizures while maintaining the privacy of the data. Further included in the central monitoring unit (104) is a seizure classification and prediction unit. This unit utilizes a DeepConvEEGgNet technique comprising multiple Conv2D layers, batch normalization layers, maxPooling layers, and a dense neural network equipped with global average pooling and dropout rate mechanisms. Such a configuration enables classification and prediction of seizure events with enhanced accuracy. Additionally, a confidence evaluation mechanism assesses the prediction confidence based on a predefined threshold mechanism, and a real-time alert unit is operative to issue alerts when the prediction exceeds the confidence threshold. Thus, system (100) serves as a comprehensive solution for proactive seizure monitoring and alerting.
In an embodiment, wherein the EEG data and EOG data is collected through a wearable helmet is described. The helmet is designed for optimal comfort and functionality, featuring a headgear lining that ensures a snug fit while housing a plurality of EEG electrode arrays for comprehensive brain activity monitoring. Flexible arms extend from the helmet, equipped with EOG sensor placements near the eyes. These sensors are carefully embedded to press gently against the temples, allowing for accurate detection of eye movement without causing discomfort to the user. A central processing unit within the helmet amplifies and filters the EEG and EOG signals before their transmission, ensuring that only the most relevant data is forwarded for analysis. This processing unit works in conjunction with a wireless communication module, enabling the seamless transmission of processed data to the central monitoring unit (104). Additionally, the helmet is fitted with adjustable straps and cushioning, promoting secure placement on the head and enhancing patient comfort during prolonged use. This design not only facilitates the continuous and non-invasive monitoring of physiological signals but also enhances the user experience by prioritizing comfort and ease of use.
In another embodiment, the system (100) involves the collection of EMG data and ECG by a wristband. This wristband is characterized by its flexible, adjustable strap designed to fit comfortably around the patient’s wrist, ensuring a secure yet non-restrictive fit. Embedded within the wristband's inner surface are ECG electrodes that make direct contact with the skin for cardiac activity monitoring. Additionally, EMG sensors are strategically placed on the strap to align with specific forearm muscles of interest, allowing for targeted electromyography. The wristband houses a miniaturized signal processing unit, tasked with conditioning and digitizing the ECG and EMG signals to prepare them for analysis. To support long-term continuous operation, the wristband is equipped with a rechargeable battery featuring wireless charging capability. Moreover, an integrated wireless transmitter within the wristband facilitates the reliable relaying of collected ECG and EMG data to the central monitoring unit (104). This innovative design enables the continuous, non-invasive monitoring of cardiac and muscle activity, essential for the proactive prediction of seizure events.
In a further embodiment, the system (100) comprises a feature extraction unit that utilizes machine learning techniques to enhance feature isolation from the physiological data. By leveraging advanced algorithms, the unit analyzes the incoming data stream from the non-invasive sensors, extracting features indicative of potential seizure activity with unprecedented precision. This approach enables the identification of subtle patterns and anomalies in the physiological data that may signify the onset of a seizure, far beyond the capabilities of traditional analysis methods. The incorporation of machine learning techniques represents a significant advancement in the system's ability to predict seizures proactively, providing a foundation for more accurate and timely alerts.
In another embodiment, the system (100) includes a headgear lining within the helmet that comprises a reservoir designed to store conductive gel. This gel is essential for improving sensor conductivity and ensuring accurate signal acquisition. Discharge mechanisms are integrated into the lining to apply the conductive gel directly onto the surface of the sensors that come into contact with the skull, thereby enhancing the quality of the EEG signals collected. This feature addresses a common challenge in EEG data acquisition by ensuring consistent signal quality, which is crucial for the reliable prediction of seizure events.
In a further embodiment, the system (100) features a noise injector within the privacy unit that is further configured to adjust the level of noise based on a privacy-utility trade-off technique. This advanced configuration allows for the fine-tuning of the noise injection process, optimizing the balance between preserving data privacy and maintaining the integrity of patterns indicative of seizures. By adjusting the level of noise injected into the extracted features, the system ensures the protection of patient data while retaining the essential information needed for accurate seizure prediction.
In another embodiment, the system (100) employs the DeepConvEEGgNet technique for seizure classification and prediction, which includes Conv2D layers with ReLU activation functions. The ReLU activation function is crucial for introducing non-linearity into the network, enabling the model to learn complex patterns in the physiological data indicative of seizure activity. This feature significantly enhances the system's predictive capabilities, allowing for the accurate classification and prediction of seizure events based on the analyzed data.
In a further embodiment, the system (100) is designed such that the MaxPooling layers are configured to follow each batch normalization layer. This arrangement enhances feature invariance to scale and orientation changes, a critical factor in the accurate classification and prediction of seizure events. By reducing the spatial size of the representation, the system can efficiently process the physiological data, ensuring timely and accurate seizure predictions.
In another embodiment, the system (100) incorporates a confidence evaluation mechanism that utilizes a dynamic threshold adaptive based on historical prediction accuracy. This mechanism assesses the reliability of seizure predictions by adjusting the confidence threshold in response to the system's past performance. Such adaptability ensures that alerts are only issued when the system's predictions surpass a confidence level informed by historical accuracy, minimizing false alarms
In an embodiment, the system (100) further comprises an intervention unit equipped to automatically administer medication upon the triggering of an alert. The intervention unit is integrated into the system (100) to provide a responsive therapeutic intervention following the proactive seizure prediction. Upon the seizure classification and prediction unit's determination that a seizure event is imminent, and when the confidence evaluation mechanism confirms the prediction's reliability exceeds the predetermined confidence threshold, the intervention unit is actuated. Said intervention unit is preloaded with medication specifically prescribed for the patient and is connected to a delivery mechanism. The delivery mechanism is configured to administer the medication in a timely manner, potentially mitigating the effects of the seizure or preventing the seizure altogether. The actuation of said intervention unit is performed in real-time and is synchronized with the alert system to ensure prompt response following the prediction of a seizure event. This aspect of the system (100) emphasizes patient safety by reducing the time between seizure prediction and therapeutic intervention, thereby providing an automated approach to seizure management. The incorporation of said intervention unit into the system (100) ensures that both monitoring and initial treatment steps are addressed within a single integrated solution, enhancing the overall efficacy of the seizure management protocol provided by the system (100).
FIG. 2 illustrates a seizure prediction system that utilizes multiple biosignals for proactive seizure detection, in accordance with the embodiments of the present disclosure. The biosignals include Electroencephalography (EEG), Electrocardiography (ECG), Electromyography (EMG), and Electrooculography (EOG). Each of these signals offers unique insights into physiological states that may precede a seizure. The collected biosignal data are subjected to feature extraction processes that involves distilling the raw biosignal data into meaningful patterns or attributes that can be analyzed effectively. Upon extraction of features, the data proceeds to the seizure classification phase, using secure DeepConvEEGNet algorithm that can distinguish between normal and pre-seizure states. The classification is based on a predictive confidence threshold mechanism, a quantitative measure that determines the likelihood of a seizure event. If the system's confidence that a seizure is forthcoming surpasses a predetermined threshold, an alarm is triggered.
FIG. 3 illustrates architecture of the secure DeepConvEEGNet algorithm, in accordance with the embodiments of the present disclosure. The neural network is designed with privacy-preserving features, using a differential privacy layer to enable that patient data is handled securely. The architecture begins with the raw input EEG data which is then passed through a privacy layer that injects noise, thereby obfuscating patient-specific details and ensuring that the data analysis adheres to privacy standards. Following the privacy layer, the data passes through multiple convolutional layers. Each convolutional layer is characterized by filters and a Rectified Linear Unit (ReLU) activation function, which identifies and activates relevant patterns within the data. The layers also include Batch Normalization (BN), which stabilizes the learning process by normalizing the input layer by adjusting and scaling the activations. Subsequent to the convolutional layers, the architecture employs max-pooling operations, which reduce the dimensionality of the data, condensing the feature set while preserving the most significant features identified by the previous layers. This step enhances the efficiency of the network by reducing computational load and helps in preventing overfitting. The network then channels the data through a dense layer which is connected to a global average pooling function. The dense layer includes a dropout mechanism with a rate of 0.5, a regularization technique used to prevent overfitting by randomly deactivating a portion of the neurons in the layer during the training process. The output layer, which is responsible for the seizure prediction. The output is derived from the processed and analyzed data, culminating in the delivery of a predictive result. This result is utilized by the system to determine whether to activate the real-time alarm.
The system of present disclosure aims to address the challenge of accurately predicting and alerting healthcare providers of impending seizure events while enabling the privacy of patient medical data. The multimodal signal processing which process physiological data from various sources including EEG, ECG, EMG, and EOG for seizure prediction. The diverse data gathered is then refined through a feature extraction process, isolating characteristics that are indicative of potential seizure activity.The system integrates a differential privacy layer early in the data processing stage. A noise injector within this layer introduces a calculated amount of random noise to the data, for masking individual identities while retaining the overall patterns that signal seizure occurrence. Further, dense neural network evaluates the data processed by the Conv2D layers, assigning classifications and forecasting potential seizure events. The inclusion of global average pooling and dropout techniques within this stage mitigates overfitting, improve the model's robustness. The reliability of the predictions is determined by a predictive confidence threshold mechanism. Only when the model's prediction exceeds a predefined confidence level does it trigger an alert, thereby reducing false alarms. Upon high-confidence detection of a possible seizure, the system activates a real-time alert, promptly notifying healthcare providers to initiate preemptive measures.
Example embodiments herein have been described above with reference to block diagrams and flowchart illustrations of methods and apparatuses. It will be understood that each block of the block diagrams and flowchart illustrations, and combinations of blocks in the block diagrams and flowchart illustrations, respectively, can be implemented by various means including hardware, software, firmware, and a combination thereof. For example, in one embodiment, each block of the block diagrams and flowchart illustrations, and combinations of blocks in the block diagrams and flowchart illustrations can be implemented by computer program instructions. These computer program instructions may be loaded onto a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions which execute on the computer or other programmable data processing apparatus create means for implementing the functions specified in the flowchart block or blocks.
Throughout the present disclosure, the term ‘processing means’ or ‘microprocessor’ or ‘processor’ or ‘processors’ includes, but is not limited to, a general purpose processor (such as, for example, a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a microprocessor implementing other types of instruction sets, or a microprocessor implementing a combination of types of instruction sets) or a specialized processor (such as, for example, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), or a network processor).
The term “non-transitory storage device” or “storage” or “memory,” as used herein relates to a random access memory, read only memory and variants thereof, in which a computer can store data or software for any duration.
Operations in accordance with a variety of aspects of the disclosure is described above would not have to be performed in the precise order described. Rather, various steps can be handled in reverse order or simultaneously or not at all.
While several implementations have been described and illustrated herein, a variety of other means and/or structures for performing the function and/or obtaining the results and/or one or more of the advantages described herein may be utilized, and each of such variations and/or modifications is deemed to be within the scope of the implementations described herein. More generally, all parameters, dimensions, materials, and configurations described herein are meant to be exemplary and that the actual parameters, dimensions, materials, and/or configurations will depend upon the specific application or applications for which the teachings is/are used. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific implementations described herein. It is, therefore, to be understood that the foregoing implementations are presented by way of example only and that, within the scope of the appended claims and equivalents thereto, implementations may be practiced otherwise than as specifically described and claimed. Implementations of the present disclosure are directed to each individual feature, system, article, material, kit, and/or method described herein. In addition, any combination of two or more such features, systems, articles, materials, kits, and/or methods, if such features, systems, articles, materials, kits, and/or methods are not mutually inconsistent, is included within the scope of the present disclosure.

Claims

I/We Claims

A system (100) for proactive seizure prediction and alerting, comprising:
an array of non-invasive sensors (102) configured to acquire continuous physiological data, including at least an Electroencephalogram (EEG), an Electrocardiogram (ECG), an Electromyography (EMG), and an Electrooculography (EOG); and
a central monitoring unit (104) comprising:
a feature extraction unit operative to process the acquired physiological data to isolate features indicative of potential seizure activity;
a privacy unit comprising a noise injector operative to apply a calculated amount of random noise to the extracted features to provide differential privacy while preserving patterns indicative of seizures;
a seizure classification and prediction unit implementing a DeepConvEEGgNet technique, configured with:
multiple Conv2D layers for feature detection,
batch normalization layers for input normalization,
maxPooling layers for spatial size reduction of the representation, and
a dense neural network with global average pooling and dropout rate mechanisms for classification and prediction of seizure events;
a confidence evaluation mechanism configured to assess the prediction confidence using a threshold mechanism; and
a real-time alert unit operative to communicate alerts to healthcare providers or caregivers when the prediction exceeds the confidence threshold.
The system (100) of claim 1, wherein the EEG data and EOG data is collected through a wearable helmet comprising:
a headgear lining;
a pair of flexible arms;
a plurality of EEG electrode arrays embedded within the helmet lining;
EOG sensor placements near the eyes, embedded within flexible arms that gently press against the temples to detect eye movement without causing discomfort.
a central processing unit that amplifies and filters the EEG and EOG signals before transmission.
a wireless communication module for transmitting processed data to a central monitoring unit (104).
adjustable straps and cushioning for secure placement and patient comfort.
The system (100) of claim 1, wherein the EMG data and EOG is collected by a wristband comprising:
a flexible, adjustable strap that comfortably fits around the patient’s wrist.
embedded ECG electrodes on the inner surface of the wristband that make contact with the skin for cardiac activity monitoring.
EMG sensors located on the strap to align with the muscles of interest on the forearm for electromyography.
a miniaturized signal processing unit within the wristband to condition and digitize the ECG and EMG signals.
a rechargeable battery with wireless charging capability for long-term continuous operation.
an integrated wireless transmitter for relaying the collected ECG and EMG data to a central monitoring unit (104).
The system (100) of claim 1, wherein the feature extraction unit utilizes machine learning techniques to enhance feature isolation from the physiological data.
The system (100) of claim 1, wherein the headgear lining comprise a reservoir to store a conductive gel, and discharge mechanisms to apply the conductive gel on surface of sensor that come in contact of skull.
The system (100) of claim 1, wherein the noise injector is further configured to adjust the level of noise based on a privacy-utility trade-off technique.
The system (100) of claim 1, wherein the DeepConvEEGgNet technique includes Conv2D layers with ReLU activation functions.
The system (100) of claim 1, wherein the MaxPooling layers are configured to follow each batch normalization layer to enhance feature invariance to scale and orientation changes.
The system (100) of any preceding claims, wherein the confidence evaluation mechanism utilizes a dynamic threshold that is adaptive based on historical prediction accuracy.
The system (100) of any preceding claims, further comprising an intervention unit equipped to automatically administer medication upon the triggering of an alert.

PROACTIVE SEIZURE PREDICTION AND ALERTING SYSTEM WITH PRIVACY-PRESERVING DATA ANALYSIS

The present disclosure provides a system for proactive seizure prediction and alerting. The system comprises an array of non-invasive sensors configured to acquire continuous physiological data, including at least an Electroencephalogram (EEG), an Electrocardiogram (ECG), an Electromyography (EMG), and an Electrooculography (EOG); and a central monitoring unit. The central monitoring unit includes a feature extraction unit operative to process the acquired physiological data to isolate features indicative of potential seizure activity; a privacy unit comprising a noise injector operative to apply a calculated amount of random noise to the extracted features to provide differential privacy while preserving patterns indicative of seizures; a seizure classification and prediction unit implementing a DeepConvEEGgNet technique, configured with multiple Conv2D layers for feature detection, batch normalization layers for input normalization, maxPooling layers for spatial size reduction of the representation, and a dense neural network with global average pooling and dropout rate mechanisms for classification and prediction of seizure events; a confidence evaluation mechanism configured to assess the prediction confidence using a threshold mechanism; and a real-time alert unit operative to communicate alerts to healthcare providers or caregivers when the prediction exceeds the confidence threshold.
, Claims:I/We Claims

A system (100) for proactive seizure prediction and alerting, comprising:
an array of non-invasive sensors (102) configured to acquire continuous physiological data, including at least an Electroencephalogram (EEG), an Electrocardiogram (ECG), an Electromyography (EMG), and an Electrooculography (EOG); and
a central monitoring unit (104) comprising:
a feature extraction unit operative to process the acquired physiological data to isolate features indicative of potential seizure activity;
a privacy unit comprising a noise injector operative to apply a calculated amount of random noise to the extracted features to provide differential privacy while preserving patterns indicative of seizures;
a seizure classification and prediction unit implementing a DeepConvEEGgNet technique, configured with:
multiple Conv2D layers for feature detection,
batch normalization layers for input normalization,
maxPooling layers for spatial size reduction of the representation, and
a dense neural network with global average pooling and dropout rate mechanisms for classification and prediction of seizure events;
a confidence evaluation mechanism configured to assess the prediction confidence using a threshold mechanism; and
a real-time alert unit operative to communicate alerts to healthcare providers or caregivers when the prediction exceeds the confidence threshold.
The system (100) of claim 1, wherein the EEG data and EOG data is collected through a wearable helmet comprising:
a headgear lining;
a pair of flexible arms;
a plurality of EEG electrode arrays embedded within the helmet lining;
EOG sensor placements near the eyes, embedded within flexible arms that gently press against the temples to detect eye movement without causing discomfort.
a central processing unit that amplifies and filters the EEG and EOG signals before transmission.
a wireless communication module for transmitting processed data to a central monitoring unit (104).
adjustable straps and cushioning for secure placement and patient comfort.
The system (100) of claim 1, wherein the EMG data and EOG is collected by a wristband comprising:
a flexible, adjustable strap that comfortably fits around the patient’s wrist.
embedded ECG electrodes on the inner surface of the wristband that make contact with the skin for cardiac activity monitoring.
EMG sensors located on the strap to align with the muscles of interest on the forearm for electromyography.
a miniaturized signal processing unit within the wristband to condition and digitize the ECG and EMG signals.
a rechargeable battery with wireless charging capability for long-term continuous operation.
an integrated wireless transmitter for relaying the collected ECG and EMG data to a central monitoring unit (104).
The system (100) of claim 1, wherein the feature extraction unit utilizes machine learning techniques to enhance feature isolation from the physiological data.
The system (100) of claim 1, wherein the headgear lining comprise a reservoir to store a conductive gel, and discharge mechanisms to apply the conductive gel on surface of sensor that come in contact of skull.
The system (100) of claim 1, wherein the noise injector is further configured to adjust the level of noise based on a privacy-utility trade-off technique.
The system (100) of claim 1, wherein the DeepConvEEGgNet technique includes Conv2D layers with ReLU activation functions.
The system (100) of claim 1, wherein the MaxPooling layers are configured to follow each batch normalization layer to enhance feature invariance to scale and orientation changes.
The system (100) of any preceding claims, wherein the confidence evaluation mechanism utilizes a dynamic threshold that is adaptive based on historical prediction accuracy.
The system (100) of any preceding claims, further comprising an intervention unit equipped to automatically administer medication upon the triggering of an alert.

PROACTIVE SEIZURE PREDICTION AND ALERTING SYSTEM WITH PRIVACY-PRESERVING DATA ANALYSIS