System and method for tracking blood biochemistry using a BIO-sensing method and calibration thereof

Abstract

The present disclosure relates to a system (100) for tracking blood biochemistry using a bio-sensing method and calibration. The system (100) may include a memory (204) and at least one processor (202) that communicates with the memory. The at least one processor (202) is configured to perform three key functions: first, receives neurophysiological signals and / or user physiological data through a receiving module; second, obtains a biochemistry estimation model that defines relationships between neurophysiological signals and multiple health markers associated with brain activity; and third, generates a tuned biochemistry estimation model specific to the user by utilizing the received signals and physiological data.

Description

SYSTEM AND METHOD FOR TRACKING BLOOD BIOCHEMISTRY USING A BIO-SENSING METHOD AND CALIBRATION THEREOF FIELD OF THE INVENTION

[0001] The present disclosure generally relates to non-invasive blood biochemistry monitoring system. More particularly, the present disclosure relates to a system and a method for tracking blood biochemistry using a bio-sensing method and calibration thereof.BACKGROUND

[0002] The information in this section merely provides background information related to the present disclosure and may not constitute prior art(s) for the present disclosure,

[0003] C onventional blood biochemistry monitoring systems, such as systems for monitoring sodium or glucose levels in blood, often rely on invasive methods that require blood samples which cause discomfort and inconvenience to a patient. Conventional continuous blood biochemistry' techniques like Continuous Glucose Monitors (CGM) also require invasive techniques wherein needles are used that are uncomfortable and painful. They also require to be replaced in periodic intervals causing discomfort.

[0004] Conventional biosensing methods like EEG can detect the neurological changes correlated with biochemical fluctuations. For example, abnormal sodium levels, such as hyponatremia, may manifest in distinct EEG patterns, such as diffuse slowing of normal background frequency. This is often observed as a shift towards slower theta and delta rhythms, indicating an impairment of brain function due to metabolic disturbance. Similarly, change in glucose levels, like hyperglycemia and hypoglycemia, leads to specific EEG alterations, particularly in theta and alpha frequency bands, as the brain reacts to a reduction in glucose which is a primary' energy source for the brain. However, achieving clinical accuracy using the non-invasive methods remains challenging.

[0005] Further, when someone abruptly stops or significantly reduces their long term heavy alcohol use, they may experience seizures as part of alcohol withdrawal syndrome.This occurs due to the brain's adaptation to alcohol's effects over time. Alcohol acts as a depressant on the central nervous system and enhances Gamma-aminobutyric acid (GABA), a neurotransmitter that inhibits brain activity while suppressing glutamate, which excites neurons. With prolonged alcohol use, the brain becomes less sensitive to the GABA and more responsive to the glutamate. Upon sudden alcohol cessation, this adapted brain chemistry' is disrupted. This imbalance creates a state of neuronal hyperexcitability that may trigger seizures, usually generalized tonic-clonic in nature, affecting the entire brain and causing full-body convulsions.

[0006] Thus, there is a need for a user-friendly system and method that can effectively monitor blood biochemical fluctuations which are indicators of health risks, and enable early detection of health-related conditions.

[0007] The drawbacks / difficulties or the disadvantages / limitations of conventional techniques explained in the background section are just for exemplary’ purposes and the disclosure would never limit its scope only such limitations. A person skilled in the art would understand that this disclosure and below mentioned description may also solve other problems or overcome the other drawbacks / disadvantages of the conventional arts which are not explicitly captured above.SUMMARY

[0008] This summary' is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention. This summary is neither intended to identify key or essential inventive concepts of the invention nor is it intended for determining the scope of the invention.

[0009] The present disclosure relates to a system and a method for tracking blood biochemistry’ using biosensing techniques like a non-invasive electroencephalography (EEG) system and calibration thereof. The present disclosure may be configured to work with invasive brain sensing techniques that involve placing electrodes inside the cranium like Electrocorticography (ECoG), subcutaneous EEG, Stereo EEG (SEEG) electrodes, an array of microelectrodes, etc., and calibration thereof to track the blood biochemistry changes and dynamics.

[0010] In one example, a system for tracking blood biochemistry may include a memory and at least one processor that communicates with the memory'. The processor is configured to perform three key functions: first, it receives neurophysiological signals and / or user physiological data through a receiving module; second, it obtains a biochemistry estimation model that, defines relationships between neurophysiological signals and multiple health markers associated with brain activity; and third, it generates a tuned biochemistry estimation model specific to the user by utilizing the received signals and physiological data.

[0011] In an embodiment, the system may calibrate the neurophysiological signals (EEG signals) through selective invasive blood measurements. The present disclosure may also include continuously capturing EEG signals produced by the brain of a subject and extracting features linked to specific biochemical states using the captured EEG signals. The present disclosure may further include periodically updating and refining a machine learning-based calibration model with data from the invasive blood measurements to maintain accuracy and adapt to individual physiological differences.

[0012] To further clarify the advantages and features of the present invention, a more particular description of the invention will be rendered by reference to specific embodiments thereof, as explained in the accompanying section.BRIEF DESCRIPTION OF THE DRAWINGS

[0013] These and other features, aspects, and advantages of the present invention will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:

[0014] Figure 1 illustrates an architecture of a system that may be utilized for tracking blood biochemistry using a biosensing method and calibration thereof, in accordance with an embodiment of the present disclosure; and

[0015] Figure 2 illustrates a detailed schematic of the system, according to an embodiment of the present disclosure;

[0016] Figure 3 illustrates a wearable device 102 configured as a headband 302 positioned around the cephalic region of the subject, according to an embodiment of the present disclosure;

[0017] Figure 4 illustrates a wearable device 102 configured as an earbud-style biosensing apparatus designed for placement within and around the ear region, according to an embodimen t of the present disclosure;

[0018] Figure 5 illustrates a wearable device configured as eyewear designed to function as a biosensing apparatus for neurophysiological monitoring, according to an embodiment of the present disclosure.

[0019] Further, skilled artisans will appreciate that elements in the drawings are illustrated for simplicity and may not have necessarily been drawn to scale. Furthermore, in terms of the construction of the device, one or more components of the device may have been represented in the drawings by conventional symbols, and the drawings may show only those specific details that are pertinent to understanding the embodiments of the present invention so as not to obscure the drawings with details that will be readily apparent to those of ordinary' skill in the art having the benefit of the description herein.DETAILED DESCRIPTION

[0020] For the purpose of promoting an understanding of the principles of the present disclosure, reference will now be made to the various embodiments and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the present disclosure is thereby intended, such alterations and further modifications in the illustrated system, and such further applications of the principles of the present disclosure as illustrated therein being contemplated as would normally occur to one skilled in the art to which the present disclosure relates.

[0021] It will be understood by those skilled in the art that the foregoing general description and the following detailed description are explanatory' of the present disclosure and are not intended to be restrictive thereof.

[0022] Whether or not a certain feature or element was limited to being used only once, it may still be referred to as “one or more features” or “one or more elements” or “at least one feature” or “at least one element.” Furthermore, the use of the terms “one or more” or “at least one” feature or element do not preclude there being none of that feature or element, unless otherwise specified by limiting language including, but not limited to, “there needs to be one or more...” or “one or more elements is required.”

[0023] Reference is made herein to some “embodiments.” It should be understood that an embodiment is an example of a possible implementation of any features and / or elements of the present disclosure. Some embodiments have been described for the purpose of explaining one or more of the potential ways in which the specific features and / or elements of the proposed disclosure fulfil the requirements of uniqueness, utility, and non-obviousness.

[0024] Use of the phrases and / or terms including, but not limited to, “a first embodiment,” “a further embodiment,” “an alternate embodiment,” “one embodiment,” “an embodiment,” “multiple embodiments,” “some embodiments,” “other embodiments,” “further embodiment”, “furthermore embodiment”, “additional embodiment” or other variants thereof do not necessarily refer to the same embodiments. Unless otherwise specified, one or more particular features and / or elements described in connection with one or more embodiments may be found in one embodiment, or may be found in more than one embodiment, or may be found in all embodiments, or may be found in no embodiments. Although one or more features and / or elements may be described herein in the context of only a single embodiment, or in the context of more than one embodiment, or in the context of all embodiments, the features and / or elements may instead be provided separately or in any appropriate combination or not at all. Conversely, any features and / or elements described in the context of separate embodiments may alternatively be realized as existing together in the context of a single embodiment.

[0025] Any particular and all details set forth herein are used in the context of some embodiments and therefore should not necessarily be taken as limiting factors to the proposed disclosure.

[0026] The terms “comprises”, “comprising”, or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process or method that comprises a list of steps does not include only those steps but may include other steps not expressly listed or inherent to such process or method. Similarly, one or more devices or sub¬ systems or elements or structures or components proceeded by “comprises... a” does not, without more constraints, preclude the existence of other devices or other sub-systems or other elements or other structures or other components or additional devices or additional sub-systems or additional elements or additional structures or additional components.

[0027] Embodiments of the present disclosure will be described below in detail.

[0028] Figure 1 illustrates an architecture of a system that may be utilized for tracking blood biochemistry using a biosensing method and calibration thereof, in accordance with an embodiment of the present disclosure. The architecture may comprise a system 100 that may interact with a wearable device 102 worn by the user 104, arranged in a manner that facilitates continuous acquisition and processing of neurophysiological signals.

[0029] The wearable device 102 may be operatively coupled to System 100 and adapted to be worn by user 104. The wearable device may include one or more biosensing brainwave sensors configured to capture neurophysiological signals such as electroencephalography (EEG) or electrocorticography (ECoG) signals. These signals may be transmitted to the system 100 for processing, feature extraction, and biochemical estimation.

[0030] The system 100 may include one or more processors and memory modules configured to execute instructions for analyzing the captured signals. The system may implement algorithms for mapping extracted EEG features - such as spectral power and time-domain characteristics - to specific biochemical states. Further, the architecture may include a calibration module that periodically updates a machine learning-basedbiochemistry estimation model using invasive measurements, thereby maintaining accuracy and adapting to individual physiological variations.

[0031] The architecture shown in Figure 1 may represent a closed-loop arrangement wherein the wearable device 102 may serve as a sensing interface for continuous neurophysiological monitoring, while the system 100 performs computational processing and calibration to enable non-invasive estimation of biochemical parameters to identify a type of biochemical anomaly. This architecture may support dynamic recalibration and adaptive learning, enhance reliability and user comfort while minimizing invasive procedures.

[0032] In an embodiment, the system 100 may be configured to generate a tuned calibration model that is personalized for the user 104. The calibration model may be derived by processing neurophysiological signals captured by the wearable device 102 in combination with user-specific physiological data, such as invasive or non-invasive measurements. The system 100 may employ predefined machine learning algorithms to map extracted EEG features - such as spectral power, frequency-domain, and time¬ domain characteristics - to corresponding biochemical states. By continuously refining this calibration model through periodic updates and adaptive learning techniques, the system 100 may ensure that identification of type of biochemical anomaly remain accurate and tailored to the individual’ s unique phy siological profi le. Thi s personalization may enhance predictive reliability and optimize therapeutic interventions.

[0033] Figure 2 illustrates a detailed schematic block diagram of the system 100, in accordance with an embodiment of the present disclosure. For instance, the system 100 may include a processor 202, a memory 204, module(s) 206, and data 208. The memory 204, in one example, may store the instructions to carry' out the operations of the modules 206. The modules 206 and the memory 204 may be coupled to the processor 202.

[0034] The processor 202 can be a single processing unit or several units, all of which could include multiple computing units. The processor 202 may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processor, central processing units, state machines, logic circuitries, and / or any devices that manipulate signals based on operational instructions. Among other capabilities, theprocessor 202 is configured to fetch and execute computer-readable instructions and data stored in the memory 204.

[0035] The memory 204 may include any non-transitory computer-readable medium known in the art including, for example, volatile memory 204, such as static random¬ access memory (SRAM) and dynamic random-access memory (DRAM), and / or non¬ volatile memory, such as read-only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.

[0036] The module(s) 206, amongst other things, includes routines, programs, objects, components, data structures, etc., which perform particular tasks or implement data types. The modules 206 may also be implemented as, signal processor 202(s), state machine(s), logic circuitries, and / or any other device or component that manipulated signals based on operational instructions.

[0037] Further, the modules 206 can be implemented in hardware, instructions executed by a processing unit, or by a combination thereof. The processing unit can comprise a computer, a processor, such as the processor 202, a state machine, a logic array, or any other suitable devices capable of processing instructions. The processing unit can be a general -purpose processor 202 which execu tes instructions to cause the general -purpose processor 202 to perform the required tasks or, the processing unit can be dedicated to performing the required functions. In another embodiment of the present disclosure, the modules 206 may be machine-readable instructions (software) which, when executed by a processor 202 / processing unit, perform any of the described functionalities. Further, the data serves, amongst other things, as a repository for storing data processed, received, and generated by one or more of the modules 206. The data 208 may include information and / or instructions to perform activities by the processor 202,

[0038] The module(s) 206 may perform different functionalities which may include, but may not be limited to, receiving information and denoising the image frame. Accordingly, the module(s) 206 may include a receiving module 210, a feature extraction and analysis module 212, a data processing module 214, a calibration module 216, an adaptive feedback module 218, and a user interface module 220. In one example, the at least oneprocessor 202 may be configured to operate by actuating the aforementioned module(s) 206.

[0039] The system 100 may be configured as a modular computing architecture that ingests, processes, and interprets neurophysiological signals obtained from a sensing interface, and it may combine those signals with user physiological data acquired via a data acquisition unit to learn and update relationships between brain-derived features and blood biochemistry markers. In accordance with Figure 2, system 100 may include at least one processor 202 and memory 204 in operative communication, a plurality of modules 206 implemented as software, firmware, hardware logic, or hybrid components, and data 208 repositories that persist raw signals, feature vectors, model parameters, calibration records, alignment indices, confidence metrics, administration events, timestamps, and audit trails. The architecture may be realized locally on a gateway device paired to wearable device 102, partly on wearable device 102 itself, or in a hybrid topology that pairs edge computation with secured cloud-assisted inference, without departing from the functional scope described herein or the claim mapping that associates features with their actor modules.

[0040] As mentioned before, the wearable device 102 may be adapted to be worn by user 104 and may include a biosensing brainwave sensor configured to continuously capture neurophysiological signals such as EEG or ECoG, depending on clinical context and deployment constraints. In an embodiment, electrodes may be positioned on the scalp or within and around the ear canal to realize an ear-EEG configuration, and impedance analysis, contact quality monitoring, and motion artifact detection may be performed by device firmware to preserve signal integrity. The device may stream digitized signals at predetermined sampling rates with per-channel calibration, and it may apply on-device packetization or compression to ensure reliable transport to receiving module 210. Device-side timestamps, temperature readings, battery metrics, and motion indexes may be exposed to receiving module 210 to assist in synchronization, quality gating, and dowmstream correction, thereby supporting robust end-to-end processing.

[0041] In one example, the receiving module 210 may acquire instantaneous neurophysiological signals continuously or at pre-set intervals and may attach system¬ side timestamps at first ingress to establish time lineage across the pipeline. The modulemay perform preliminary conditioning such as DC offset removal, band-limiting, and notch filtering to reduce power-line interference and high-frequency noise while preserving the band structure relevant to brain rhythms. Receiving module 210 may ingest user physiological data obtained from the data acquisition unit -represented by invasive measurements such as blood glucose or sodium and non -invasive measurements including pulse oximetry' or self-testing kits - and it may normalize units, validate ranges, and stamp measurements with acquisition metadata including device type, site of collection, and operator annotations. In scenarios involving pharmacological intervention, receiving module 210 may register a signal indicative of administration time, dosage, route, and product identifiers so that subsequent changes in neurophysiological signals may be appropriately contextualized for model update and interpretation.

[0042] Further, the feature extraction and analysis module 212 may operate on the signal data prepared by the receiving module 210, computing multi-layered feature sets tailored to reveal physiological correlates within the neurophysiological signals. Spectral power may be computed for canonical bands (delta, theta, alpha, beta, gamma) using methods that may include periodograms, Welch’s averaging, or autoregressive spectral estimation, with normalization to handle inter-session amplitude variability. Frequency-domain features may include peak frequency within defined bands, band ratios (e.g., theta / alpha), spectral entropy to capture distribution uniformity, and spectral edge measures that indicate cumulative power thresholds. Time-domain features may include amplitude statistics such as mean, variance, skewness, and kurtosis, zero-crossing rates, event- related potential summaries, and temporal stability metrics. Complexity features may include approximate entropy, sample entropy, multiscale entropy, and fractal dimension measures such as Higuchi’s or Katz’s estimators, and, where multi-channel signals are available, connectivity descriptors such as coherence or phase-locking values may be computed to augment the feature space. Feature extraction and analysis module 212 may produce per-window feature vectors aligned to signal segments and may attach metadata including window length, quality indices, and synchronization markers, thereby facilitating robust downstream model operations.

[0043] According to the present disclosure, the system 100 may maintain a biochemistry estimation model that encodes learned relationships between neurophysiological signalfeatures and health markers associated with brain activity and systemic physiology. These health markers may include sodium, glucose, oxygen saturation (SpO2), calcium, magnesium, and hormone-related indicators. Calibration module 216 may instantiate and manage the biochemistry estimation model using an Artificial Intelligence (Al) technique, which may comprise convolutional neural networks for spatio-temporal pattern recognition, transformer-based architectures for context-aware interpretation, and ensemble regressors that fuse diverse feature families to produce robust marker estimates. Model parameters may be tracked and versioned in memory 204 and data 208 to enable reproducibility, rollback, and auditability. Calibration module 216 may apply regularization and guardrails to prevent overfitting, and it may enforce clinically safe responses to atypical data by bounding updates and providing confidence rationale narratives.

[0044] In one embodiment, the biochemistry estimation model may be a machine learning framework that includes, but is not limited to, an asynchronous federated learning model, large language models (LLMs), and convolutional neural networks (CNNs), to derive insights from combined local and cloud-based models. For each user, the system 100 employs, for example, the Edge Al technology to run local machine learning models that continuously analyze personal EEG signals and biochemical data, estimating key health markers such as glucose, sodium levels, and insulin sensitivity. In an embodiment, the local models, utilizing CNNs for pattern recognition and LLMs for context-aware data interpretation, may be periodically updated based on new invasive calibration data and evolving EEG patterns. The present disclosure may include aggregating anonymized data from multiple users into a cloud-based global model to improve the overall system’s predictive accuracy. The global model, leveraging federated learning, refines population-wide trends and insights that may not be identifiable from individual datasets, ensuring both personalized and collective accuracy in monitoring and managing diabetes.

[0045] The data processing module 214 may orchestrate inference by accepting neurophysiological signals and the corresponding feature vectors and by invoking the biochemistry estimation model to produce marker estimates. The module may incorporate sliding-window logic and overlap strategies to balance responsiveness with stability, andit may compute a continuous blood biochemistry composition for user 104 by fusing instantaneous estimates with user physiological calibration data using smoothing filters, Bayesian fusion, or state-space model approaches. Data processing module 214 may compare feature patterns and model outputs against expected norms, historical baselines, and therapy-aware trajectories to identify deviations that may constitute biochemical anomalies. These comparisons may rely on decision boundaries, statistical tests, and learned threshold functions that are adjusted per user by calibration module 216, thereby enabling precise identification of anomaly types while maintaining trustworthiness through logged rationale.

[0046] In an embodiment, the data processing module 214 may employ transformer-based architectures such as large language models (LLMs) to interpret complex biosignal and contextual data. The LLMs facilitate context-aware analysis by capturing patterns beyond simple biomarker estimation, offering a more detailed understanding of the user’ s health status. These models, with their larger context windows, are periodically updated with global and personalized data, ensuring they remain adaptive, robust, and effective in making predictions and guiding personalized therapeutic strategies.

[0047] Adaptive feedback module 218 may compute a confidence interval for each estimated marker and for the continuous biochemistry composition at window, session, and rolling-interval granularities. The confidence interval may be understood as a delta between a degree of accuracy in identifying the type of biochemical anomaly versus a minimum level of accuracy in identifying the type of biochemical anomaly. The adaptive feedback module 218 may integrate signal quality indicators from receiving module 210, feature stability measures from feature extraction and analysis module 212, posterior uncertainty from the biochemistry estimation model, alignment with known physiological trends (e.g., circadian rhythms), pharmacological context, and consistency across redundant signal channels if available. The adaptive feedback module 218 may compare confidence interval against a threshold — static or dynamically governed — and may generate a calibration signal to trigger model update whenever confidence is less than the threshold or indicates drift beyond an acceptable tolerance band. Confidence governance may be logged with justifications and summary traces stored in data 208 for clinical audit and interpretability, and may be rendered by user interface module 220 in forms suitable for review.

[0048] The calibration module 216 may process calibration signals by aggregating newly available user physiological data from the data acquisition unit, recent neurophysiological windows, and contextual inputs such as meal composition and timing, exercise patterns, sleep intervals, and therapy administration events, and it may use these inputs to adjust model parameters. The calibration module 216 may apply incremental learning when new data volumes are modest, transfer learning when leveraging population-level patterns, or Bayesian updating when uncertainty quantification and prior knowledge are critical. Over time, these updates may produce a tuned biochemistry estimation model personalized for user 104, with parameter sets and decision boundaries reflecting the user’s unique physiological profile. The tuned biochemistry estimation model may include risk-aware guards to ensure updates do not degrade performance under atypical conditions, and it may maintain dynamic envelopes for marker trajectories to support early-warning triggers.

[0049] In an embodiment, the tuned biochemistry estimation model may be adapted through incremental learning, federated learning, or probabilistic updating based on newly acquired invasive glucose reference measurements. The tuned biochemistry estimation model may tuned by dynamically adjusting model weights to account for physiological drift, environmental variability, or user-specific neurophysiological patterns, thereby enhancing robustness and accuracy in estimating glucose concentration over time.

[0050] The data processing module 214 may implement automated retro-calibration (ARC) to facilitate calibration when user physiological data is obtained asynchronously relative to neurophysiological monitoring. ARC may begin with timestamp normalization, whereby receiving module 210 and data processing module 214 may collect timestamps from both sources — neurophysiological signal streams from wearable device 102 and physiological measurements from the data acquisition unit — and normalize them to a common reference such as the gateway’s system time. For, ARC the data processing module 214 may estimate clock drift by comparing known synchronization points such as device handshake events logged during session start, periodic heartbeat signals exchanged between wearable device 102 and system 100, andmetadata associated with invasive measurement uploads. The difference between expected and actual offsets over time may provide a drift rate, and the data processing module 214 may apply dynamic offset correction to align timestamps using linear interpolation for uniform drift or polynomial / spline-based correction for non-linear drift patterns. Once drift is quantified, data processing module 214 may correct timestamps and extract neurophysiological signal windows corresponding to the physiological measurement period, with optional expansion of the window to absorb uncertainty and application of feature similarity checks to validate alignment.

[0051] The feature extraction and analysis module 212 may compute features for the aligned windows, and calibration module 216 may pair these feature sets with the physiological measurements to create labeled calibration instances. Calibration module 216 may then update the biochemistry estimation model using incremental learning or Bayesian updating to incorporate the retroactively aligned data, thereby correcting drift, improving personalization, and maintaining accuracy without requiring real-time co-recording. ARC may also log drift correction parameters, alignment confidence scores, update rationales, and any residual uncertainty envelopes into data 208 for compliance and traceability, and user interface module 220 may present alignment summaries to clinicians upon request.

[0052] The system 100 may further leverage stimulus feedback measurement and calibration to reduce dependence on invasive measurements, by introducing controlled stimuli known to provoke predictable physiological responses that are detectable in neurophysiological signals. For glucose monitoring, ingestion of a standardized carbohydrate load may serve as a stimulus, and for sodium monitoring, controlled intake of sodium-containing solutions may be used. Receiving module 210 may record neurophysiological signals continuously around stimulus administration windows, and feature extraction and analysis module 212 may detect changes in bands such as alpha and theta associated with metabolic processing. Calibration module 216 may compare observed EEG responses with model expectations and adjust parameters and thresholds accordingly, ensuring that non-invasive estimations remain accurate.

[0053] According to an embodiment, the system 100 may optionally integrate Rapid Invisible Frequency Tagging (RIFT) to enhance stimulus-response detection bypresenting high-frequency visual or auditory cues imperceptible to the user yet capable of eliciting measurable steady-state neural responses in neurophysiological signals. These RIFT-evoked responses may be captured by wearable device 102 and analyzed for synchronization with metabolic changes, so that calibration module 216 may utilize RIFT-linked neurophysiological markers to refine biochemistry estimation models with high temporal precision. By combining biochemical stimuli with RIFT, system 100 may achieve robust calibration without invasive sampling, improving accuracy and user comfort while enabling closed-loop adjustments when adaptive feedback module 218 indicates widening confidence intervals.

[0054] In another embodiment, the present disclosure may include utilizing ear- EEG technology (e.g., biosensing earphones) to capture the EEG signals through electrodes strategically placed within and around the ear canal. This embodiment can be in the form of a Biosensing Earphone, Smart Glass / VR headset, Headphone or a neckband.

[0055] In embodiments involving diabetes management, system 100 may quantify insulin effectiveness using the tuned biochemistry model that correlates neurophysiological features with glucose reduction dynamics. The model may be expressed by defining the glucose change as ΔG(t) = G(0) − G(t), where G(0) is the baseline blood glucose and G(t) is the blood glucose level at time t after insulin administration, and by relating this change to neurophysiological features and dose via the following relation:ΔG(t) = η · Id· f(EEGf(t), t) …… (1)

[0056] In equation (1), Idrepresents the administered insulin dose (units), η represents the insulin effectiveness coefficient that may be dynamically adjusted over time to account for tolerance buildup or reduced responsiveness, and f(EEGf(t), t) represents a function mapping neurophysiological spectral changes and their temporal evolution to glucose regulation dynamics. Calibration module 216 may continuously recalibrate η and, where appropriate, the functional mapping f by analyzing deviations between expected and observed ΔG(t) and by reconciling these deviations with ARC-aligned measurement pairs and stimulus-response windows. In doing so, system 100 may optimize therapeuticdosing, improve prediction fidelity for hypoglycemia risk, and enhance confidence governance through adaptive feedback module 218.

[0057] The data processing module 214 may orchestrate end-to-end inference and anomaly identification by accepting instantaneous neurophysiological signals, passing the corresponding feature vectors through the biochemistry estimation model to produce marker estimates, and comparing these outputs to historical baselines, expected physiological ranges, pharmacological-aware profiles, and pattern constraints. Where deviations exceed learned bounds, data processing module 214 may classify anomaly types—for example, signatures indicative of hyponatremia via diffuse slowing and increased low-frequency power, hypoxia via modulations in delta / theta bands, electrolyte imbalances via shifts in complexity features, and endocrine-linked changes via alpha / beta rhythm alterations — and it may signal early warnings, recalibration prompts, or measurement suggestions through user interface module 220, all with associated confidence scores derived by adaptive feedback module 218.

[0058] In one example, user interface module 220 may present real-time estimations, historical trajectories, confidence intervals, calibration prompts, ARC alignment summaries, RIFT stimulus-response overlays, pharmacological administration logs, and anomaly flags to user 104 and authorized clinicians. It may support role-based access control, privacy mechanisms, and consent management aligned to applicable regulations, with export options for clinical review in standardized formats. The interface may allow user input of contextual information including meal timing and composition, activity levels, sleep intervals, and subjective symptoms, and calibration module 216 may incorporate these inputs as covariates or priors. The interface may also surface rationale narratives for confidence changes, calibration triggers, drift correction parameters, and ARC alignment quality, thereby enhancing transparency and enabling clinician oversight of model evolution.

[0059] To strengthen robustness, system 100 may incorporate artifact mitigation strategies spanning hardware and software. Wearable device 102 may provide electrode impedance checks, motion detection, and environmental compensation; receiving module 210 may implement glitch filters and dropout handling; feature extraction and analysis module 212 may apply independent component analysis for ocular or muscular artifactsor supervised artifact classifiers trained on labeled datasets; and data processing module 214 may reject or down- weight windows failing quality thresholds. Adaptive feedback module 218 may reduce confidence when artifacts compromise feature reliability and may prompt recalibration only after artifact-free periods to avoid spurious updates. ARC may reflect artifact influences by incorporating alignment confidence scores that consider signal quality in the target window and by expanding windows cautiously when drift or artifacts could obscure true physiological correspondence.

[0060] According to the present disclosure, the tuned biochemistry estimation model may be used by the system 100 to process subsequent instantaneous to other kinds of biochemical anomalies based on the identification of associated health marker. Exemplary embodiments highlighting different health markers are now explained.

[0061] In an embodiment, the system 100 may estimate a health marker, such as blood sodium concentration by continuously analyzing instantaneous neurophysiological signals (having EEG signals) to detect neurophysiological patterns associated with sodium imbalance. Variations in blood sodium levels, including hyponatremia or hypernatremia, induce measurable changes in cortical electrical activity characterized by diffuse slowing, increased theta and delta power, and alterations in neural synchrony. In such an example, the feature extraction and analysis module 212 may extract spectral and temporal EEG features that correlate with neuronal osmotic stress responses, and mapping these features to sodium concentration levels using the tuned biochemistry estimation model. The tuned biochemistry estimation model may be periodically updated through invasive or non-invasive reference sodium measurements and adaptive machine learning algorithms in a manner explained above, thereby enabling accurate, non-invasive estimation of sodium fluctuations and supporting early detection of sodium-related neurological / seizure risks.

[0062] In another embodiment, the system 100 may estimate another health marker, such as blood glucose concentration across a physiological range by analyzing neurophysiological signatures in instantaneous neurophysiological signals (having EEG signals) that arise from glucose-dependent modulation of cortical metabolic activity. Variations in glucose levels, encompassing hypoglycemic, normoglycemic, andhyperglycemic states, produce measurable changes in cortical electrophysiology, including shifts in alpha, theta, and low-beta power, alterations in phase-amplitude coupling, and modifications in transient event-related oscillatory responses. In one example, the feature extraction and analysis module 212 may process the EEG signals using spectral decomposition, time-frequency analysis, and functional connectivity metrics to extract features indicative of glucose-influenced neuronal energetics, such as reduced alpha coherence, enhanced theta synchronization, or attenuation of higher- frequency components under metabolic stress.

[0063] Thereafter, the data processing module 214 may map extracted features to glucose concentration values using the tuned biochemistry estimation model trained with invasive glucose reference measurements. The data processing module 214 may employ the tuned biochemistry estimation model trained to estimate glucose concentration by processing EEG features associated with glucose-dependent neurophysiological activity. To further improve the determination, the feature extraction and analysis module 212 may extract a plurality of EEG-derived parameters, including but not limited to spectral band power, time–frequency coefficients, entropy measures, phase synchronization indices, and functional connectivity graphs, and providing said parameters as inputs to a predefined machine learning model. The tuned biochemistry estimation model may comprise one or more deep learning architectures such as convolutional neural networks, recurrent neural networks, transformer-based models, or hybrid models configured to learn non-linear mappings between EEG features and glucose concentration values.

[0064] In yet another embodiment, the system 100 may estimate the glucose marker utilizing a controlled glucose stimulus to refine or validate the glucose estimation model. IN such an approach, a predefined quantity of glucose or carbohydrate-containing material as a stimulus known to elicit, a predictable metabolic response is administered to the user. At the same time, the wearable device 102 may continuously record EEG signals during the pre-stimulus, absorption, and post-absorption phases and communicate the same to the receiving module 210. In response to the stimulus, EEG signals may exhibit measurable changes in oscillatory dynamics, including alterations in alpha desynchronization, modulation of theta-band activity, and changes in event-related spectral perturbations associated with cerebral glucose uptake. The data processing module 214, upon feature extraction by the feature extraction and analysis module 212,may compare the observed EEG response with expected model-derived glucose trajectories and adjusting the calibration model when discrepancies exceed a predefined threshold. Such stimulus-feedback calibration enables the system to account for individualized metabolic processing rates and ensures long-term precision in glucose estimation without requiring synchronous invasive measurement at each calibration instance.

[0065] In an embodiment, the system 100 may estimate another health marker, such as effects of metabolic therapies on blood biochemistry’, including but not limited to metformin and GLP-1 receptor agonists. The system 100 may detect EEG changes associated with the uptake, action, and efficacy of these medications, and estimating real¬ time blood glucose dynamics from the corresponding neurophysiological patterns. In some embodiments, the system may generate alerts for optimal medication timing based on the detected EEG signatures, including signatures associated with glucose metabolism, insulin sensitivity, and appetite regulation. Thus, the system 100 may precisely manage both metformin therapy and GLP-1 receptor agonist administration by optimizing dosing and timing based on real-time physiological responses.

[0066] In an embodiment, the system 100 may estimate another health marker, such as blood alcohol concentration by analyzing EEG signatures that arise due to alcohol- induced modulation of cortical excitability and neurotransmission. Alcohol alters inhibitory and excitatory balance within the brain, resulting in measurable electrophy siological effects including reductions in alpha-band coherence, enhancement of low-frequency theta activity, suppression of higher-frequency beta rhythms, and disruptions in phase synchrony across cortical regions. During rising alcohol levels, the EEG may exhibit attenuated event-related desynchronization and reduced peak oscillatory amplitude, while declining alcohol levels or withdrawal states may present with heightened cortical hyperexcitability, increased delta-theta coupling, and instability in functional network connectivity. Hereto, feature extraction and analysis module 212 may extract spectral power distributions, phase-amplitude coupling metrics, entropy measures, and connectivity -based graph features from the EEG signals, and map these features to estimated blood alcohol concentration values through the tuned biochemistry estimation model trained using invasive or breath-based alcohol reference measurements. As mentioned before, the tuned biochemistry estimation model may be periodicallyrefined through incremental or retroactive learning to accommodate individual variability in alcohol metabolism, tolerance, and neurophysiological response, thereby enabling accurate, real-time estimation of blood alcohol dynamics and early detection of neurological instability associated with withdrawal or intoxication.

[0067] In an embodiment, the system 100 may estimate another health marker, such as blood oxygenation status by analyzing EEG signatures indicative of altered cerebral oxygen delivery arising from variations in oxygenated hemoglobin percentage, reduced cerebral blood flow, diminished total hemoglobin concentration, or other physiological conditions affecting oxygen transport. Changes in oxygen availability, whether caused by hypoxemia, ischemia, anemia, or impaired perfusion, produce characteristic neurophysiological effects including increased delta and theta activity, attenuation of higher-frequency bands, reductions in cortical complexity, and disruptions in functional connectivity patterns, the feature extraction and analysis module 212 may extract EEG- derived parameters such as spectral power ratios, time–frequency coefficients, non-linear complexity metrics, and connectivity indices sensitive to oxygen-dependent, metabolic modulation of neural tissue. The tune biochemistry estimation model may be configured to map these EEG features to oxygenation-relevant parameters, which may include SpO2values, relative oxygen delivery capacity, or oxygenation anomaly indicators, using invasive or non-invasive reference measurements for periodic recalibration. Thus, the system 100 may perform non-invasive estimation of cerebral oxygenation dynamics and detection of oxygen-related abnormalities that may arise independently of peripheral saturation levels, including conditions attributable to reduced hemoglobin concentration or impaired blood flow.

[0068] In an embodiment, the system 100 may estimate yet another health marker, such as blood calcium levels by analyzing EEG signatures associated with hypocalcemia or hypercalcemia. Abnormal calcium levels alter neuronal excitability and synaptic transmission, resulting in EEG changes such as high-amplitude slow-wave activity, increased cortical irritability, and shifts in theta and delta frequency bands. Hereto, the feature extraction and analysis module 212 may extract EEG spectral and temporal characteristics indicative of calcium-dependent neuronal excitability and the data processing module 214 may correlate these features with invasive calcium measurements using a personalized calibration model. This enables continuous, non-invasive estimationof calcium concentration levels and supports early detection of calcium-related neurological instability.

[0069] In an embodiment, the system 100 may estimate another health marker, such as blood magnesium concentration by detecting EEG alterations arising from hypomagnesemia or magnesium-related neuronal dysfunction. Magnesium plays a critical role in neurotransmission and NMDA receptor regulation, and deviations from normal levels produce EEG features such as increased cortical hyperexcitability, reduced inhibitory modulation, and susceptibility to seizure-like activity, the feature extraction and analysis module 212 may extract EEG features representative of magnesium-dependent neuronal behavior and the data processing module 214 may process the extract features using the tuned biochemistry estimation model trained through invasive magnesium measurements to estimate magnesium concentration non-invasively.

[0070] In an embodiment, the system 100 may estimate monitor pituitary / function as well, specifically focusing on detecting patterns associated with anomalies of hypopituitarism. The system 100 may identify the EEG patterns indicative of altered cognitive processing and neural activity characteristic of hormonal deficiencies due to pituitary dysfunction. This includes analyzing variations like spectral slowing, reduced alpha coherence, altered theta activity during wakefulness, disruptions in sleep spindle dynamics, as well as changes in event-related oscillatory responses that correspond to pituitary-related physiological states.

[0071] According to the present disclosure, the system 100 may be used for tracking a source of energy metabolism in the blood using neurophysiological signals (for example, the EEG signals, the ECoG signals, etc.). The system 100 may include continuously monitoring the brain activity through instantaneous neurophysiological signals (EEG signals) of the user and detecting neural patterns associated with shifts in energy substrate utilization. By analyzing specific frequency bands and neural oscillations extracted by the feature extraction and analysis module 212, the data processing module 214 may identify transitions between different metabolic states, such as energy derived from the breakdown of glucose, glycogen (glucagon-induced), ketone, fat, or muscle protein. The data processing module 214 may, in a manner explained above, process instantaneous neurophysiological signals using the biochemistry estimation model to distinguishbetween distinct metabolic pathways, recognizing changes in brainwave patterns corresponding to each energy source.

[0072] In one example, the data processing module 214 may detect transitions between distinct metabolic fuel-source states by analyzing characteristic variations in the user’s neurophysiological signals. Such metabolic states may include glucose-dominant metabolism, glycogen-derived glucose metabolism, ketone-dominant metabolism, systemic fat-oxidation metabolism, and protein-derived metabolism. Each of these states may produce differentiable EEG signatures arising from altered neuronal energy availability and mitochondrial acetyl-CoA production pathways. For example, glucose- dominant and glycogen-derived glucose states may be associated with stable alpha-band activity and balanced theta-alpha ratios, whereas early systemic fat-oxidation may manifest as increased theta power, reduced alpha coherence, and elevated low-frequency variability indicative of transitional energy stress. In contrast, ketone-dominant metabolism may be characterized by increased alpha stability, reduced delta-theta activity, higher entropy measures, and improved neural-efficiency metrics. Protein- derived glucose metabolism or late-deficit states may exhibit elevated theta activity, diminished alpha power, reduced signal complexity, and increased spectral variability.

[0073] To further facilitate the data processing module 214, the feature extraction and analysis module 212 may extract spectral, temporal, complexity, and neural -efficiency features from the EEG signals and processing the extracted features using a predefined metabolic-state estimation model to determine the active metabolic fuel-source state. The data processing module 214 may further utilize the tuned biochemistry estimation model to associate user-specific EEG patterns with corresponding biochemical pathways, thereby generating real-time, continuous insights into the user’s metabolic dynamics, including fat-oxidation status, caloric-deficit state, metabolic flexibility, and fuel-source switching behavior. Based on the detected metabolic fuel-source state, the data processing module 214 may generate metabolic insights and provide optimized dietary, nutritional- timing, fitness, or therapeutic recommendations. When paired with the wearable device 102 such as an ear-EEG biosensing earphone, smart glass, headphone, headband, or other head-worn device, the present disclosure offers personalized fat-loss guidance, muscle-gain optimization, and metabolic-state coaching derived from real-time neurophysiological monitoring.

[0074] Figure 3 illustrates a wearable device 102 configured as a headband 302 positioned around the cephalic region of the subject, according to an embodiment of the present disclosure. The headband is designed to integrate one or more electrodes 304 that maintain optimal contact with the scalp, enabling continuous acquisition of neurophysiological signals such as EEG. This arrangement ensures high-quality signal capture while offering comfort and stability for real-time monitoring and data transmission to the processing system for biochemical estimation and calibration.

[0075] In a non-limiting example, the headband 302 may correspond to any head-worn accessory adapted for seamless integration of biosensing components. Its design facilitates stable electrode placement, minimizes motion artifacts, and supports real-time data transmission to the processing system for biochemical estimation and calibration. This configuration offers a practical, non-invasive solution for continuous monitoring of brain activity and related health markers.

[0076] Figure 4 illustrates a wearable device 102 configured as an earbud-style biosensing apparatus designed for placement within and around the ear region, according to an embodiment of the present disclosure. The wearable device 102 integrates electrodes 402 positioned to capture neurophysiological signals such as EEG, supported by an electrode holder 404 that ensures stable contact and minimizes motion artifacts. A housing 406 encloses the structural components, providing secure placement and durability while maintaining user comfort. This configuration enables continuous signal acquisition and real-time data transmission to the processing system for biochemical estimation and calibration, offering a compact, non-invasive solution for brain activity monitoring.

[0077] Figure 5 illustrates a wearable device 102 configured as eyewear designed to function as a biosensing apparatus for neurophysiological monitoring, according to an embodiment of the present disclosure. The frame 506 supports multiple integrated electrodes 502 strategically positioned along the inner surfaces of the temples and near the contact points to ensure stable acquisition of EEG signals. These electrodes arearranged to maintain consistent contact with the skin around the cephalic region, enabling accurate and continuous signal capture during regular use.

[0078] The wearable device 102 also incorporates a data acquisition unit 504 embedded within the frame, which processes and transmits the captured neurophysiological signals to the connected system for real-time biochemical estimation and calibration. This configuration combines practicality with advanced sensing capabilities, offering a non-invasive solution that seamlessly integrates into everyday eyewear. By minimizing motion artifacts and ensuring user comfort, the design supports continuous monitoring of brain activity and related health markers without disrupting normal routines.

[0079] According to the present disclosure, deployment topologies may vary according to regulatory and operational constraints. In a pure edge configuration, inference and calibration may occur on a local gateway paired to wearable device 102, minimizing data movement outside the user’s control domain. In a hybrid configuration, anonymized features or parameter gradients may be exchanged with a secure cloud service to obtain population-level improvements via asynchronous federated learning. Calibration module 216 may selectively incorporate cloud-delivered parameter updates into the tuned biochemistry estimation model whenever updates are consistent with the user’s profile and pass confidence validation, while adaptive feedback module 218 may monitor post-update stability by inspecting confidence trajectories and drift signals. ARC operations may continue to function under intermittent connectivity by performing local alignment with later reconciliation upon synchronization, thereby preserving calibration fidelity across connectivity disruptions.

[0080] Security and privacy governance may be enforced across modules. Data in transit may be encrypted, data at rest may be protected with hardware-backed keys, and access to sensitive artifacts may be restricted through role-based permissions and consent gates. ARC alignment may operate primarily on timestamps, session identifiers, and derived features, avoiding unnecessary exposure of raw neurophysiological signal content beyond protected boundaries. Audit records maintained in data 208 may include model version histories, calibration events, ARC alignment factors and confidence scores, RIFT stimulus sessions, pharmacological administration logs, and user consent metadata,facilitating compliance and enabling reconstruction of decision pathways for clinical assessment.

[0081] Operational flows may begin with acquisition by wearable device 102 and ingestion by receiving module 210, continue through feature computation by feature extraction and analysis module 212, and proceed to inference and composition calculation by data processing module 214 using the biochemistry estimation model under the selected Al technique. Adaptive feedback module 218 may assess confidence and generate calibration signals when thresholds are not met, and calibration module 216 may incorporate user physiological data, ARC-aligned pairs, and stimulus-response windows -including those enhanced by RIFT - to update parameters and yield the tuned biochemistry estimation model. The updated model may then drive subsequent estimations with improved alignment to user physiology and therapy context, and user interface module 220 may report these outputs in real time with comprehensive rationale narratives and alignment quality indicators.

[0082] Over extended monitoring horizons, system 100 may perform scheduled recalibrations even when confidence remains adequate, accounting for gradual changes in electrode placement, device aging, seasonal influences, or alterations in lifestyle. Calibration module 216 may employ change-point detection to identify shifts in feature distributions and initiate targeted recalibrations only for affected sub-models, thereby preserving stability in unaffected components. Adaptive feedback module 218 may log confidence evolution, compute exponentially weighted moving averages, and adjust thresholds to stabilize user experience while maintaining sensitivity to real physiological changes. ARC may use refined drift estimation and nonlinear correction policies to sustain alignment quality across multi -week recording campaigns, and RIFT may be scheduled periodically to provide stimulus-anchored benchmarks for neurophysiological responsiveness without perceptual burden.

[0083] In therapy-heavy contexts, system 100 may implement agent-aware guardrails that prevent model instability when consecutive administrations lead to overlapping pharmacodynamic effects. Receiving module 210 may tag overlapping epochs; feature extraction and analysis module 212 may segment windows to isolate additive and washout phases; data processing module 214 may reduce aggressiveness of parameterchanges during confounded periods; and calibration module 216 may defer major updates until unconfounded data is available, using ARC to recover calibration opportunities post¬ event. Adaptive feedback module 218 may check for lowered confidence during these periods, raise verification prompts for the data acquisition unit, and restore thresholds once ARC indicates sufficient alignment quality,

[0084] In one example, wearable device 102 may include firmware routines that coordinate sampling rates with expected physiological events, extending battery life while preserving signal quality. The device may switch to higher sampling during high-value periods—such as post-therapy windows or overnight monitoring -and may down-sample during low-variance intervals. The receiving module 210 may record such state changes and ensure downstream feature extraction accounts for sampling variability through resampling or band-specific normalization. The device-system handshake may include integrity checks to protect against packet loss and reconstruct missing windows when feasible, with ARC alignment logic resilient to minor gaps through window-expansion strategies and alignment confidence indexing.

[0085] Data formats within system 100 may be standardized to ensure interoperability across modules and deployment environments. Neurophysiological signals may be represented as multi-channel arrays with timestamps and metadata headers; features may be serialized as typed vectors with schema identifiers; model parameters may be stored with version hashes and training provenance; calibration events may be logged with inputs, outputs, and policy references; and confidence records may contain quantitative indices and qualitative narratives. Data processing module 214 may maintain lineage across transformations, and user interface module 220 may expose filtered views appropriate for clinical review and user comprehension.

[0086] From a clinical workflow perspective, system 100 may integrate with electronic health records via secure interfaces that transfer summary estimations, confidence measures, ARC alignment summaries, RIFT session markers, and calibration events. Clinicians may review proposed recalibrations, annotate anomaly episodes, and schedule follow-up measurements through the data acquisition unit, and calibration module 216 may incorporate clinician feedback as constraints or priors, enabling medical oversight to shape the tuned biochemistry estimation model in high-risk cases. Adaptive feedbackmodule 218 may provide time-indexed confidence histories that support clinical decisions regarding therapy titration, dietary modifications, or additional diagnostics.

[0087] The industrial applicability of system 100 may include diabetes management including gestational diabetes, where neurophysiological signatures complement invasive glucose measurements to reduce burden, electrolyte surveillance where early detection of sodium, calcium, or magnesium imbalances may prevent adverse events, hypoxia monitoring where SpO2estimations may guide activity and recovery, and endocrine tracking where subtle neurophysiological changes may prompt timely clinical attention. The combination of continuous sensing, ARC-enabled calibration with explicit clock-drift correction, RIFT-enhanced stimulus calibration, pharmacological agent-aware modeling, and confidence governance may result in a system that is both user-friendly and clinically meaningful while remaining adaptable across diverse environments.

[0088] While specific language has been used to describe the present disclosure, any limitations arising on account thereto, are not intended. As would be apparent to a person in the art, various working modifications may be made to the method in order to implement the inventive concept as taught herein. The foregoing description give examples of embodiments. Those skilled in the art will appreciate that one or more of the described elements may well be combined into a single functional element. Alternatively, certain elements may be split into multiple functional elements. Elements from one embodiment may be added to another embodiment.

Claims

I claim:

1. A system (100) comprising:a memory (204); andat least one processor (202) in communication with the memory, the at least one processor (202) configured to:receive at least one of neurophysiological signals and user physiological data;receive a biochemistry estimation model, wherein the biochemistry estimation model includes relationships between the neurophysiological signals and a plurality of health markers associated with brain activity;generate a tuned biochemistry estimation model for the user using the at least one of neurophysiological signals and the user physiological data.

2. The system (100) as claimed in claim 1, wherein the at least one processor (202) is configured to:receive instantaneous neurophysiological signals of user; compare the instantaneous neurophysiological signals with the tuned biochemistry estimation model; andidentify a type of biochemical anomaly based on the processing of the instantaneous neurophysiological signals using the tuned biochemistry estimation model.

3. The system (100) as claimed in claim 1, wherein the user physiological data includes one or more of a user input, an invasive measurement test of the user, and a non- invasive measurement test.

4. The system (100) as claimed in claim 2, wherein, to compare the instantaneous neurophysiological signals, the at least one processor (202) is configured to:process instantaneous neurophysiological signals and the tuned biochemistry estimation model using an Artificial Intelligence (Al) technique.

5. The system (100) as claimed in claim 2, wherein, to process the instantaneous neurophysiological signals, the at least one processor (202) is configured to:generate signal data from the instantaneous neurophysiological signals, extract features from the signal data, the features comprising spectral power, frequency-domain, time-domain, entropy and / or fractal-dimension metrics; map extracted features to blood biochemistry estimations for one or more health markers including sodium, glucose, blood alcohol content, oxygen saturation, calcium, magnesium, and / or hormone-related indicators; and compare extracted features with biochemistry estimation model to identify the type of biochemical anomaly.6, The system (100) as claimed in claim 2, wherein the at least one processor (202) is configured to:determine, for each blood biochemistry estimation, a confidence interval, wherein the confidence interval is indicative of deviation in accuracy in identifying the type of anomaly;compare the confidence interval with a confidence threshold; and generate a calibration signal to initiate the generation of the tuned biochemistry estimation model when the confidence interval is less than the confidence threshold.

7. The system (100) as claimed in claim 5, wherein the at least one processor (202) is configured to:receive one of user input and data from the user physiological data, and update the biochemistry estimation model, using the Al technique, based on the user physiological data, such that a subsequent confidence interval is more than the calibration threshold.

8. The system (100) as claimed in claim 7, wherein, the at least one processor (202) is configured to:determine the confidence interval of the biochemistry estimation model periodically after lapse of predefined intervals.

9. The system (100) as claimed in claim 7, wherein, to update the biochemistry estimation model, the at least one processor (202) is configured to:implement an Automated Retro-Calibration (ARC) that retroactively aligns user physiological data with corresponding segments of continuously recorded neurophysiological signals; andupdate the biochemistry estimation model based on the alignment of user physiological data with corresponding segments of continuously recorded neurophysiological signals.

10. The system (100) as claimed in claim 7, wherein the at least processor (202) is configured to:receive a signal indicative of administration of pharmacological agent to alter biochemical state of the user;record a change in neurophysiological signals caused by the pharmacological agent; andupdate the biochemistry estimation model based on the recorded change in neurophysiological signals.

11. The system (100) as claimed in claim 1, wherein the at least one processor (202) is configured to:process the neurophysiological signals and the tuned biochemistry estimation model using the user physiological calibration data in combination with the Artificial Intelligence technique to compute a continuous blood biochemistry composition for the user.

12. The system (100) as claimed in claim 1, comprising a wearable device having a biosensing brainwave sensor adapted to continuously capture instantaneous neurophysiological signals, the signals including at least one of Electroencephalography (EEG) and Electrocorticography (ECoG)13. A system (100) comprising:a memory (204); andat least one processor (202) in communication with the memory, the at least one processor (202) configured to:receive at least one of neurophysiological signals and user physiological data;extract, from the at least one neurophysiological signal, a plurality of EEG features including spectral power distribution, time–frequency coefficients, entropy measures, and neural -efficiency metrics;process the extracted EEG features using a tuned biochemistry estimation model;estimate, using the metabolic-state estimation model, a brain metabolic fuel-source state selected from glucose-dominant metabolism, glycogen- derived glucose metabolism, ketone-dominant metabolism, fat-derived systemic metabolism, and protein-derived metabolism;generate, from the estimated metabolic fuel-source state, at least one metabolic insight indicative of user’s fat-oxidation status, caloric-deficit state, and metabolic flexibility, andoutput one or more recommendations or alerts to optimize fat-loss, nutritional timing, or therapeutic intervention based on the generated metabolic insight.

Images