Minimally Invasive Glucose Status Systems, Devices, and Methods

Abstract

A glucose system including a scalp-worn minimally invasive device including first and second sensors, the system adapted to determine one or more glucose states based on sensed EEG signals using the scalp-worn device.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Application No. 63 / 378,834, filed October 7, 2022, and U.S. Provisional Application No. 63 / 381,078, filed October 26, 2022, the entire disclosures of which are incorporated herein by reference for all purposes. Incorporation by Reference

[0002] All publications and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

[0002]

[0003] This application incorporates by reference in their entirety US Pat. Nos. 6,572,542, 8,118,741, and 1,102,035 for all purposes.

[0003]

[0004] This application incorporates by reference WO 2023 / 183798 in its entirety and for all purposes.

[0005] By way of example only, glucose monitor and / or insulin administration devices and uses thereof, such as those described in U.S. Pat. Nos. 9,585,607, 10,827,956, 11,744,943, 11,672,422, and 11,064,917, are hereby incorporated by reference in their entirety for all purposes.

[0004]

[0006] WO 2023 / 034820 is incorporated herein by reference in its entirety for all purposes. The entire disclosure of U.S. Provisional Application No. 63 / 238583, filed August 30, 2021, which claims priority to PCT International Application PCT / US2022 / 075695 (WO 2023 / 034820), is also incorporated herein by reference in its entirety for all purposes. Huang, Y., Wang, JB, Parker, JJ et al. "Spectro-spatial features in distributed human intracranial activity proactively encode peripheral metabolic activity," Nat Commun 14, 2729 (2023). https: / / doi.org / 10.1038 / s41467-023-38253-7, which is also incorporated herein by reference in its entirety for all purposes. [Background technology]

[0005]

[0007] Blood glucose is one of the most important blood parameters to measure, and abnormal values ​​can cause serious complications. For example, blood glucose levels above or below normal can cause serious medical problems, including medical emergencies.

[0006] Hypoglycemia, which requires medical attention, is a condition in which blood sugar (glucose) levels are below the normal range, often considered as low as 70 milligrams per deciliter (mg / dL) or 3.9 millimoles per liter (mmol / L). While hypoglycemia can be related to diabetes management, hypoglycemia can also occur in association with non-diabetic conditions and / or settings (e.g., ICU patients). Treatment for hypoglycemia may include high-sugar foods and drinks (e.g., juice) and / or medications to raise blood glucose levels. If left untreated, hypoglycemia can lead to loss of consciousness and seizures.

[0007]

[0002] Hyperglycemia is a condition in which blood glucose levels are higher than the normal range, often considered to be greater than 180 milligrams per deciliter (mg / dL). Hyperglycemia, commonly associated with diabetes, occurs when the body produces too little insulin (the hormone that transports glucose into the blood) or when the body is unable to use insulin properly. However, hyperglycemia can also be associated with non-diabetic conditions and / or settings (e.g., ICU patients). If left untreated for extended periods, hyperglycemia can damage nerves, blood vessels, tissues, and organs.

[0008]

[0003] Conventional technology can test current blood glucose levels in near real time. For example, a continuous glucose monitor ("CGM") includes a small sensor inserted under the skin, such as on the abdomen or arm. The sensor measures interstitial glucose ("ISG") levels, which indicate blood glucose levels. The sensor can test ISG every few minutes. A transmitter can wirelessly transmit information to a monitor, which may be part of an insulin pump or a separate device such as a smartphone or tablet. CGMs essentially detect existing or current blood glucose levels. Also, because ISG levels track blood glucose levels within a few minutes, blood glucose values ​​provided by CGMs actually represent past blood glucose levels. Insulin pumps can automatically initiate insulin infusion if glucose levels become too high. A threshold glucose value is set (stored) in the CGM, so insulin can be administered when glucose levels reach the pre-set threshold. Individualized thresholds can be set for patients, and thresholds may need to be reset over time (currently during outpatient visits) if changes in the patient's life require the threshold to be reset (e.g., as they reach puberty). Currently, the optimal insulin pump for diabetes management achieves a peripheral glucose target range of 70-180 mg / dL for <70% of the time, but this is not ideal and the glycemic range is still quite wide.

[0009]

[0004] Additionally, there are non-diabetic conditions and circumstances in which it is desirable or necessary to know blood glucose levels. By way of example only, intensive care unit ("ICU") patients, whether diabetic or not, typically have their blood levels checked frequently, usually using blood drawn from a finger prick that can be tested with a small strip equipped with a meter that indicates the glucose level.

[0010]

[0005] U.S. Patent No. 6,572,542 describes utilizing ECG and EEG signals to determine whether a hypoglycemic event is occurring or imminent. While determining whether a glycemic event is occurring (e.g., using a CGM) or is imminent is useful, it may generally be necessary to determine real-time or near-real-time glucose status and / or be able to predict future blood glucose values ​​or the risk of a future glycemic event in advance (e.g., further in advance than an impending event). Summary of the Invention [Means for solving the problem]

[0011] One aspect of the present disclosure is a method for predicting a future glucose state of a subject, the method including: non-invasively sensing EEG signals from an ear hook location on the scalp of the subject; inputting the EEG signals or processed EEG signals to a trained computer-executable method trained to predict the subject's future glucose state; predicting the subject's future glucose state based at least in part on the non-invasively sensed EEG signals; and outputting instructions to initiate a communication based on and in response to the predicted future glucose state. This aspect may include any other appropriately combinable methods or steps herein.

[0012] One aspect of the present disclosure is a computer-executable method stored on a non-transitory medium that, when executed by a processor, is adapted to perform the following steps: receiving as input non-invasively sensed EEG data or processed EEG data from ear hook locations on a subject's scalp; predicting a future glucose state of the subject based at least in part on the EEG data or processed EEG data; and initiating an output adapted to communicate information indicative of the subject's predicted future glucose state based on and responsive to the subject's predicted future glucose state. This aspect may include any other suitably combinable method herein.

[0013] One aspect of the present disclosure is a non-transitory computer-readable storage medium having instructions stored thereon, the instructions executable by a processor to perform a method, the method including receiving as input non-invasively sensed EEG data or processed EEG data from an ear hook location on a subject's scalp, predicting a future glucose state of the subject based at least in part on the EEG data or processed EEG data, and initiating an output adapted to communicate information indicative of the subject's predicted future glucose state based on and responsive to the subject's predicted future glucose state. This aspect may include any other suitably combinable devices, features, and / or methods herein.

[0014] One aspect of the present disclosure is a method for predicting a future glucose state of a subject, the method including: non-invasively sensing EEG signals with a single-channel EEG sensor on the scalp of the subject; inputting the EEG signals or processed EEG signals to a trained computer-executable method trained to predict the subject's future glucose state; predicting the subject's future glucose state based at least in part on the non-invasively sensed EEG signals; and outputting instructions to initiate communication based on and in response to the predicted future glucose state. This aspect may include any other appropriately combinable methods herein.

[0015] One aspect of the present disclosure is a computer-executable method that, when executed by a processor, is adapted to perform the following steps: receiving as input EEG data or processed EEG data sensed non-invasively using a single-channel EEG sensor on a subject's scalp; predicting a future glucose state of the subject based at least in part on the EEG data or processed EEG data; and initiating an output adapted to communicate information indicative of the subject's predicted future glucose state based on and in response to the subject's predicted future glucose state. This aspect may include any other suitably combinable method herein.

[0016] One aspect of the present disclosure is a system for predicting a subject's future glucose state. The system includes a wearable EEG sensor configured and arranged to be wearable on the scalp via ear hooks, and a computer-executable method stored on a non-transitory medium of a personal device that, when executed by a processor in the personal device, is adapted to perform the following steps: receiving non-invasively sensed or processed EEG data as input from the EEG sensor; predicting the subject's future glucose state based at least in part on the EEG data or the processed EEG data; and initiating an output adapted to communicate information indicative of the subject's predicted future glucose state based on and in response to the subject's predicted future glucose state. This aspect may include any other appropriately combinable devices, features, and / or methods herein.

[0017]

[0014] One aspect of the present disclosure is a method for predicting a subject's future glucose level, the method including: sensing EEG signals from the subject using an ear-mounted EEG device; processing the sensed EEG signals; analyzing the processed EEG signals using a trained prediction model using an application on a personal device to predict the subject's future glucose level; and causing the personal device to visually present information indicative of the predicted future glucose level on a display. This aspect may include any other appropriately combinable method herein.

[0018] One aspect of the present disclosure is a computer-executable method stored in a non-transitory memory of a personal device, the method including receiving sensed EEG data from a subject or information indicative of the sensed EEG data from the subject, analyzing the processed EEG signals using a trained predictive model to predict future glucose levels for the subject, and causing the personal device to visually present information indicative of the predicted future glucose levels on a display. This aspect may include any other suitably combinable method herein.

[0019] One aspect of the present disclosure is a glucose prediction system (GFS). The system includes a minimally invasive EEG device including a first sensor and a second sensor, the minimally invasive EEG device being sized, configured, and adapted to be ear-wrapped by a subject and to sense EEG signals with the first and second sensors; and a personal device adapted to communicate with the EEG device, the personal device being further adapted to receive and process the sensed EEG signals or information indicative of the sensed EEG signals and to analyze the processed EEG signals using a trained prediction model to predict future glucose values ​​for the subject. This aspect may include any other appropriately combinable devices, features, and / or methods herein. In this regard, the personal device may be further adapted to visually present information indicative of the predicted future glucose values ​​or glucose states on a display of the personal device.

[0020] One aspect of the present disclosure is a method for training a computer-executable prediction method for predicting a future glucose state of a subject, the method including: noninvasively sensing training EEG signals from one or more subjects, processing the sensed training EEG signals, sensing or receiving training real-time or near real-time blood glucose values ​​or information indicative of the real-time or near real-time blood glucose values ​​from the one or more subjects, and correlating the sensed or processed training EEG signals with the sensed or received training real-time or near real-time blood glucose values ​​or information indicative of the real-time or near real-time blood glucose values ​​to train the method for predicting a future glucose state of the individual based on the noninvasively sensed EEG signals from the individual.

[0021] One aspect of the present disclosure is a computer-executable method (e.g., an "app") adapted to present predicted future blood glucose values ​​or glucose states, the computer-executable method being stored in non-transitory memory. The method includes receiving extracranially sensed EEG data or information indicative of the extracranially sensed EEG as input from a subject and causing a visual representation of the predicted future blood glucose values ​​to be displayed on a display of a device. The visual representation may include a graph showing time on a first axis and the predicted blood glucose values ​​or glucose states on a second axis. The visual representation may include textual information indicating times or time ranges during which blood glucose values ​​are likely to be undesirable. [Brief explanation of the drawings]

[0022] [Figure 1]

[0019] A system including an intracranial device is shown. [Figure 2]

[0020] FIG. 1 is a block diagram of an exemplary controller herein, aspects of which may be incorporated into any of the devices and systems herein. [Figure 3]

[0021] An exemplary flowchart of an exemplary process for predictively managing glucose levels is shown. [Figure 4] 1 is an exemplary flowchart of an exemplary process for predicting glucose levels based on brain activity data. [Figure 5]

[0023] 1 is a non-limiting example of a visual representation of a predicted blood glucose value or glucose state on the display of a device (optionally a personal device), such as a smartphone or other computing device. [Figure 6]

[0024] 1 illustrates an exemplary glucose prediction system including exemplary and non-limiting components. [Figure 7]

[0025] 1 illustrates an exemplary continuous glucose prediction system. [Figure 8]

[0026] A method is presented that includes using sensed interstitial glucose information (eg, ISF glucose values) to predict information indicative of a future glucose state. [Figure 9]

[0027] A method is presented that includes using sensed interstitial glucose information (e.g., ISF glucose values) to manage future glucose conditions. [Figure 10]

[0028] A method is shown that includes correlating or associating one or more aspects of the sensed interstitial glucose information with one or more aspects of the sensed EEG signal to create a correlation therebetween. [Figure 11]

[0029] 10 illustrates exemplary steps for calibrating or recalibrating a glucose monitor using at least one of EEG data or information indicative of sensed EEG data. [Figure 12]

[0030] 1 illustrates an exemplary method that includes calibrating or recalibrating a blood glucose quantification or blood glucose prediction method. [Figure 13]

[0031] 1 illustrates an exemplary bidirectional calibration method herein. [Figure 14]

[0032] 1 illustrates an exemplary method that includes predicting a subject's future glucose status. [Figure 15]

[0033] 1 illustrates an exemplary method for training a method for predicting an individual's future glucose state. [Figure 16]

[0034] A method is presented that includes predicting a future glucose status of a subject based at least in part on a non-invasively sensed EEG signal. [Figure 17]

[0035] 1 illustrates a portion of an exemplary personal device. [Figure 18]

[0036] 1 illustrates a portion of an exemplary system. [Figure 19]

[0037] Additional exemplary and optional components of any of the devices and systems (eg, personal devices and / or systems including personal devices) herein are shown. DETAILED DESCRIPTION OF THE INVENTION

[0023]

[0038] The present disclosure relates to a glucose prediction system and method of use that includes a behind-the-ear (or otherwise near-ear) scalp-worn EEG device (EEG sensor). Much of the disclosure of WO 2023 / 034820 (published March 9, 2023) is incorporated herein by reference in its entirety, including, but not limited to, the following paragraphs:

[0039]

[0055] and Figures 1-4 are expressly incorporated herein. This may (but does not necessarily) provide exemplary support and foundation for one or more aspects of the glucose prediction system and method of use herein, including a minimally invasive behind-the-ear, scalp-worn EEG sensor. The entire disclosure of U.S. Provisional Application No. 63 / 238583, filed August 30, 2021, which claims priority to PCT International Application PCT / US2022 / 075695 (International Publication No. WO 2023 / 034820), is also incorporated herein by reference in its entirety for all purposes. Huang, Y., Wang, JB, Parker, JJ et al., in connection with International Publication No. WO 2023 / 034820. "Spectro-spatial features in distributed human intracranial activity proactively encode peripheral metabolic activity." Nat Commun 14, 2729 (2023). The entire article (including, but not limited to, any methodology) at https: / / doi.org / 10.1038 / s41467-023-38253-7 is incorporated herein by reference for all purposes.

[0024] Metabolic syndrome and diabetes are increasingly prevalent health conditions, now affecting a broader age range of the global population. Specifically, the significant morbidity and mortality associated with diabetes pose a significant burden to healthcare systems, including enormous personal and societal costs in the form of medical expenses for the disease itself and lost workforce productivity due to disability associated with disease progression. Therefore, the ability to prevent the development of diabetes and its subsequent complications represents a high-impact public health area for intervention. Furthermore, the physiological regulation of dietary behavior and the balance of body metabolism and weight involves a complex interplay of hormonal signaling and behavior. In this context, close monitoring and control of blood glucose levels has been shown to be one of the optimal and most reliable methods for preventing both hypoglycemic and hyperglycemic complications.

[0025]

[0040] Blood glucose control is important beyond diabetes, as hyperglycemia and hypoglycemia in hospitalized and critically ill patients are associated with increased costs, length of hospital stay, morbidity, and morale. Patients, particularly those in intensive care units, can suffer from stress-related hyperglycemia as a result of severe injury or illness, such as traumatic brain injury (TBI), intracranial hemorrhage, stroke, or subarachnoid hemorrhage (SAH). Conservative glycemic control is associated with improved outcomes for these patients.

[0026]

[0041] Current iterations of continuous glucose monitors (CGMs) rely on interstitial glucose measurements as a surrogate for blood glucose levels. These have an inherent lag time and are prone to interference from medications and extreme blood glucose levels. Furthermore, traditional CGMs fail to anticipate abnormal glucose levels. As a reactive modality, traditional CGMs can only respond to hypoglycemia or hyperglycemia after an abnormality has occurred. Additionally, critically ill patients may experience chronically elevated glucose levels that are resistant to traditional treatment options, such as continuous insulin infusion. The systems and methods described herein attempt to remedy these limitations by predicting a patient's glucose levels over the next several hours. This information can be used to administer preemptive therapy to the patient to prevent harmful glucose levels from occurring.

[0027]

[0042] In many embodiments, the predictive glucose management (PGM) systems and methods described with respect to FIGS. 1-4 decode brain activity to predict a patient's future glucose levels. In various embodiments, brain activity is measured using a non-invasive modality such as electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), magnetoencephalography (MEG), or any other modality depending on the requirements of a particular application of the embodiments herein. However, brain activity can also be recorded using intracranial sensors, if available, such as, but not limited to, deep brain stimulation (DBS) systems or ECoG. In various embodiments, the PGM system is wearable or otherwise minimally invasive to a patient's life outside of the clinical setting.

[0028]

[0043] In many embodiments, the systems and methods described herein include closed-loop management of glucose levels, in which preemptive therapy is provided to a patient to avoid hyperglycemia or hypoglycemia. For example, a patient may be administered long-acting insulin, insulin analogs, and / or any other hyperglycemic control medication in anticipation of future glucose changes, depending on the requirements of a particular application. As a further example, brain stimulation may be provided to perturb the glucose-encoding network in the brain to alter blood glucose levels over the next few hours. Depending on the implantation site, brain stimulation may be provided by an already implanted DBS electrode, any other type of implanted electrode, or via a non-invasive brain stimulation modality, such as, but not limited to, transcranial magnetic stimulation (TMS), transcranial direct current stimulation (tDCS), transcranial focused ultrasound (tFUS), and / or any other modality depending on the requirements of a particular application. In various embodiments, the brain stimulation modality can be used as a supplement to insulin when a patient is refractory to standard treatment. The PGM system architecture is described in further detail below.

[0029]

[0044] The PGM system records and interprets brain activity to estimate a patient's likely glucose levels within the next few hours. Typically, predictions are accurate for at least two to eight hours, though this number may increase depending on the patient and condition. In many embodiments, the PGM system provides these predictions to the patient and / or a medical professional. However, in various embodiments, the PGM system is further capable of closed-loop glucose control by continuously predicting future glucose levels and modifying therapy (e.g., drug delivery rate, brain stimulation, etc.) to avoid predicted harmful glucose changes. In this way, the PGM can function as an artificial pancreas system with superior glucose management capabilities. In some embodiments, patient and / or medical professional permission is required before therapy is implemented and / or modified by the PGM.

[0030]

[0045] FIG. 1 illustrates an exemplary PGM system architecture according to an embodiment. PGM system 100 includes a brain activity recorder 110. In the illustrated embodiment, brain activity recorder 110 is a deep brain stimulation system. However, as can be readily appreciated, any brain activity recorder can be used, including non-invasive ones, as discussed above. In some embodiments, the brain activity recorder is a wearable device rather than an implanted device. PGM system 100 further includes a CGM 120. In many embodiments, a CGM is used to continuously verify the accuracy of interstitial blood glucose predictions and may also serve as a redundant alert modality. However, a CGM may not be present in all PGM systems depending on the requirements of a particular application.

[0031]

[0046] A controller 130 is communicatively coupled to the brain activity recorder 110, the CGM 120, and the insulin infusion pump 140. In many embodiments, communication between different components may not be direct. For example, the brain activity recorder may provide data to the CGM, which in turn provides data to the controller, rather than communicating directly with the controller. Indeed, as one skilled in the art will appreciate, any communication architecture may be used without departing from the scope or spirit of the disclosure herein.

[0032]

[0047] In many embodiments, the controller processes the recorded brain activity to generate a prediction regarding the patient's glucose level. The controller may provide a prediction for only one of interstitial glucose values ​​or blood glucose values. In some embodiments, predictions of both interstitial glucose values ​​and blood glucose values ​​are calculated. Furthermore, the controller may be implemented using any of a variety of computing platforms. In various embodiments, the controller is a smartphone, a smartwatch, a tablet computer, a personal computer, and / or any other personal wearable device. In some embodiments, the controller may be integrated into a medical device or a medical server system, such as a hospital computer network or a cloud medical system.

[0033]

[0048] In various embodiments, the insulin infusion pump can variably infuse insulin as directed by the controller. Additionally, depending on the patient's needs, other drugs besides insulin may be provided via similar infusion pumps. As can be readily appreciated, many PGM systems may not include an infusion pump if drug delivery is not advisable for a particular patient. Similarly, the PGM may further include a method for delivering brain stimulation as an alternative therapy. In various embodiments, the brain activity recorder may also function as a brain stimulation device. Indeed, depending on the requirements of a particular application, any number of different PGM system architectures may be used depending on the needs of a particular patient.

[0034]

[0049] Referring now to FIG. 2, a block diagram of one of the controllers herein is shown. Controller 200 includes a processor 210. In many implementations, the processor is a logic circuit capable of executing instructions, such as, but not limited to, a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), and / or any combination thereof. In many embodiments, multiple processors can be used. Controller 200 further includes an input / output (I / O) interface 220. The I / O interface can be used to communicate with various PGM system components and / or third-party components, such as, but not limited to, a display, a speaker, a CGM, a brain activity recorder, a stimulation device, an infusion pump, a mobile phone, a medical device, a computer, and / or any other component, via wired or wireless connections. Processor 210 may be any of the other processors described herein, or vice versa.

[0035]

[0050] The controller 200 further includes memory 230. The memory 230 can be volatile memory, non-volatile memory, or any combination thereof. The memory 230 includes a glucose management application 232. The glucose management application can instruct the processor to execute various PGM processes as described herein. In many embodiments, the memory 230 further includes brain activity data obtained from a brain activity recorder. The brain activity data can describe brain activity as a signal or a set of signals. In some embodiments, the brain activity data includes waveforms recorded by sensor electrodes. In various embodiments, one or more waveforms are recorded for each electrode (“channel”). In some embodiments, the brain activity data describes a spectral profile of broadband brain activity. In various embodiments, the glucose management application configures the processor to function as a multivariate decoder of the brain activity data. As can be readily appreciated, the controller can be implemented in a variety of ways using similar computing components without departing from the scope or spirit of the present disclosure. The PGM process is described in further detail below.

[0036]

[0051] Predictive Glucose Management. The PGM process involves collecting and using brain activity to predict a patient's future glucose levels. In many embodiments, treatment recommendations or treatment itself can result from the prediction of hyperglycemia or hypoglycemia to stabilize glucose levels in a healthier range. Peripheral glucose levels tend to follow circadian dynamics and are strongly aligned with intracranial high-frequency activity (HFA, 70-170 Hz) across multiple brain regions. Therefore, overall brain activity can be used in the predictive modeling process. In some embodiments, brain activity data from known glucose sensors, such as the hypothalamus, amygdala, and hippocampus, are used instead of or in combination with brain activity from other regions and / or the entire brain.

[0037]

[0052] A machine learning model can be trained to process data from one or more brain activity recorders. In some embodiments, the training process is performed using data obtained from the patient, and the trained model is used on that data. In various embodiments, the model can be pre-trained on standardized data, and training can be completed using patient data. In various embodiments, the model is continuously refined using predictions measured using CGM and subsequent validation. While linear models are often considered less predictive than more modern machine learning models, in many embodiments, a linear model is sufficient for accurate predictions. However, in various embodiments, more complex predictive machine learning models can be used, such as, but not limited to, other types of regression models, neural networks, etc., depending on the requirements of a particular application.

[0038]

[0053] FIG. 3 shows an example flowchart of an example PGM process for predictively managing glucose levels. Process 300 includes recording brain activity using a brain activity recorder (310) and providing the brain activity data to a trained predictive model (320). The predictive model predicts future glucose values ​​(330). In many embodiments, the certainty of the prediction may decrease the further into the future the prediction is made. In various embodiments, multiple predictions are provided at different points in time, and only those above a predetermined confidence threshold determined by a medical professional are used. In some embodiments, a hard limit is set on how far into the future the predictions are made. If a hypoglycemic and / or hyperglycemic condition is predicted, a medical intervention is provided (340) to avoid unhealthy dips or spikes in glucose levels, respectively. In various embodiments, the medical intervention is provided automatically, for example, via infusion pump control and / or brain stimulation. In various embodiments, an alert is provided informing the patient and / or medical professional that unhealthy glucose levels have been predicted. In some embodiments, confirmation is required before the medical intervention is provided.

[0039]

[0054] While a particular process is depicted in Figure 3, it is readily apparent that various modifications can be made without departing from the scope or spirit of the present disclosure. For example, pre-recorded brain activity data can be provided and used to make predictions. Furthermore, intervention need not be recommended or provided in all cases; in many situations, a warning alone may be beneficial.

[0040]

[0055] FIG. 4 is an exemplary flowchart of an exemplary PGM process for predicting glucose levels based on brain activity data. Process 400 includes generating a feature vector across all electrode (or sensor) channels (410). In many embodiments, all frequency bands across all channels are flattened into a single feature vector. A subset of features is selected from the feature vector (420) using a least absolute shrinkage and selection operator (LASSO) model and regularized (430). The regularized features are provided to a trained machine learning model (440) to generate one or more predictions. In many embodiments, a similar process is used to train the model using labeled training data from the patient and / or other patients. While a particular machine learning model is described herein, many different machine learning models can be used without departing from the scope or spirit of the present invention.

[0041]

[0056] The following disclosure, including exemplary FIG. 5, describes a system and method adapted to non-invasively predict future glucose levels using the location of a behind-the-ear scalp EEG sensor. The entire disclosure of U.S. Patent No. 1,102,035 is incorporated herein by reference for all purposes with respect to an exemplary wearable EEG sensor. Any aspect of this sensor can be incorporated into the EEG device (sensor) shown in FIG. 5 or any other wearable EEG sensor or sensing device herein.

[0042]

[0057] FIG. 5 illustrates an exemplary glucose forecasting system (“GFS”) that includes multiple components. The GFS includes a wearable EEG sensor (labeled “mini-EEG”) that can include first and second paired electrodes. The EEG is adapted, sized, and configured to be worn behind the ear and measure EEG signals from the scalp. The GFS system can also optionally include (but does not necessarily include) a stomach-worn interstitial glucose-sensing insulin pump (e.g., a glucose monitor (such as a continuous glucose monitor) and an insulin pump with an infusion site). The GFS can also include an application or “app” stored on a smartphone or other personal device (or smart wearable device) that is adapted and configured to continuously receive and analyze EEG and glucose data (and optionally administered insulin) to generate “predictions.” The “predictions” can optionally be used by the insulin pump to administer insulin. The app can optionally communicate with an online secure patient management portal. From this portal, doctors in remote clinics and patients at home or in clinics can view insulin dosing trends and EEG signal patterns and the relationship between them.

[0043]

[0058] The GFS herein can be adapted and configured to predict glucose values ​​and fluctuations up to six hours before actual value changes. When a glucose shift or change is predicted, the app can optionally communicate commands to an insulin pump to administer the appropriate amount of insulin needed before glucose values ​​become extreme. The GFS herein can optionally incorporate a glucose monitor (i.e., the GFS may not include a glucose monitor) until EEG-guided insulin titration is further validated for the individual. The GFS may also rely on or utilize additional standard monitoring technologies, such as finger-prick blood glucose monitoring. An exemplary advantage of the GFS over standard pumps is its ability to process trends hours before approaching dangerous levels (either high or low glucose values). In fact, hypoglycemia is the most common problem seen with insulin-based therapies today. The GFS herein allows for safer upper and lower predictive algorithm titration. Furthermore, an optional closed-loop approach can prevent glucose values ​​outside of dangerous ranges, optionally avoiding even the need to monitor one's own glucose levels. The GFS herein may also be adapted to inform optimal doses of long-acting insulin injections.

[0044]

[0059] The GFS herein may optionally be configured to provide information related to or regarding glucose predictions. By way of example only, an app may be adapted to display at least a portion of the glucose prediction and / or the predicted time or period during which the predicted glucose value will fall below or rise around a threshold on a screen of a personal device (e.g., phone, wearable).

[0045]

[0060] FIG. 5 is a non-limiting example of a visual representation of predicted blood glucose values ​​on a display of a device, such as a smartphone or other computing device. For example, the app may include a computer-executable method stored on a non-transitory medium (e.g., memory) adapted, when executed by a processor, to present predicted future blood glucose values. The method includes receiving extracranial sensed EEG data or information indicative of the extracranial sensed EEG as input from a subject (e.g., from a wearable EEG sensor behind the ear) and causing a visual representation of the predicted future blood glucose values ​​to be displayed on the device's display. In the exemplary FIG. 5, time is on the x-axis and blood glucose levels are on the y-axis. The left side of the x-axis can be considered 3 a.m., with predicted values ​​outside the range indicated. Visual indicators (e.g., red icons) can indicate predicted peaks and valleys. The range is user-adjustable on the display, allowing the user to change the high or low levels of the range indicated as "in range."

[0046]

[0061] As previously mentioned, existing closed-loop systems for detecting glucose levels and automatically administering insulin operate in real time. The closed-loop systems react to glucose peaks and dips as they change. The GSF described herein uses a wearable ear-mounted EEG sensor to sense EEG, predict shifts in glucose levels hours before they occur, and optionally titrate insulin based on this prediction before glucose levels become extreme. Furthermore, titrating insulin over time can dramatically reduce the risk of overcorrecting and causing symptomatic hypoglycemia, the most common and disabling side effect of all current insulin-based therapies.

[0047]

[0062] The following disclosure relates to the above disclosure in one or more ways. The following description can be integrated and combined with the above examples and embodiments, and vice versa.

[0048]

[0063] One aspect of the present disclosure relates to predicting blood glucose values. As used herein, the term "forecasting" generally refers to predicting or determining a subject's (e.g., human) future glucose value or glucose state (interstitial and / or blood) and / or risk of a future glycemic event, optionally one hour or more in advance. Prediction may be performed to predict a future glycemic event and, optionally, to prevent the predicted glycemic event from occurring (examples of which are described herein). A glycemic event generally refers to a glucose value or state above or below a particular value (or within certain predefined parameters), such as above or below a normal glucose value.

[0049]

[0064] The methods herein may optionally be adapted to provide a prediction including one or more generalized risk levels or risk indices for entering a future glycemic state within one or more specific time periods in the future. By way of example only, the prediction may include multiple risk indices, such as low risk, medium risk, and high risk. The optional risk indices may optionally be visually presented on a display (wherein the computer-executable method is stored on a non-transitory medium), such as a green light indicating a future time at low risk, a red light indicating a future time at high risk, and a yellow light indicating a future time at medium risk.

[0050]

[0065] Prediction herein may include sensing EEG data (by way of example only, optionally using scalp and / or sub-scalp (subgaleal) electrodes) or receiving sensed EEG data and analyzing the EEG data to predict future blood glucose values ​​or conditions and / or risk indicators. However, as described in more detail herein, it is contemplated that prediction may be made without sensing any EEG data. For example, if a CGM has been trained with sensed EEG data, an existing CGM may optionally be modified and adapted to predict future glucose values ​​and / or risk indicators based on an existing process of sensing ISG values. Furthermore, as part of the process of making a prediction, one or more patient parameters (with or without EEG data), such as, for example, without limitation, heart rate (HR), heart rate variability, skin conductance, blood pressure, body temperature, exercise level, etc., may be sensed / obtained and analyzed. Any of these patient parameters may be referred to herein as user input.

[0051]

[0066] In any of the methods herein, the optional EEG data can be sensed from the subject continuously. In any of the methods herein, the EEG data can be sensed from the subject periodically. In some examples, the EEG data is sensed periodically (e.g., every minute, every five minutes, etc.), and if a higher risk for future glycemic events is predicted, the EEG data can be sensed continuously (or relatively frequently) for a period of time.

[0052]

[0067] Prediction may optionally include providing a prediction of future blood glucose values ​​or risk indicators for a period of time into the future, such as one hour or more, such as 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, or more, which may be thought of as similar to a 7-day weather forecast.

[0053]

[0068] The disclosure herein describes exemplary methods of prediction and management. The disclosure also includes example devices and systems that may be adapted to implement one or more of the prediction and / or management methods. The devices and systems herein are exemplary, and it is contemplated that other device and system configurations may implement or perform one or more of the methods herein.

[0054]

[0069] While some of the disclosures herein may relate to predicting future blood glucose values ​​or risk indicators, one aspect of the disclosure herein relates to optional real-time (or near-real-time) blood glucose quantification. The term "real-time" as used herein also includes near-real-time detection / quantification, such as within 5 to 15 minutes of the actual glucose value. Real-time blood glucose values ​​can also be quantified using EEG data, optionally sensed using one or more scalp or sub-scalp electrodes. Thus, the systems herein can optionally be adapted to determine actual and / or predicted blood glucose values, or information indicative thereof. Thus, the systems herein can also be considered CGMs, similar to existing monitors.

[0055]

[0070] One or more aspects of the prediction methods herein may optionally be implemented or performed on a personal device, such as a smartphone, tablet, or smartwatch, which may include one or more processors adapted to execute one or more computer-executable methods / algorithms stored on a non-transitory medium on the personal device. For example, a computer-executable application ("app"), when executed by the processor, may cause the processor to receive raw and / or processed EEG data sensed from a subject (or information indicative of the raw or processed data) and may be adapted to perform a prediction process. In some alternatives, one or more processing steps may occur within a sensing device (such as a scalp or sub-scalp device), which is an example of how the methods herein may be implemented in one or more different devices.

[0056]

[0071] One or more aspects of the predicted and / or actual glucose values ​​or status may optionally be visually represented or presented on a display of a device (e.g., a smartphone, tablet, smartwatch, electronic ophthalmic device such as contact lenses, etc.). By way of example only, an executable application ("app") may be adapted to visually present risk indicators and / or predicted glucose values ​​for the next three hours (or other time), which may be updated (continuously or periodically) so that the prediction always includes predicted values ​​for the next three hours (or other time). Further, by way of example only, the app may be adapted to visually present a specific time (e.g., 4:17 PM) at which a glycemic event is predicted to occur. Further, by way of example only, the app may be adapted to present a timer with a countdown indicating the time remaining until the predicted glycemic event. The methods herein may optionally be adapted to provide "predictions" of blood glucose measurements for a relatively short period of time in advance (e.g., 1 hour, 2 hours, 3 hours in advance, etc.) and "risk predictions" of blood glucose for a longer period of time in advance (e.g., 10 hours, 11 hours, 12 hours in advance, etc.). Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to communicate actual and / or predicted future glucose values ​​and / or risk indicators to a patient / care team, optionally to one or more different devices (one example of which is described herein as a portal). Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to provide a user / patient with a relatively long-term (e.g., 10 hours or more) prediction of risk indicators, and may also optionally be adapted to provide suggested (e.g., optimal) times for performing certain activities, such as exercise, eating, taking medication, etc., based on the predictions. Any of the methods herein (e.g., an app on a personal device) can optionally be adapted to visually present (e.g., plot) the amount of time (e.g., percent) spent in an optimal / preferred glucose range versus the time spent outside the optimal / preferred range, which range can be adjustable and / or personalized.Optionally, the method (e.g., an app) is adapted to allow personalized adjustment and setting of ranges via interaction with a display of the personal device. One aspect of the present disclosure is a computer-executable method adapted to present interactive features (e.g., on-screen icons, up / down arrows, audible instructions, etc.) that allow a user (e.g., a patient and / or a care team member) to adjust the range of glucose values.

[0057]

[0072] Any of the methods herein may optionally be adapted to provide or initiate an alert when a current glycemic event or condition is detected and / or when a future glycemic event is predicted. For example, when a future glycemic event is predicted, any of the alerts herein may be triggered (such as an audible alert, a text alert, an email, an alert to the patient and / or caregiver / care team, etc.). Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to provide or initiate an alert when glucose indicates a trend toward going out of range and / or is currently out of range. Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to provide or initiate a relatively high-frequency pitch (47-65 Hz) alert when glucose indicates a trend toward going out of range, which can alert a service animal of the predicted or detected glycemic event. One aspect of the present disclosure is an executable method adapted to provide or initiate an alert with a relatively high frequency pitch (e.g., 47-65 Hz) when a medical event is detected and / or predicted, such as a glucose value trending out of range, a seizure being detected or predicted, or a loss of consciousness being predicted.

[0058]

[0073] One or more aspects of the actual and / or predicted blood glucose value (or information indicative thereof or related thereto) may optionally be communicated to the subject and / or a third party (caregiver, family member, etc.). One or more aspects of the actual and / or predicted blood glucose value or status may optionally be communicated to one or more devices that may be the same as or different from the device that determines the actual and / or predicted blood glucose information.

[0059]

[0074] The methods and systems herein may optionally be adapted to continuously stream real-time EEG data to another device, such as any of the personal devices herein (smartphone, watch, etc.). Thus, an app herein can receive the continuous or near-continuous real-time EEG data being sensed from a patient and make predictions using this continuously streamed real-time EEG data.

[0060]

[0075] The methods and systems herein may optionally be adapted to receive one or more non-EEG patient parameters, such as skin conductance, heart rate, blood pressure, etc. Any of these parameters may be input to the prediction and / or detection methods herein. Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to receive one or more inputs from one or more of the patient or caregiver. For example, any of the methods herein (e.g., an app on a personal device) may optionally be adapted to allow user input of information such as, for example, insulin administration / dosage, medication timing, food, exercise, stress, sleep, illness, etc. Any of the methods herein (e.g., an app on a personal device) may optionally be adapted to allow a user (patient and / or caregiver) to mark or indicate events on the raw EEG trace (or processed EEG data).

[0061]

[0076] Portions of the present disclosure include methods adapted to use one or more inputs to improve or train (e.g., increase the accuracy of) any of the predictive algorithms herein, and optionally, the inputs can alternatively or additionally improve or train any of the CGMs herein. For example, without limitation, inputs include, without limitation, inputs manually entered via a user and / or inputs received from other devices (such as a pump, glucose meter, and / or CGM). For example, an app herein may be adapted to receive information (e.g., blood glucose measurements / values) from a glucose meter and / or CGM and use the information to improve the predictive algorithm. A user may optionally input a blood glucose value. Inputs in this context may further include, without limitation, food ingested and / or the time at which food was ingested, the time and / or dosage of insulin administered, or any other input related to the user's life that may improve the predictive method.

[0062]

[0077] There may (but is not necessarily) be some patient variability in predictions. By way of example only, there may be some variability in the time period for which accurate predictions are possible. By way of example only, some patients may be most accurately predicted 8 hours in advance, while others may be most accurately predicted 2 hours in advance. Thus, the systems and methods herein may be somewhat personalized for individual patients.

[0063]

[0078] Additionally, there may be some variability in the time at which a patient can be provided with a peak or optimal prediction. For example, a first patient may have an optimal prediction 2-3 hours before a glycemic event (or up to 3 hours before the event), while a second patient may have an optimal prediction 4-5 hours before the event (or up to 5 hours before the event). Thus, the systems and methods herein may be somewhat personalized for individual patients.

[0064]

[0079] One aspect of the present disclosure relates to managing blood glucose levels. In some examples, managing blood glucose levels includes preventing glycemic events, such as hypoglycemic events and hyperglycemic events. In some examples, managing blood glucose levels also optionally includes halting or minimizing the severity of a current glycemic event or condition. Management methods herein may include guiding medication and insulin (long / short acting) dosage based on glucose predictions.

[0065]

[0080] It will be appreciated that some of the above concepts, such as alerts and outputs (optionally provided by an app), may also be considered part of an overall approach to managing blood glucose levels. For example, communicating insulin and medication needs to a user based on real-time and predicted blood glucose may be considered part of management. Additionally, by way of example only, automated dietary and exercise recommendations to a subject (optionally communicated via an app) may also be considered part of management.

[0066]

[0081] It is understood that aspects of the present disclosure related to predicting / detecting blood glucose levels may or may not be incorporated into aspects of the present disclosure related to managing blood glucose levels.

[0067]

[0082] Merely exemplary management methods may include closed-loop functionality and, optionally, a pump that may be adapted to administer one or more medications (e.g., glucagon, insulin, etc.). For example, one aspect of the present disclosure is a method for controlling a patient's future glucose levels in response to measuring / detecting EEG signals and / or patterns (although other inputs, such as HR, skin conductance, etc., may also be used as part of the control process). In this exemplary aspect, controlling may include administering insulin or glucagon to the patient before the glucose levels deviate from a desired range / limit. In this regard, controlling may include administering insulin to the patient when the glucose levels are still within a safe range (e.g., 70-180, 80-170, etc.). In this regard, controlling may include maintaining the glucose levels within a safe / desirable range. In this regard, controlling may include administering a medication (e.g., insulin) before the time when the glucose levels are predicted to deviate from the safe range. In this regard, controlling may include administering a particular dose of insulin based on the time when the glucose levels are predicted to deviate from the safe range. In this regard, control may include administering a different dose of insulin than that administered in response to a real-time glucose level monitoring process, such as via a CGM.

[0068]

[0083] The systems herein may optionally include a multi-chamber pump, such as a dual-chamber pump adapted to administer insulin and glucagon to manage future glucose levels.

[0069]

[0084] The systems herein can be adapted and configured to integrate with Bluetooth-enabled insulin pumps (eg, Omnipod).

[0085] Patients in an ICU environment have their blood glucose levels measured periodically, typically via finger prick testing with a glucose meter. This process can take 10 minutes to complete, resulting in a delay of approximately 10 minutes for each result. Being able to predict blood glucose levels can be extremely useful in an ICU environment, both in terms of patient care and hospital resource management. ICU patients typically already have EEG electrodes attached, so their EEG can be sensed and analyzed to predict (or help predict) future blood glucose levels. Therefore, any of the methods and systems described herein can be advantageously used in an ICU environment. Currently, insulin infusions are typically provided to ICU patients, using a sliding scale of doses as part of their current treatment.

[0070]

[0086] Any of the systems and methods herein may optionally be adapted to administer insulin when a hyperglycemic event is predicted. Any of the systems and methods herein may optionally be adapted to administer glucagon (e.g., intranasally, using a pump, etc.) when a hypoglycemic event is predicted. As an example of managing hypoglycemia herein, any of the methods herein may include, for example, an alert or recommendation to consume a high-sugar drink.

[0071]

[0087] Any of the systems and methods herein may include one or more of oral drug administration, subcutaneous administration, or administration via a pump.

[0088] As described, the present disclosure includes an exemplary system, but it is understood that systems having different configurations may be adapted to implement many, if not all, of the methods herein. FIG. 6 provides an illustrative, non-limiting diagram of a glucose prediction system, which may optionally not include all of the components shown. The "mini-EEG" in FIG. 6 is merely representative of an exemplary wearable sensor. The "glucose monitor" in FIG. 6 is merely representative of an exemplary CGM. The smartphone shown is merely representative of a personal device on which an app may be stored. An optional pump is also shown. Wireless capabilities (e.g., Bluetooth) are also shown.

[0072]

[0089] In an illustrative example only, the system herein may be a continuous glucose forecasting system in which one or more patient parameters (such as EEG) are continuously monitored. Figure 7 shows an exemplary continuous glucose forecasting system (CGFS). In alternative systems, monitoring may be periodic or a combination of continuous and periodic.

[0073]

[0090] Optionally, the systems herein include one or more wearable, rechargeable sensors (sometimes referred to as wearable sensing devices or sensing devices), such as multiple (e.g., two) electrodes, optionally ear-hanging. The sensing devices herein may optionally have Bluetooth or other wireless communication capabilities. The sensing devices herein may be adapted to measure EEG signals continuously, near-continuously, and / or periodically. U.S. Pat. No. 1,102,035 is merely an example of a wearable EEG monitoring / recording system, any features of which may be incorporated into any of the wearable sensing devices, systems, and / or methods of use herein.

[0074]

[0091] The wearable sensing devices herein may or may not have storage capabilities. The wearable sensing devices herein may or may not have signal processing / analysis capabilities.

[0075]

[0092] Any of the wearable sensing devices herein may be worn on the scalp of a subject, or in other instances may be subcutaneous (subgaleal) in configuration.

[0093] Any of the wearable sensing devices herein may include a microneedle array that may be adapted to sense ISF (similar to CGM). The needle array may be positioned such that the needles extend into the skin.

[0076]

[0094] By way of example only, wearable sensors herein may include one or more of the following: - Completely non-invasive - Multi-electrode (two-channel, single-channel, etc.) patches, silicone (patches) and stainless steel ("dry" electrodes) adapted for continuous ear-hook wear over several days (e.g., 30 days or more), with replaceable stickers to hold the patch in place on the skin - May include a memory chip for at least temporarily storing data (e.g., if the personal device is not nearby, the sensing device may need to be able to store data until it can be transmitted to the personal device). - Bluetooth communication function - waterproof - Rechargeable - Optional microneedles / microarrays integrated into the patch for ISG measurements

[0095] Optionally, the systems herein include a personal device (such as a smartphone, watch, or tablet) on which an app is optionally stored. The app can be accessed by the subject or a caregiver. The app is optionally adapted to receive / collect data from one or more wearable sensors, process the received information (to some extent), and optionally display actual and predicted blood glucose measurements or risk indicators to the user. This is described in more detail above. The app is also optionally adapted to display trends of past blood glucose measurements and user insights regarding blood glucose values. The app is optionally adapted to store EEG data (raw and / or processed), optionally until transferred to a different device, such as an optional online portal described below. Personal devices may have larger data storage and thus may be suitable for storing more data, allowing wearable sensors to have a smaller form factor.

[0077]

[0096] Any of the apps on a personal device herein may be adapted to perform any of the methods herein or may cause a processor to perform the app's computer-executable instructions (e.g., executable methods such as prediction).

[0078]

[0097] Optionally, the systems herein include an online portal, optionally available to the subject and any other individual authorized by the user (such as a physician, caregiver, care team, family member, etc.). The portal may be adapted to display any of the information or data described herein, including historical data regarding blood glucose trends and user insights regarding blood glucose values. The portal may optionally be adapted to display raw EEG signals. The online portal may have signal processing and analysis capabilities.

[0079]

[0098] By way of example only, the online portal may include any of the following features or capabilities: - Communicate actual and predicted future glucose values ​​to the patient / care team - Communicating trends of past glucose readings - Communicating insulin and medication information to users based on real-time and predicted blood glucose -Providing alerts when glucose is trending out of range or goes out of range - Providing users with long-term (e.g., 10+ hour) predictions of "at risk" and suggesting optimal times for exercise, meals, and medications based on glucose predictions - Adapted to display app information from external insulin pumps, glucose monitors and Bluetooth-linked systems - be adapted to display raw EEG traces using user-input markers, and be capable of displaying any other relevant health data;

[0099] One aspect of the present disclosure relates to training a predictive method or model (e.g., a computer-executable predictive method or model) for predicting a subject's future glucose state (e.g., future blood glucose values ​​or risk indicators for future blood glucose values). The training may include sensing one or more patient parameters (e.g., EEG signals) from one or more subjects, optionally noninvasively and optionally from a scalp location. The training may optionally include processing the sensed EEG signals. The training may include sensing real-time or near-real-time blood glucose values ​​(which may be interstitial blood glucose values ​​representative of blood glucose values) or information indicative of the real-time or near-real-time blood glucose values ​​from one or more subjects, and then creating an association between the one or more sensed parameters (e.g., EEG signals) and the sensed real-time or near-real-time blood glucose values ​​or information indicative of the real-time or near-real-time blood glucose values. In this manner, the predictive method can be trained to predict (at least with some accuracy) in advance when a future glycemic event will occur or is likely to occur.

[0080]

[0100] Any of the methods / algorithms herein may be trained for one or more of normal conditions, hyperglycemic conditions (diabetes, hyperglycemic ICU, sepsis, traumatic brain energy, diabetic ketoacidosis, etc.), or hypoglycemic conditions.

[0081]

[0101] The systems, devices, and methods herein may be adapted to be combined with glucose monitors, such as CGMs, and / or their methods of use. For example, existing CGMs may be modified and adapted to incorporate sensed EEG data (by way of example only, any of the sensing concepts / methods of U.S. Pat. Nos. 6,572,542 and / or 8,118,741) and / or predictive concepts herein to improve performance. By way of example only, CGM sensed data may be analyzed along with patient EEG data, and predictive EEG data may be trained with interstitial glucose (ISG) data, so that the CGM can be adapted to better predict future blood glucose conditions using ISG measurements. Exemplary method steps are shown in FIG. 8 and may be combined with any other suitable method steps herein. For example, certain patterns of EEG-trained ISG data (readings well before an impending event) may then be used to predict future glycemic events. Thus, it is understood that any existing CGM may be modified and adapted to incorporate any of the features or methods herein. For example, the CGM can be adapted to communicate with an app to remind the subject that they should prepare to drink a sugary drink within a certain time period, such as within the next hour, or that a hypoglycemic event is likely to occur in 2.5 hours. Further, the CGM can be modified to deliver insulin much earlier than previous techniques, for example, to deliver long-acting insulin well in advance of a hyperglycemic event. Figure 9 shows exemplary method steps in which sensed ISG can be used to manage a subject's future glucose status.

[0082]

[0102] 10 illustrates exemplary steps that may be included in a method for training ISG data using EEG data, which may be implemented by a glucose monitor to sense ISG to 1) predict information indicative of future glucose conditions (e.g., FIG. 8) and / or 2) facilitate management of future glucose conditions (e.g., FIG. 8).

[0083]

[0103] Additionally, by way of example only, any of the EEG data and methods herein may optionally be used to assist in the calibration and / or recalibration of a glucose monitor (e.g., a CGM) (which may require recalibration over time), thereby eliminating the need to use a glucose meter or finger prick to recalibrate a glucose monitor such as a CGM. FIG. 11 illustrates an exemplary method for calibrating or recalibrating a glucose monitor (optionally a CGM). The method includes calibrating or recalibrating the glucose monitor using at least one of EEG data sensed from a subject or information indicative of EEG data sensed from the subject. Additionally, a glucose monitor (e.g., a CGM) and / or glucose meter may similarly be used to calibrate any of the EEG prediction methods (e.g., algorithms) herein. An example is illustrated in FIG. 12.

[0084]

[0104] One aspect of the present disclosure is an optional bidirectional calibration method and / or system, an example of which is shown in Figure 13. In one example, the bidirectional calibration method may include sensing interstitial glucose of a subject with a glucose monitor (optionally a CGM), sensing EEG signals from one or more subjects, optionally with any of the wearable devices described herein, and at least one of, and optionally both, using the sensed EEG signals and / or information indicative of the sensed EEG signals to calibrate (or recalibrate) the glucose monitor or using the sensed ISG or information indicative of the sensed ISG to calibrate a method adapted to determine current blood glucose values ​​or predict future blood glucose values ​​from the sensed EEG signals and / or information indicative of the sensed EEG signals.

[0085]

[0105] Furthermore, as mentioned above, CGM thresholds (when resetting is necessary) are currently reset during outpatient visits. Sensing EEG data, with its predictive nature, may even allow thresholds to be reset without the need for an outpatient visit, such as by using the online portal herein.

[0086]

[0106] Additionally, the CGM can optionally predict future glucose status in conjunction with any of the other health data / parameters (e.g., user input) herein (e.g., heart rate, blood pressure, skin conductance - easily sensed by existing devices such as smartwatches and Fitbits).

[0087]

[0107] Exemplary CGMs, features, and methods of use thereof that may be incorporated herein include those by Dexcom (such as the G6 CGM System), Medtronic (such as the Guardiam™ Connect), Abbott (such as any FreeStyle Libre), and Eversense® E3 CGM.

[0088]

[0108] Glucose meters (glucometers), features, and methods of use, such as those exemplified by Abbott's LifeScan OneTouch®, Accu-Chek®, and FreeStyle Lite, may be incorporated herein.

[0089]

[0109] Insulin pumps, features, and methods of use, such as those exemplified by pumps from Medtronic Minimed™, Tandem, and Omnipod® pumps, may be incorporated herein.

[0090]

[0110] Any feature from any of the examples or embodiments herein can be combined with any other feature unless expressly stated otherwise herein. For example, any of the methods herein may or may not be implemented by a system or device.

[0091]

[0111] 14 shows an exemplary method including the steps as shown. Any of the methods herein may include receiving as input non-invasively sensed EEG data or processed EEG data from a subject, predicting a future glucose state of the subject based at least in part on the EEG data or processed EEG data, and initiating an output adapted to communicate information indicative of the subject's predicted future glucose state based on and in response to the subject's predicted future glucose state, as shown in FIG.

[0092]

[0112] 15 shows an exemplary training method including the steps as shown. Any of the training methods herein may include sensing or receiving training real-time or near real-time blood glucose values ​​or information indicative of the real-time or near real-time blood glucose values ​​from one or more subjects, sensing training EEG signals from the one or more subjects, processing the sensed training EEG signals, and correlating the sensed or processed training EEG signals with the sensed or received training real-time or near real-time blood glucose values ​​or information indicative of the real-time or near real-time blood glucose values ​​to train a method for predicting an individual's future glucose state based on subsequent and non-invasively sensed EEG signals from the individual, as shown in FIG.

[0093]

[0113] 16 shows an exemplary method including the steps as shown. Any of the methods herein may include non-invasively sensing EEG signals from a subject's scalp, inputting the EEG signals or processed EEG signals to a trained computer-executable method trained to predict a future glucose state of the subject, predicting a future glucose state of the subject based at least in part on the non-invasively sensed EEG signals, and outputting instructions to initiate communication based on and in response to the predicted future glucose state, as shown.

[0094]

[0114] 17 illustrates exemplary components of a personal device. This personal device may be any personal device described herein and may include additional components (such as, but not limited to, any of the exemplary hardware components of FIG. 19). A personal device described herein may include one or more processors (illustrated as a single "processor") and one or more non-transitory memories or media. The media may store any of the computer-executable methods (e.g., apps) described herein.

[0095]

[0115] 18 shows an exemplary system including at least one wearable EEG sensor (such as any of the behind-the-ear and / or single-channel EEG sensors described herein) and a personal device. The description of the personal device in FIG. 17 is fully incorporated herein by reference into the description of FIG. 18. The system in FIG. 18 may be any of the systems described herein and may include other components, such as any of the components described herein.

[0096]

[0116] 19 illustrates merely examples of one or more components that may be included in any of the sensors and / or computing devices (such as the personal devices) herein. Reference labels are understood to refer to textual descriptions of the components illustrated in FIG. 19.

[0097]

[0117] Unless specifically indicated, one or more of the methods or techniques described in this disclosure (e.g., any computer-executable method adapted to be executed by a processor, which may be stored on a non-transitory medium) may be implemented, at least in part, in hardware, software, firmware, or any combination thereof. For example, various aspects of the techniques or components may be implemented within one or more processors including one or more microprocessors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), programmable logic circuits, etc., either alone or in any suitable combination. The terms "processor" or "processing circuitry" may generally refer to any of the above circuits alone or in combination with other circuits or any other equivalent circuitry.

[0098]

[0118] Such hardware, software, or firmware may be implemented within a single device or separate devices to support the various operations and functions described in this disclosure. Furthermore, any of the described units, modules, or components may be implemented together or separately as discrete, yet interoperable, logic devices. While various features are depicted as modules or units, this is intended to emphasize different functional aspects and does not necessarily imply that such modules or units must be realized by separate hardware or software components. Rather, functionality associated with one or more modules or units may be performed by separate hardware or software components or integrated within common or separate hardware or software components.

[0099]

[0119] When implemented in software, the functions pertaining to the systems, devices, and techniques described in this disclosure may be embodied as instructions on a computer-readable medium, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically erasable programmable ROM (EEPROM), flash memory, etc. These instructions may be executed by a processor to support one or more aspects of the functions described in this disclosure.

Claims

1. 1. A method for predicting a future glucose status of a subject, comprising: non-invasively sensing EEG signals from ear hook locations on the subject's scalp; inputting the EEG signals or processed EEG signals into a trained computer-executable method that is trained to predict a future glucose status of the subject; predicting a future glucose status of the subject based at least in part on the non-invasively sensed EEG signal; outputting instructions to initiate a transmission based on and in response to the predicted future glucose state; A method comprising:

2. 10. The method of claim 1, wherein the step of non-invasively sensing EEG signals comprises sensing EEG signals with a device including a plurality of sensing electrodes, the plurality of sensing electrodes including only a first sensing electrode and a second sensing electrode.

3. 10. The method of claim 1, further comprising wirelessly transmitting sensed EEG signals to a personal device, wherein the computer-executable method is stored on the personal device and the predicting step occurs on the personal device.

4. 10. The method of claim 1, wherein the step of outputting an instruction to initiate the transmission comprises outputting an instruction to initiate the transmission in the form of information displayed on a display.

5. The method of claim 4 , wherein the display is a display on a personal device.

6. The method of claim 4 , wherein the communicating comprises causing a personal device to visually present information indicative of the predicted future glucose state on a display.

7. The method of claim 1 , wherein predicting the subject's future glucose status comprises predicting one or more future interstitial glucose values.

8. 10. The method of claim 1, wherein the step of outputting instructions to initiate a transmission comprises outputting instructions to initiate a transmission that causes administration of insulin to the subject.

9. A computer-executable method stored on a non-transitory medium, which, when executed by a processor, receiving as input non-invasively sensed or processed EEG data from ear hook locations on a subject's scalp; predicting a future glucose status of the subject based at least in part on the EEG data or processed EEG data; initiating an output adapted to communicate information indicative of the predicted future glucose state of the subject based on and responsive to the predicted future glucose state of the subject; A computer-executable method adapted to perform the steps of:

10. 10. The method of claim 9, wherein the receiving step includes receiving as input non-invasively sensed or processed EEG data from a wearable device at the ear hook location, the wearable device including a plurality of electrodes having only a first sensing electrode and a second sensing electrode.

11. 10. The method of claim 9, further comprising wirelessly transmitting the sensed EEG signals to a personal device, wherein the computer-executable method is stored on the personal device and the predicting step occurs on the personal device.

12. The method of claim 9 , wherein initiating the output adapted to convey information comprises initiating an output that causes a personal device to convey the information.

13. The method of claim 12 , wherein the information is visually represented on a display of the personal device.

14. 10. The method of claim 9, wherein predicting the subject's future glucose status comprises predicting future interstitial glucose values.

15. 10. The method of claim 9, wherein the step of initiating an output adapted to communicate information comprises initiating an output adapted to communicate information that causes administration of insulin to the subject.

16. A non-transitory computer-readable storage medium having instructions stored thereon, the instructions being executable by a processor to perform a method, the method comprising: receiving as input non-invasively sensed or processed EEG data from ear hook locations on a subject's scalp; predicting a future glucose status of the subject based at least in part on the EEG data or processed EEG data; initiating an output adapted to communicate information indicative of the predicted future glucose state of the subject based on and responsive to the predicted future glucose state of the subject; 1. A non-transitory computer-readable storage medium comprising:

17. The medium of claim 16 , further comprising a personal device, the non-transitory computer-readable storage medium being stored on the personal device, the device further comprising the processor.

18. 17. The medium of claim 16, wherein the personal device comprises a display, and initiating an output adapted to communicate information comprises initiating an output that causes the personal device to communicate the information.

19. The method of claim 18 , wherein the information is visually represented on a display of the personal device.

20. 17. The medium of claim 16, wherein predicting the subject's future glucose status comprises predicting future interstitial glucose values.

21. 17. The medium of claim 16, wherein the step of initiating an output adapted to communicate information comprises initiating an output adapted to communicate information that causes administration of insulin to the subject.

22. 1. A method for predicting a future glucose status of a subject, comprising: non-invasively sensing EEG signals using a single channel EEG sensor on the subject's scalp; inputting the EEG signals or processed EEG signals into a trained computer-executable method that is trained to predict a future glucose status of the subject; predicting a future glucose status of the subject based at least in part on the non-invasively sensed EEG signal; outputting instructions to initiate a transmission based on and in response to the predicted future glucose state; A method comprising:

23. 23. The method of claim 22, further comprising any one or more of the method steps of any one or more of claims 2 to 8.

24. A computer-executable method that, when executed by a processor, receiving as input EEG data sensed non-invasively using a single channel EEG sensor on the scalp of a subject or processed EEG data; predicting a future glucose status of the subject based at least in part on the EEG data or processed EEG data; initiating an output adapted to communicate information indicative of the predicted future glucose state of the subject based on and responsive to the predicted future glucose state of the subject; A computer-executable method adapted to perform the steps of:

25. 25. The method of claim 24, further comprising any one or more of the method steps of any one or more of claims 10 to 15.

26. 1. A system for predicting a future glucose status of a subject, comprising: a wearable EEG sensor configured and arranged to be wearable on the scalp; A computer-executable method stored on a non-transitory medium of a personal device, which, when executed by a processor within said personal device, comprises: receiving as input non-invasively sensed or processed EEG data from the EEG sensor; predicting a future glucose status of the subject based at least in part on the EEG data or processed EEG data; initiating an output adapted to communicate information indicative of the predicted future glucose state of the subject based on and responsive to the predicted future glucose state of the subject; a computer-executable method adapted to perform the A system comprising:

27. 27. The system of claim 26, wherein the wearable EEG sensor includes a plurality of sensing electrodes, the plurality of electrodes including only a first sensing electrode and a second sensing electrode.

28. 27. The system of claim 26, wherein the wearable EEG sensor is a single channel EEG sensor.

29. 27. The system of claim 26, wherein the personal device includes a display, and the step of initiating an output adapted to convey information includes initiating an output adapted to convey information displayed on a display of the personal device.

30. 30. The system of claim 29, wherein the information is visually represented on the display of the personal device.

31. 31. The system of claim 30, wherein the information visually represented on the display indicates the predicted future glucose state.

32. 1. A method for predicting future glucose levels in a subject, comprising: sensing EEG signals from a subject using a behind-the-ear EEG device; processing the sensed EEG signals; using an application on a personal device to analyze the processed EEG signals using a trained predictive model to predict a future glucose status of the subject; causing the personal device to visually present on a display information indicative of the predicted future glucose state; A method comprising:

33. 33. The method of claim 32, wherein the information indicative of the predicted future glucose state includes a future time period and a predicted glucose state during the future time period.

34. 34. The method of claim 33, wherein the information indicative of the predicted future glucose state comprises a graph showing time on a first axis and predicted glucose state on a second axis.

35. 33. The method of claim 32, further comprising the step of communicating the sensed EEG signals to the personal device.

36. 1. A method for predicting future glucose levels in a subject, comprising: sensing EEG signals from a subject using a behind-the-ear EEG device; processing the sensed EEG signals; using an application on a personal device to analyze the processed EEG signals using a trained predictive model to predict a future glucose status of the subject; A method comprising:

37. 37. The method of claim 36, further comprising the step of communicating the sensed EEG signals to the personal device.

38. 1. A computer-executable method stored in a non-transitory memory of a personal device, comprising: receiving sensed EEG data from the subject or information indicative of the sensed EEG data from the subject; analyzing the processed EEG signals using a trained predictive model to predict a future glucose status of the subject; causing the personal device to visually present on a display information indicative of the predicted future glucose state; A method comprising:

39. 39. The computer-executable method of claim 38, wherein the causing step comprises causing the personal device to visually present on the display a future time period and the predicted glucose state for the future time period.

40. 40. The computer-executable method of claim 39, wherein the presenting step includes causing the personal device to visually present on the display a graph showing time on a first axis and the predicted glucose state on a second axis.

41. A glucose prediction system (GFS) comprising: a minimally invasive EEG device including a first sensor and a second sensor, the minimally invasive EEG device being sized, configured, and adapted to be ear-mounted on a subject and to sense EEG signals with the first sensor and the second sensor; a personal device adapted to communicate with said EEG device, receiving and processing the sensed EEG signal or information indicative of the sensed EEG signal; analyzing the processed EEG signals using a trained predictive model to predict future glucose levels for the subject; a personal device further adapted to A glucose prediction system (GFS) comprising:

42. 42. The glucose prediction system of claim 41, wherein the personal device is further adapted to visually present information indicative of the predicted future glucose state on a display of the personal device.

Images