EEG recording and analysis

Abstract

One embodiment provides a method that includes acquiring EEG data from one or more single-channel EEG sensors worn by a user, classifying the EEG data using a processor as one of nominal and abnormal, and providing an indication associated with the classification of the EEG data. Other embodiments are described and claimed.

Description

[Technical Field]

[0001] This application relates to the recording and analysis of EEG. [Background technology]

[0002] There are thousands of hospitals in the United States. Many of these hospitals are regional or regional hospitals. These regional or regional hospitals are traditionally part of a hospital system or network. One example of one such network includes several regional hospitals with one major tertiary hospital. Regional or regional hospitals outside of the larger hospital network typically contract with the larger tertiary hospital for emergency and intensive care solutions outside the regional or regional hospital's area of ​​expertise.

[0003] Electroencephalogram (EEG) monitoring has traditionally only been available in large tertiary hospitals that support neurology departments with EEG services. Many hospitals do not offer EEG monitoring. These hospitals make arrangements with larger tertiary hospitals or their partners when such monitoring is necessary or desired for a patient. This traditionally takes the form of referring the patient to a tertiary hospital for specialist services. In many cases, this involves transferring or transporting the patient to the tertiary hospital for services. Summary of the Invention [Means for solving the problem]

[0004] In summary, one embodiment provides a method that includes acquiring EEG data from one or more single-channel EEG sensors worn by a user, classifying, using a processor, the EEG data as one of nominal and abnormal, and providing an indication associated with the classification of the EEG data.

[0005] Another embodiment provides a system including an output device, a processor operably coupled to the output device, and a memory storing instructions executable by the processor to acquire EEG data from one or more single-channel EEG sensors worn by a user, classify the EEG data as one of nominal and abnormal, and provide instructions associated with the classification of the EEG data.

[0006] A further embodiment provides a method that includes acquiring EEG data from two or more single-channel EEG sensors worn by a user, transmitting the EEG data to a remote device, and providing data from the remote device to a display associated with the remote user, the data including a montage of the EEG data.

[0007] Those skilled in the art will appreciate that the foregoing is a summary and, as such, may contain simplifications, generalizations, and omissions of detail; thus, the summary is illustrative only and is not intended to be in any way limiting.

[0008] For a better understanding of the embodiments, together with other and further features and advantages, reference is made to the following description taken in conjunction with the accompanying drawings. [Brief explanation of the drawings]

[0009] [Figure 1A] FIG. 1 illustrates an exemplary sensor. [Figure 1B] FIG. 1 illustrates an exemplary sensor. [Figure 2] FIG. 1 illustrates an exemplary system. [Figure 3] FIG. 1 illustrates an example of EEG monitoring and instruction. [Figure 4A] FIG. 1 illustrates exemplary EEG data. [Figure 4B] FIG. 1 illustrates exemplary EEG data. [Figure 5]FIG. 10 is a diagram illustrating an example of classification of EEG data. [Figure 6] FIG. 1 illustrates an exemplary method of EEG monitoring and instruction. [Figure 7A] FIG. 1 illustrates an example of an application view or screen. [Figure 7B] FIG. 1 illustrates an example of an application view or screen. [Figure 8] FIG. 1 illustrates an exemplary system. DETAILED DESCRIPTION OF THE INVENTION

[0010] It will be readily understood that the components of the embodiments, as generally described and illustrated in the Figures herein, can be arranged and designed in a wide variety of different configurations in addition to the illustrated exemplary embodiments. Thus, the following more detailed description of the exemplary embodiments represented in the Figures is not intended to limit scope of the claims, but is merely representative of such embodiments.

[0011] To optimize treatment after initial seizure diagnosis, epileptologists ideally obtain high-quality, longitudinal EEG studies in hospitals using 19+ channel wired EEG placement in the standard international "10-20" system. Such studies can be difficult to perform because the process is prohibitively expensive, time-consuming, and extremely inconvenient for patients. Additionally, the time spent in the epilepsy monitoring unit may not capture any or all of the seizure activity a person typically experiences over an extended period of time. However, optimal treatment often depends on identifying the full extent of a patient's convulsive (clinical) and non-convulsive (subclinical) seizure activity. Currently, the only technique for collecting a history of seizure activity outside of the hospital is the seizure diary, a home-based self-report record that can be incomplete. Seizure diaries can be difficult to maintain accurately, and diaries can be inaccurate, making clinical decisions regarding appropriate medication difficult. Furthermore, traditional EEG techniques do not adequately address the need for monitoring in many emergency medical scenarios, particularly those encountered by paramedics or clinicians in small or rural hospitals.

[0012] Referring now to the drawings, representative exemplary embodiments will be described to provide a better understanding of the appended claims.

[0013] Referring to FIG. 1A , one embodiment provides a sensor 101a. While the sensor 101a can take various forms, the example in FIG. 1A shows a wearable sensor that adheres to a patient's skin. The sensor 101a may include a data analysis platform that takes EEG monitoring out of the hospital. The sensor 101a features a small size and can be placed in various locations, reducing user strain during prolonged wear. This enables EEG data collection using a small, convenient sensor 101a that provides real-time EEG data capture, as shown in FIG. 1A . This contrasts with the complex, wired electrode configurations typically used in hospital environments, which feature many electrodes and measurement locations.

[0014] In the example shown in FIG. 1A, a small, connected health wearable sensor 101a provides single-channel EEG data that can be analyzed for the occurrence of both convulsive (clinical) and non-convulsive (subclinical) seizures. The sensor 101a senses the user's EEG data by detecting brain electrical activity, for example, by sensing the voltage difference between two electrode contacts. The sensor may include two electrodes separated by a sensor housing, forming a single bipolar channel ("single channel") in a typical EEG montage. In one embodiment, the sensor 101a is a single-channel differential amplifier transmitter and data logger with 6 mm diameter gold electrodes and 18 mm electrode spacing, similar to a bipolar pair in high-density EEG. The electrode contacts may be located on the surface of a device that is placed on the patient's skin, for example, adhered to the patient via a sticker or other adhesive containing a material such as hydrogel that enables voltage sensing. The sensor 101a is placed in an appropriate location on the user, for example, on the scalp below the hairline, to sense and record single-channel EEG data. The EEG data may be analyzed on-board, for example, through the application of analytical or machine learning models stored on the sensor 101a, or may be analyzed by a local device, a remote device, or a combination of the above. By way of example, the sensor 101a may communicate to a local device using a personal area network (PAN), such as communicating data via a wireless network, e.g., secure Bluetooth Low Energy (BLE), to a smartphone or tablet. Similarly, the sensor 101a may communicate with a remote device using a wide area network (WAN), such as communicating EEG data via the internet to a remote server or cloud device, whether or not communicating through an intermediary device such as a local device.

[0015] In one embodiment, the EEG monitoring sensor 101a is a self-contained recording patch that includes a first electrode and a second electrode that cooperate to measure a voltage. The sensor 101a includes circuitry for generating an EEG signal from the measured voltage, amplifying the EEG signal, digitizing the EEG signal, and retrievably storing the EEG data in memory. The sensor 101a may also include a power source and an enclosure that houses the circuitry, power source, and first and second electrodes in an integrated package. The sensor 101a is worn on the user's scalp, e.g., on the forehead or both crowns, and can capture EEG data over an extended period of time while the user performs normal daily activities.

[0016] As shown in FIG. 1B, in one embodiment, sensor 101b is designed to be unobtrusive and water-resistant, allowing for continuous use in all aspects of a person's normal daily life. In the example of FIG. 2, sensor 101b may be placed below the hairline in a location such as behind the ear. Sensor 101b placement may be assisted by prior diagnosis, for example, following a formal assessment using a full sensor montage (via monitoring with a typical wired sensor array in a cap) or via refined position estimation facilitated by adjusting the position of sensor 101b or a set of sensors over time. By selecting appropriate placement of sensor 101b, the usefulness of single-channel EEG data is improved by placing sensor 101b in close proximity to foci of brain activity or seizure activity in the brain.

[0017] As a non-limiting example, a patient may first receive a diagnosis of seizure activity, e.g., from a clinician, at a prescribed location. Thereafter, the patient may be asked to monitor seizure activity using sensor 101b. In one embodiment, a data source indicating the location of the diagnosed seizure activity is accessed, such as an electronic medical record (EMR) or in an application that receives user input to a location map image or illustration. The location data may be provided to the patient, e.g., via a companion mobile application that facilitates pairing and data communication between sensor 101b and a local device. In another example, the location information is provided to an online application that can be viewed by a user (e.g., a patient or clinician). Thus, the location data may be provided to the user in the form of instructions indicating acceptable placement of a sensor, e.g., sensor 101b, to maximize the likelihood of detecting an upcoming seizure using a single-channel sensor, e.g., sensor 101b.

[0018] Due to the small size of sensor 101b, e.g., on the order of 1 inch in length and width and ½ inch in depth dimensions, sensor 101b may be worn continuously for several days before needing to be removed, e.g., to charge an on-board power source such as a rechargeable battery. This allows for the capture, recording, and analysis of large amounts of single-channel EEG data using a detection process that identifies seizures, including seizures that may not be consciously known to the user wearing the sensor, such as seizures that occur during sleep.

[0019] The sensors 101a and 101b shown in Figures 1A and 1B are suitable for use in any patient, adult, adolescent, child, or neonate. A single sensor 101a and 101b may be used to facilitate more accurate recording of EEG data for real-time or later analysis. The sensors are also waterproof or water-resistant, making them suitable for wear during activities where the sensors 101a and 101b may get wet. This further facilitates long-term wear and comprehensive EEG data collection for seizure diary, seizure prediction, and seizure warning.

[0020] A consideration in making the sensors 101a, 101b viable long-term wearable sensors is power consumption. An exemplary goal is for the sensors 101a, 101b to operate for approximately three days without recharging. To enable continuous monitoring, a user may have two (or more) sensors 101a, 101b, using one while the other is recharged. Such a configuration allows for continuous EEG data capture and monitoring.

[0021] To facilitate long-term wear, several techniques may be used. For example, a power-consuming operation of the sensors 101a, 101b is to transmit EEG data from the sensors 101a, 101b to another device. To reduce the power used for data communication, the sensors 101a, 101b may transmit captured EEG data at intervals. For example, the sensors 101a, 101b may capture EEG data for a predetermined amount of time (e.g., 6 seconds) and then transmit the captured EEG data, e.g., one or more pages of EEG data. By transmitting data at intervals, the sensors 101a, 101b only need to activate their transmission capabilities for a short period of time (e.g., 1 second). As another example, a power-efficient microprocessor may be selected for use with the sensor. For example, some microprocessors may include a sleep processor core or capability while transmitting data via DMA to low-power SRAM for data communication. This feature may significantly reduce power consumption. Additionally, sensors 101a, 101b may have some components omitted, e.g., wireless radio, and may include other components, e.g., USB, or other data communication elements such as near-field or RFID, in various combinations, to facilitate power conservation and appropriate data transfer for a given use case. If a physical communication port is included, it may be covered to prevent ingress of water or contaminating elements, such as by placing sensors 101a, 101b under a removable hydrogel or sticker that adheres to the user's skin. The thickness of the sticker may be varied; for example, the thickness of the sticker or its material (e.g., hydrogel region) may be increased to accommodate use situations such as its placement on curved skin, while thinner stickers may be used on relatively flat surfaces.

[0022] The sensors 101a, 101b are suitable for use by adults, adolescents, children, and neonates. In the example of Figure 2, a system for monitoring in a clinical or emergency care environment is shown.

[0023] The example in Figure 2 uses an exemplary patient, but this non-limiting example can be extended to other clinical or non-clinical scenarios. Seizures are common in emergency care scenarios or asphyxiated newborns (especially within the first 2 days of life). Full-montage clinical EEG systems use many (11 or more) tethered (wired) electrode leads for monitoring. These leads must be placed by an EEG technician and can take up to 60 minutes to place. Reduced-set (3-lead) amplitude-integrated electroencephalogram (aEEG) recording systems provide real-time EEG from two channels along with a history of EEG activity displayed as a filtered, rectified, and averaged signal. However, electrodes are traditionally placed by a specialist. aEEG leads are placed in both parietal regions in standard 10-20 EEG systems. aEEG can be used to diagnose seizures as well as other background EEG abnormalities associated with encephalopathy. Persistently abnormal aEEG for as little as 48 hours is associated with adverse neurodevelopmental outcomes.

[0024] 2 illustrates one embodiment in which multiple single-channel sensors (collectively designated 201) are spaced apart on a patient's scalp, generally above the eyes in the parietal region and behind each ear, to create an EEG montage. Alternatives are possible, for example, two sensors may be placed approximately over the parietal region of a 10-20 EEG system to generate a hybrid of three channels: (1) C3-P3, (2) C4-P4, and (3) C3P3-C4P4. The output of sensors 201 can be synchronized, organized, and viewed in real time by an application (as described further herein) and converted by software, for example, into aEEG.

[0025] Similar to post-diagnosis instructions, instructions for placing the sensor 201 may be provided using a device 202, such as a desktop computer, tablet, other hospital monitor, or mobile device, running an application 203 that displays the instructions as a graphic showing placement information for the sensor 201 on the patient's scalp, e.g., on the forehead, behind the ears, a combination thereof, or other or additional locations. The placement information in this case may be generalized, e.g., approximate locations for seizure monitoring in a patient suspected of having suffered a given type of brain injury or traumatic event, or may be customized in some way if access to additional data is available, e.g., for a particular type of accident suspected for the patient, or in another clinical scenario, such as in the case of an adult or pediatric emergency care patient. As a specific example, one embodiment may provide graphical instructions such as those shown in FIG. 2, where an emergency medical screening is performed on a patient using four sensors: two on the forehead and two behind the ears. In this four-sensor configuration, one embodiment can create a desired montage (as described herein) by, for example, subtracting EEG signals from one sensor to another to create a 10-channel "vertical" montage, as further described in connection with FIG. 4B.

[0026] In the example shown in FIG. 2 , application 203 runs on device 202, e.g., a medical-grade tablet. While the application may be programmed to facilitate the collection of EEG data for emergency medical care, a version of the application may also be used for home use in a seizure diary context. In one example, application 203 provides instructions for EEG data collection by non-specialists, such as clinicians at a local or regional hospital who are unfamiliar with EEG monitoring. As an example, application 203 may provide a graphic, such as that shown in FIG. 2 , showing the placement and orientation of four single-channel sensors 201. In some examples, single-channel sensors 201 may be directional and keyed, such as by including markings. In the example of FIG. 2 , the graphic shows gold dots 207 that correspond to similar markings on single-channel sensors 201, allowing for correct orientation of the electrodes on the underside of the device (not shown) for placement on the patient's scalp. That is, the keyed device allows the user to properly align single-channel sensors 201 on the patient.

[0027] The command and wireless single-channel sensor allows for rapid collection and analysis (even remote analysis) by non-specialists in the field or clinical setting. This avoids or reduces the need to transport the patient to another location, such as a larger hospital with traditional EEG monitoring equipment and specialists. For example, upon a suspected seizure, emergency department staff at a local hospital can place four sensors 201 on the scalp and below the hairline, as instructed by the graphic shown in FIG. 2. The sensors begin transmitting EEG data to the same device displaying the graphic, i.e., a tablet or other device, which may be 202 in the example of FIG. 2. The tablet or other device 202 then relays the EEG data and patient information to, for example, a secure cloud server 204 running EEG review platform software. The emergency department staff then directs a neurological consultation to an EEG service at a tertiary care hospital, either within or outside the hospital network. An on-call epileptologist at a tertiary hospital can log on to a mobile application 206 running on device 207 to review the EEG in real time or substantially real time, while also using the quantitative EEG analysis features in the EEG review platform. Other or additional data may be provided by other sensors or devices as well, such as, for example, images or video of the patient captured by a camera, heart rate, pulse oximetry, or temperature data captured by an appropriate device. An embodiment may provide alerts or predictions based on the EEG data, such as, for example, a preliminary diagnosis, suggested options such as transferring the patient, continuing or discontinuing a drug treatment or intervention, or simply indicate areas within the EEG tracing data where a seizure or other EEG abnormality is suspected.

[0028] In one example, application 203 connects to sensor 201 via BLE, receives EEG data, buffers the EEG data, and transmits the EEG data via WIFI to cloud 204 where it can be retrieved and viewed by another clinician at a remote location. For example, the EEG data can be retrieved from cloud 205 and reviewed by a remote specialist on a medical-grade tablet or other device 205. This facilitates review of the EEG data displayed in application 206 running on device 205. The EEG data from both sensors 204 can be synchronized, for example, by a BLE command from tablet 202.

[0029] The cloud device 204 may provide a server instance running EEG review software that allows a specialist, such as a neonatologist or pediatric epileptologist, to log on to a tablet 205 running a mobile application 206 and view the EEG in real time as an aEEG. The aEEG data may include instructions such as the markings described with respect to FIG. 4A.

[0030] The general principles of the example of FIG. 2 can be extended to other scenarios. For example, in one embodiment for intensive care in children and adults, two sensors, four sensors, eight sensors, or various combinations of sensors may be used. Data flow and operation are similar to the example of FIG. 2, and an adult or pediatric epileptologist can log on to device 205 and monitor EEG data in real time, enabling them to make more accurate decisions more quickly and advise local hospital staff on appropriate care. Similarly, in a triage use case, as in the example of FIG. 2, four sensors, approximately F7, F8, TP9, and TP10, provide a total of eight electrodes that approximate the international 10-10 system and generate ten channels of EEG, as further described with respect to FIG. 4B. In one embodiment, EEG data acquired by various single-channel sensors can be synchronized and combined to create the desired differential montage. This may be accomplished, at least in part, with an application or software program that, for example, determines the difference montages available given the number and placement of sensors (e.g., 10-10), displays this to the user, and allows the user to select the desired configuration, or automatically configures the montage for the user.

[0031] In one embodiment, for example, an application running on device 202 visually guides a user, e.g., emergency medical personnel, through a step-by-step visual. This may include, but is not necessarily limited to, guiding the user through scanning the barcode on the patient bracelet and each sensor, positioning the sensors on the scalp, and ensuring that quality signals are recorded and relayed, e.g., to the cloud. In some embodiments, additional or reduced data may be provided. For example, EEG data may not be shown at the point of care, e.g., on device 202. This may be done to accommodate emergency room staff and physicians, who may consider such data display distracting. In other embodiments, such data may be displayed in connection with, for example, seizure type detection, suggested treatment, or alerts for specific actions, such as transport options. Thus, in one embodiment, an application running on device 202 may be designed to interact with the user only when there is a problem, such as insufficient signal quality from an individual sensor, in the form of a user alert. These user alerts are designed to, for example, flash a red or blue LED on each sensor to indicate to staff that interaction with the application is required to guide them through resolving the problem, such as obtaining guidance to reposition or reorient the sensor as described with respect to Figure 6.

[0032] Collection of EEG data by one or more single-channel sensors allows the EEG data to be reviewed along with event markers (as described further herein) to quickly determine areas within the EEG data that indicate seizures. As described further herein, the type and nature of detection, analysis, or classification the EEG data undergoes can vary depending on the use case or desired outcome. For example, for triage event marking, a model with a higher false positive rate can be used compared to a use case where real-time seizure prediction is desired. Similarly, for use in a home seizure diary, a simple thresholding process can be suitable to generate seizure counts and markers on the EEG data traces.

[0033] Referring to Figure 3, an example of EEG monitoring and instruction is shown. It is often difficult to diagnose seizure disorders through short- or long-term monitoring in epilepsy monitoring units using video-EEG. Furthermore, it is estimated that 20-30% of individuals who visit epilepsy centers are diagnosed with psychogenic non-epileptic seizures (PNES). Therefore, mechanisms that facilitate long-term monitoring, such as in-home seizure monitoring, are desirable.

[0034] In one embodiment, EEG data is acquired at 301 from a single-channel EEG sensor. As described herein, this may be a single sensor placed post-diagnosis, a single sensor placed pre-diagnosis, or multiple sensors used in various situations. The use of a single sensor may be more appropriate for home use, while multiple sensors may be more appropriate for clinical or supervised scenarios. Each such sensor provides single-channel EEG data.

[0035] The single-channel EEG data is classified at 302. The classification performed at 302 may be implemented using a variety of devices. For example, the classification at 302 may be performed by the sensor, on a local device that communicates with the sensor over a PAN, on a remote or cloud device that communicates with the sensor over a WAN, or on a suitable combination of the foregoing.

[0036] The classification performed in 302 can take a variety of forms. For example, a detection model may take the form of simple thresholding to detect brain activity over a specific amount or duration for general seizure detection. One or more models may be designed to detect specific types of brain activity known to be useful in specific clinical settings, such as the emergency medical environment described with respect to FIG. 2. For example, many signal processing techniques have been investigated over the past 40 years to analyze and extract information from captured EEG data. Signal processing techniques can be used to distinguish between ictal (seizure) and interictal (non-seizure) states, and there is a wide range of scientific literature and research describing various signal processing techniques that can be applied to EEG data. The most common information extracted from EEG data is spike wave generation, time-domain characteristics (e.g., range, variance, skew), frequency-domain characteristics, time-frequency-domain characteristics (e.g., wavelet decomposition), complexity measures (e.g., entropy, fractal dimension), correlation measures (e.g., inter-channel correlation), and state dynamics.

[0037] One embodiment employs machine learning techniques for use with single-channel EEG data at 302. In one embodiment, the machine learning model may be trained using EEG data from one or more sensors, such as sensor 101a, in combination with other EEG data, for example, collected via a wired or tethered EEG system. By way of example, as further described with respect to FIG. 5, two-second segments of interictal (pre-ictal) and ictal (ictal) EEG data may be extracted from the sensor recordings and used to identify seizure occurrence.

[0038] In one embodiment, in identifying an ictal state by classifying the EEG data at 302, correlation with additional data, such as historical data (e.g., patterns or trends, medical record information obtained from an EMR, etc.), environmental data (e.g., weather data), or user activity (e.g., behavior), may be used to assist in determining the onset of seizure activity or indicating its past occurrence, as shown at 304. For example, psychological, behavioral, and environmental cues may be identified as potential information correlating with the user's ictal (seizure) state. This additional data may optionally be used to classify the EEG data or to improve the reliability of an independently made classification.

[0039] As an example, user feedback may be used to improve the seizure detection process. First, at 302, automatic detection of general seizure activity may be performed. When an embodiment detects general seizure activity, user-supplied input may be used to improve the accuracy of the detection (e.g., regarding time, severity, etc.). For example, an input interface, such as a small button included on sensor 101a or an input element included in a mobile application, may be used to provide an indication of when the user feels that seizure activity is occurring, about to occur, or has just occurred. Similarly, the interface may allow the user to record the severity, duration, or other data related to the event. This feedback is subjective and may not be desirable to use as a reliable source for determining the occurrence of seizure activity. However, user-supplied data may provide insight into the user's experience of non-epileptic seizures. Thus, it may be used to confirm a detected seizure or lack of detection. For example, if EEG data is recorded with high quality and automatic analysis indicates no seizures, the user's repeated indication that a seizure is occurring may indicate that the user is experiencing something else, e.g., a non-epileptic seizure, a psychological event. Conversely, such data feedback may indicate that model or threshold tuning is necessary or desirable.

[0040] After EEG data is classified at 302, for example, as nominal (e.g., non-seizure, pre-seizure) or abnormal (e.g., pre-seizure or seizure), an embodiment may provide output in the form of instructions. In one embodiment, EEG data, such as pre-seizure EEG data, may be classified as nominal or abnormal depending on the features being used for classification, the context (e.g., a mode in which pre-seizure activity is ignored and classification as nominal in favor of a lower false positive rate may be selected, etc.), a variable threshold, etc. In one embodiment, if a seizure is detected as a result of the classification at 302, this classification at 303 may be used to provide instructions at 306, such as generating an alert (e.g., to the patient or clinician), marking the EEG segment that triggered the detection or correlates with additional data, incrementing a seizure count, or generating a report.

[0041] As a specific example, referring to FIG. 4A , one embodiment may provide, at 306, instructions in the form of marking a segment of the EEG data displayed in trace 401 a. Instructions 402 a may highlight the area of ​​the EEG trace that triggered the classification of a seizure event. This may facilitate review by an epileptologist or another clinician. For example, color coding on the trace or text or other graphical indicators may be automatically provided to facilitate identification of important or interesting portions of EEG trace 401 a. Additionally or alternatively, an automated program may provide, for example, a link or location marker to navigate to this portion of EEG trace 401 a, automatically or in response to manual input. This may facilitate rapid review of large amounts of EEG trace data 401 a, such as over several days, where the patient is continuously wearing sensors and a remote clinician wishes to quickly review significant events at periodic intervals.

[0042] EEG data from two or more sensors can be displayed in various ways. As shown in the example of FIG. 4B, EEG trace data from each sensor and differential EEG data obtained through comparison with another sensor can be displayed. For example, in FIG. 4B, four sensors, such as four single-channel EEG sensors similar to sensor 201, are used to record EEG data at locations near F7, TP9, F8, and TP10. These locations may correspond to the forehead (front left and right) and behind the ears (left and right), respectively, utilized in emergency medical settings, as described herein. In the example of FIG. 4B, four channels of EEG trace data from the associated sensors are listed from top to bottom. Then, differential traces, e.g., F8-TP10, F7-F8, TP9-TP10, F7-TP10, and F8-TP9, are listed. In one embodiment, the order of these EEG traces may be modified, for example, based on user preference (identified by user input, such as dragging and dropping traces), or by creating more or fewer differential traces or individual sensor traces. As a specific example, in a scenario where two sensors are used, two sensor traces, e.g., F7 and F8, may be displayed, along with one or more differential traces created from these sensor readings. In one embodiment, the creation or display of traces may be automated, for example, by an application such as that described with respect to FIGS. 6 and 4B determining sensor locations and automatically associating sensor pairs to create differential traces of interest.

[0043] Referring again to FIG. 3 , if the classification at 302 results in an abnormal classification, such as pre-seizure, an embodiment can provide an indication in the form of a forecast or prediction at 305. In one example, a pre-seizure classification can be made when the EEG data is abnormal, such as EEG data where changes in frequency and / or amplitude exceed a threshold obtained from a nominal EEG data trace but are not sufficient to confidently classify as a seizure event. Similarly, a pre-seizure event can occur when the EEG data matches a known pattern leading up to a seizure, for example, a pattern showing a characteristic frequency of changes in the EEG data that precede a seizure, or a characteristic amplitude change in the EEG data, or a combination thereof. As described in the example of FIG. 5 , features of the EEG data that can be useful in identifying such EEG trace data can be obtained from human-labeled training data. As with other classifications, this decision can be aided by reference to additional data 304, such as psychological, behavioral, or environmental data.

[0044] The instructions provided at 305 may include a prediction provided to a user, such as the wearer of the sensor. The instructions provided at 305 may take the form of a real-time prediction (e.g., generated within a second or two) that changes as the EEG data or other data 304 changes. The instructions provided at 305 may take other forms, such as hourly, daily, weekly, or other time period predictions. Time period predictions may be influenced by historical data accessed at 304; for example, increasing or decreasing trends in seizure frequency may help form or modify the prediction. Predictions may take various forms, such as a score or color displayed in the mobile application related to the likelihood of a seizure during a certain time period, such as imminent, likelihood in a certain day, within the next week, etc. Similarly, predictions may take the form of haptic, audio, or visual effects generated by the sensor or a connected local device, remote device, etc. Predictions may also be provided to other or additional users, such as a clinician or another user. In the case of a prediction provided to a clinician, the prediction may include an indication of a relevant diagnosis, such as hypoxic-ischemic encephalopathy, and associated actions, such as suggested treatment, e.g., therapeutic hypothermia, or automated or semi-automated actions, such as requesting consultation with an on-call specialist.

[0045] Real-time or impending seizure forecasting or prediction adds further complexity in that it must distinguish between interictal (no seizure coming), preictal (an ictal event will occur in the next X minutes / hours), and ictal (a seizure is occurring) states in a time-dependent manner. Similar to seizure detection, information features and machine learning techniques have been extensively tested for seizure prediction and detailed in the scientific literature. The success of seizure prediction is often measured by sensitivity (whether the warning of an upcoming seizure was correct or missed) and by the hourly false alarm rate or "time-in-warning" (how frequently the warning occurs). Not all people with epilepsy are able to identify psychological, behavioral, or environmental seizure triggers. However, more than half of epilepsy patients report at least one seizure trigger. Emotional stress is the leading seizure trigger, followed by behavioral factors that are well known to trigger seizures, such as sleep deprivation and fatigue. Other behavioral factors include alcohol consumption, anti-seizure medication non-adherence, and physical exercise. Environmental factors such as time of day, light flicker, and weather (e.g., ambient temperature, relative humidity) have been shown to increase susceptibility to seizures.

[0046] Similar to seizure detection, seizure prediction can take the form of classification performed at 302. Similarly, additional data, such as historical data, environmental data, or user-provided data, can be used to generate or modify the classification of the EEG data as nominal or abnormal. This data can be used to form the forecast or prediction provided at 305.

[0047] The additional data used for classification at 302 may include, but is not limited to, historical data (e.g., seizure tendency data, etc.), environmental data (e.g., weather, stimulus data such as exposure to flickering light, etc.), and user data (e.g., behavioral data). User data may be provided directly or indirectly by the user. For example, user data may be entered directly by the user, such as by entering self-assessment data into a mobile application. Non-limiting examples include stress levels, sleep quality ratings, sleep scores according to known scales, sleep duration, etc. User or other data may be obtained indirectly, for example, from a linked health app, from medical records, from input by another user on another device such as a doctor, or from another device such as a mobile phone or smartwatch providing accelerometer data.

[0048] In one embodiment, the instructions may take the form of a report, as shown, for example, at 306. For example, a digital seizure diary report may be provided to the patient or clinician. The epileptologist will have an accurate, quantitative record of the patient's seizure activity, will know if treatment is effective, and can adapt the patient's treatment more quickly and successfully.

[0049] Improvements in seizure detection or prediction depend heavily on the quantity and quality of collected EEG data. Currently, no practical EEG database exists. While there are several laboratory EEG databases (e.g., the MIT database), the EEG data in these databases is too clean to be used for prediction, as it does not represent the quality of EEG data collected in the real world. The ease of EEG data collection improves EEG data availability, such as through the use of EEG data collection sensors 101a, 101b and their minimal impact on a user's daily routine. Thus, one embodiment features the use of single-channel EEG data collection sensors 101a, 101b, accompanying EEG data analysis, and seizure prediction techniques. While single-channel EEG data collection may not be able to accurately identify where seizure activity always or initially occurs in a user's brain, simple detection of the onset of seizure activity presents a valuable tool to help users manage and treat epilepsy. As described, the location or placement of the sensor may be refined over time, e.g., with respect to preliminary or subsequent full EEG montages, with respect to analysis of sensor data quality, etc. Additionally, the predictive functionality may provide the user with an improved quality of life, as the user may be able to perform activities with reduced anxiety that an unexpected attack may occur.

[0050] One aspect of identifying and / or predicting seizures via classification in 302 may include distinguishing between various seizure types. For example, absence seizures typically occur many times per day, and the electronic signature of such seizures is the same across all ages. Therefore, it may be relatively easy to collect large-scale EEG data and improve seizure prediction models through machine learning to detect absence seizures. Other types of seizures may occur less frequently, such as once a month, and therefore may be difficult to accurately predict due to a lack of EEG data on which to train a machine learning model. However, starting with the creation of generalized seizure prediction models for common seizure types, an embodiment may be expanded and refined by using models covering other seizure types, especially as users continue to wear sensors 101a, 101b over long periods of time. This EEG data may be stored and used, with the user's permission, to build a database suitable for forming future seizure detection and prediction models.

[0051] Referring now to Figure 5, an example of EEG data classification is shown. The exemplary classification technique shown in Figure 5 may be used to provide classification as part of a larger processing technique, for example, as shown in Figure 3.

[0052] In one embodiment utilizing a machine learning process to classify EEG data, the training phase may include the process outlined at 501-505 in FIG. 5. By way of example, as shown at 501, a patient wears multiple sensors, e.g., four single-channel EEG sensors, such as sensors 201 of FIG. 2. In one example, the four sensors 201 may be positioned on the forehead and behind the ears, e.g., one sensor on the left forehead, one sensor on the right forehead (F7 / F8), one sensor behind the left ear, and one sensor behind the right ear (TP9 / TP10). The sensors may be worn for a period of time to collect training EEG data; for example, a patient may wear sensors 201 for seven days during a stay in an epilepsy monitoring unit (EMU).

[0053] In one example, the training data may include both single-channel EEG data collected using sensor 201 and EEG data collected using a regular 10-20 or 10-10 multi-channel wired EEG sensor or headset as part of standard care, i.e., both sensor 201 and the wired EEG sensor or headset may be worn by the same patient at the same time to obtain a set of EEG training data.

[0054] During the EMU stay, the review software and one or more of the epileptologist identify potential seizure events in the wired EEG data recordings. The patient and / or family in the room may also indicate a seizure by, for example, pressing a button. The epileptologist reviews the entire multi-day wired EEG data with the review software and user-provided event markers to determine when a seizure occurred (as is conventional). The epileptologist may also review the EEG data (as is conventional) to identify what type of seizure event occurred. If the seizure was focal in origin, the epileptologist may indicate which wired EEG electrode was the center of the focus. The epileptologist may also indicate the EEG start / stop of the seizure, whether electronic seizures were visible on the wired EEG at each sensor location, and the start / stop of electronic obscuring artifacts (e.g., patent movements). This information indicates when a seizure is electrically visible (as is conventional), which can be used to compare with data acquired via the four sensors 201.

[0055] Preprocessing of the raw EEG data is performed, as shown at 502. For example, noise removal or filtering may be applied, such as removal of 50 / 60 Hz line noise, low-pass filtering to remove electromyographic (EMG) muscle activity, and standardization normalization to account for inter-patient and inter-sensor differences in data amplitude. Other or additional signal processing performed at 501 may include, for example, electro-ocular artifact removal (to remove the effects of eye movement) from specific sensor placements, such as F7 / F8 placement sensors.

[0056] At 503, the pre-processed data is segmented into short-term segments (e.g., 0.5-10 seconds). In one embodiment, each segment is labeled as a seizure or non-seizure based on, for example, its origin from a time previously noted during a seizure in the patient application phase of 501, and whether the seizure was visible at the location of the sensor 201.

[0057] Feature extraction is performed at 504. In one embodiment, one or more of the following features are extracted from the segmented data: time domain (minimum, maximum, mean, median, range, variance, standard deviation, skew, kurtosis), frequency domain (Fast Fourier Transform, EEG eigenbands (s, delta, θ, α, β, γ)), time-frequency domain (wavelets), complexity domain (sample / spectral entropy, nonlinear energy operator, Hjorth parameter, fractal dimension), transforms (principal component analysis, linear discriminant analysis), and history (past segment values ​​that may be weighted).

[0058] Model tuning is performed at 505. Because EEG sensor data is highly imbalanced (e.g., a 100:1 ratio of non-seizure segments to seizures), a subset of the EEG data may be used (e.g., a 3:1 EEG data / model). A machine learning model, such as a random forest as shown in FIG. 5, or a support vector machine, an artificial neural network (shallow or deep), etc., is trained at 505 and tuned on the training data. For example, tuning may include hyperparameter tuning, feature relevance determination, cross-validation (e.g., using a leave-one-out (LOO) method), etc.

[0059] Metrics such as the area under the receiver operating curve (ROC) curve (AUC), specificity, sensitivity, positive predictive value, false positive rate, or any combination of the foregoing can be used to determine the best model. In some cases, for example, where seizure detection is paramount and false positives are acceptable, e.g., in the context of a seizure diary, a particular model may be selected over use in another scenario, e.g., where false positives should be minimized, such as providing automated medication recommendations or suspected diagnoses. In one embodiment, the model used may be replaced or modified, for example, by adjusting parameters such as a probability threshold to fit the usage context. As an example, one embodiment may adjust the model used by providing an input interface for end-user data, such as displaying a selectable element indicating the usage context and, after selection, loading a predetermined model or set of parameters for the indicated context. For example, a context such as a seizure diary, real-time alerts, or emergency medical care may be indicated via selection, and a model may be loaded or model parameters, e.g., a probability threshold that may be modified to adjust sensitivity, may be adjusted to match the indicated context. More experienced users may interface more directly with model parameters.

[0060] Following tuning at 505, the tuned model may be saved for use in detection. For example, at steps 506-510, a model, such as a previously tuned model, is accessed and run on patient data collected using one or more sensors 201, e.g., in a treatment scenario. As described herein, this may include the process of segment detection, as outlined at 506-510. By way of example, at 506, a patient, e.g., previously diagnosed with a seizure disorder, wears one or more sensors 201, e.g., using placement guided by the patient's epileptologist as most likely to detect seizure events. No wired EEG data is recorded, and the patient wears the sensors 201, e.g., during daily activities. In one example, the patient wears the sensors 201 up to 24 hours a day, and may wear them for several days.

[0061] EEG data collected by sensor 201 may be streamed to a remote device such as a cloud platform, as in the example of Figure 2. Raw sensor EEG data preprocessing is performed at 507, EEG segments are identified at 508, and features are extracted at 509, which may be similar to the processing performed at 502, 503, and 504, respectively.

[0062] At 510, the unlabeled segmented feature set is run through a trained model, the output of which may be utilized in the overall seizure detection process, for example, as outlined at 511-514.

[0063] In the example of FIG. 5 , the output of the model performed at 510 may include a segment probability 511 for a seizure event for each segment or set of segments. In one embodiment, the output of the model performed at 510 is a likelihood between 0 and 1 that the segment occurs during a seizure event. In some embodiments, the specific type of seizure may or may not be determined, for example, if a machine learning model tailored to a particular seizure type is used, such as by user selection of such a model. In one embodiment, for example, as outlined at 510-514, the general seizure detection model and discrimination process resembles a series of models with probabilities for each seizure type, combined or considered together to make a seizure / non-seizure decision.

[0064] At 512, segment stitching is performed. For example, segment probabilities are combined to create a start / stop time for an (overall) seizure event, i.e., consisting of multiple segments. This can be achieved in many ways. By way of example, segments may be stitched together or combined via individual segment thresholding (e.g., comparing changes in frequency and amplitude EEG data to a threshold for each segment), a multi-segment thresholding and windowing process (e.g., combining or considering multiple segment probabilities together compared to one or more thresholds), or joint windowing (e.g., weighted, leaky, etc.). Windows are typically as short as a few seconds (e.g., absence seizures) or up to several minutes in duration (e.g., focal seizures).

[0065] At 513, annotations are generated for the determined seizure event start / stop times, e.g., for the patient's medical record or EEG trace display. This annotation list data can be utilized in a variety of ways, e.g., as instructions according to FIG. 3. For example, one embodiment can present an annotation list (a list of segments of EEG data associated with seizures or metadata for identifying such EEG data) directly to the patient via an application running on a local device, or directly to a clinician (e.g., if the clinician is remote) via a local or remote device. The annotation list data, along with the raw EEG trace data, can also be stored on a cloud platform for review by the clinician.

[0066] At 514, an epileptologist review stage is shown. In one embodiment, a clinician can review the annotations and / or raw EEG data and make a clinical decision. The process of epileptologist or simply user review at 514 can be influenced by context. For example, in one embodiment, a low threshold that leads to many false positives can be set (as described herein). In such an embodiment, an overall seizure / no-seizure decision may or may not be made. Rather, the annotations generated at 513 may be provided to a clinician, for example, for more rapid review of potentially interesting EEG data, and a seizure / no-seizure decision may be made by the clinician based on this review. In other contexts, the review may be performed by another user, for example, a user at home creating a seizure diary. In this context, additional or different data, for example, time or location context data indicating when potentially interesting EEG data was recorded, may be displayed to facilitate user review.

[0067] Referring to FIG. 6 , to facilitate the use of one or more EEG sensors, such as sensor 201, by non-expert users, one embodiment provides an application, such as a mobile application for use on a mobile device, that guides the user through placing the sensor and recording EEG data. By way of example, as shown in FIG. 6 , one embodiment allows a user to easily capture sensor or patient data using a mobile device. In the example of FIG. 6 , at 601, the application acquires sensor data, for example, via capturing an image of a barcode, QR code, or other coded data, for example, provided with each sensor. This allows one embodiment to automatically identify the sensor. Similarly, patient data may be captured automatically or semi-automatically by the application. Of course, data may also be entered manually, similar to other data entry described herein. An exemplary screen or application view of a user capturing sensor data from a barcode is provided in FIG. 7A . As shown in FIG. 7A , the application can capture an image of the barcode and automatically enter the captured sensor information (e.g., identification information formed from the barcode or other captured data) into a display screen. The application may further indicate to the user the number of sensors used for the application or context. In the example of Figure 7A, four sensor locations are shown, two of which are successfully identified, which helps the user determine the number of sensors to be used in a scenario, e.g., four sensors for an emergency medical scenario.

[0068] In one embodiment, the application may further display instruction steps to the user. For example, the application may display instructions for pairing the sensor, verifying that the sensor is connected to the mobile device, verifying that the mobile device is connected to a remote device (e.g., a cloud platform), preparing the sensor to be adhered to the patient, determining an appropriate location for sensor placement on the patient, repositioning the sensor within the application, recording data, and storing data (locally or remotely), which may be automatic or semi-automatic routines accomplished by user input such as turning on or activating the sensor, pressing a button, etc. Additionally, the application may include additional or alternative display features, such as the ability to conduct live video calls with specialists, clinicians, etc., the latter of which may assist in a live or real-time troubleshooting or diagnostic context.

[0069] When a user activates a sensor and connects it to a mobile device running the application, this is confirmed by the application, as shown at 602. For example, the application can display the sensor in a location diagram, as shown in FIG. 7B. This helps the user determine whether the sensor is appropriately positioned on the patient for a given context, and this is accurately reflected in the application view. In the example of FIG. 7B, the sensor is positioned at the forehead and behind the ear, as shown. If a given sensor ID in the diagram provided by the application does not match the actual location of the physical sensor on the patient (e.g., visible on a barcode or otherwise identifiable on the physical sensor), the user can select (e.g., by touch, in the case of a mobile touchscreen) the target sensor icon to reassign its location in the diagram. As an example, if sensor 502 in FIG. 7B is shown by the application as being located on the right front, but was actually placed by the user on the left front, the user can simply reposition it by interfacing with the application. This can take a variety of approaches. In the example of Figure 7B, a user can touch the selectable icon shown at 702 for sensor 502 to invoke a menu for swapping its position with another sensor in the array, as shown at 701. Similarly, another mechanism, such as dragging and dropping icon 702, can be used to rearrange sensors within an application. This ensures that the EEG data collected by the actual sensor is known to the application, for example, to create differential EEG data via subtraction from another sensor position.

[0070] 6 , an embodiment may verify that the sensor position is appropriate at 603, and then, if a decision to reposition the sensor does not come, as shown at 604, begin a recording session, as shown at 605. The recording session may be terminated at a predetermined time, as determined at 606, or based on another factor, such as user selection or interface. Thereafter, as shown at 607, the EEG data for the session may be stored locally, remotely, or both. As described herein, other activities may be performed by an embodiment during the recording at 605. For example, the sensor EEG data may be analyzed locally or remotely, for example, by a cloud platform, a remote clinician, etc.

[0071] It will be readily understood that some embodiments may be implemented using any of a wide variety of devices or combinations of devices. Referring to FIG. 8 , an exemplary system-on-chip (SoC) included in a computer 800 is shown, which may be used in implementing one or more embodiments. The SoC outlined in FIG. 8 or similar circuitry may be implemented in a variety of devices in addition to computer 800; for example, similar circuitry may be included in a sensor 870 or another device or platform 870a. In addition, circuitry other than the SoC, an example of which is provided in FIG. 8 , may be utilized in one or more embodiments. The SoC of FIG. 8 , as shown, includes functional blocks integrated onto a single semiconductor chip to meet specific application requirements.

[0072] The central processing unit (CPU) 810, which may include one or more graphics processing units (GPUs) and / or microprocessing units (MPUs), includes an arithmetic logic unit (ALU) that performs arithmetic and logical operations, an instruction decoder that decodes instructions and provides information to the timing and control unit, and registers for temporary data storage. The CPU 810 may comprise a single integrated circuit that comprises several units, the design and arrangement of which will vary depending on the selected architecture.

[0073] The computer 800 also includes a memory controller 840, which may comprise, for example, a direct memory access (DMA) controller for transferring data between the memory 850 and hardware peripherals. The memory controller 840 includes a memory management unit (MMU) that functions to handle cache control, memory protection, and virtual memory. The computer 800 may communicate with various communication protocols (e.g., I / O, 2 It may include a controller for communicating using a network interface (e.g., USB, USB, etc.).

[0074] The memory 850 may include various memory types, volatile and non-volatile, for example, read-only memory (ROM), random-access memory (RAM), electrically erasable programmable read-only memory (EEPROM), flash memory, and cache memory. The memory 850 may include embedded programs and downloaded software, such as EEG processing software. By way of example and not limitation, the memory 850 may also include an operating system, application programs, other program modules, and program data.

[0075] The system bus enables communication between various components of the computer 800. An I / O interface 830 and a radio frequency (RF) device 820, e.g., WIFI and telecommunications radio, BLE devices, etc., are included to enable the computer 800 to send and receive data to the sensor 870 or remote device 870a using wired or wireless mechanisms. The computer 800 may operate in a networked or distributed environment using logical connections to one or more other remote computers or databases. The logical connections may include a network such as a personal area network (PAN), a local area network (LAN), or a wide area network (WAN), but may also include other networks / buses. For example, the computer 800 may communicate data between the sensor 870 and the remote device 870a, as well as between the sensor 870 and the remote device 870a, via the Internet.

[0076] Thus, computer 800 can execute program instructions configured to store and analyze EEG data and perform other functions of the embodiments, as described herein. A user can interface with computer 800 (e.g., input commands and information) via input devices, which may be connected to I / O interface 830. A display or other type of device may also be connected or coupled to computer 800 via an interface selected from I / O interface 830.

[0077] It should be noted that various functions described herein may be implemented using executable instructions stored in a memory, e.g., memory 850, that are transmitted to and executed by a processor, e.g., CPU 810. Computer 800 includes one or more storage devices that persistently store programs and other data. A storage device, as used herein, is a non-transitory storage medium. Some additional examples of non-transitory storage devices or media include, but are not limited to, storage integrated into computer 800, such as a hard disk or solid-state drive, and removable storage, such as an optical disk or memory stick.

[0078] The program code stored in a memory or storage device may be transmitted using any suitable transmission medium, including but not limited to wireless, wireline, optical fiber, cable, RF, or any suitable combination of the foregoing.

[0079] The program code for performing the operations may be written in any combination of one or more programming languages. The program code may run entirely on a single device, partially on a single device as a standalone software package, partially on a single device and partially on another device, or entirely on another device. In some cases, the devices may be connected via any type of connection or network, including short-range wireless mechanisms such as a LAN, WAN, PAN, short-range communication mechanisms, or the connection may be made through another device (e.g., over the Internet using an Internet Service Provider) using a wireless connection or via a hardwired connection such as a USB connection.

[0080] Exemplary embodiments are described herein with reference to figures illustrating exemplary methods, devices, and program products according to various exemplary embodiments. It will be understood that operations and functions may be implemented at least in part by program instructions. These program instructions may be provided to a processor of a device to create a special-purpose machine, such that the instructions, executed via the processor of the device, implement the specified functions / operations.

[0081] It should be noted that while particular elements are used in the figures and particular orderings of elements are shown, these are non-limiting examples. In certain contexts, two or more elements may be combined, elements may be divided into two or more elements, or some elements may be rearranged, reorganized, or omitted, as appropriate, since the explicitly shown examples are used for illustrative purposes only and should not be construed as limiting.

[0082] While illustrative embodiments have been described herein with reference to the accompanying drawings, it should be understood that this description is not intended to be limiting and that various other changes and modifications can be effected therein by those skilled in the art without departing from the scope or spirit of the present disclosure. [Explanation of symbols]

[0083] 101 Sensors 201 Sensors 202 devices 203 Applications 204 Cloud 204 Sensors 205 Cloud 205 devices 206 Mobile Applications 207 devices 304 Additional Data 800 computers 810 Central Processing Unit (CPU) 830 I / O interface 840 memory controller 850 memory 870 Sensor 870a Remote Device

Claims

1. acquiring EEG signals from one or more single bipolar channel wireless EEG sensors placed on the patient's scalp, the EEG sensors configured to sense a voltage difference between two electrode contacts; transmitting the EEG signals to a remote device; The remote device: automatically creating a vertical-horizontal montage from the EEG signals by bipolar derivation, and subtracting, for each of the sensors, each corresponding EEG signal from the corresponding EEG signals of the remaining sensors to create the vertical-horizontal montage; detecting seizure events based on the longitudinal and transverse montage; providing information associated with the seizure event; A method comprising:

2. The method of claim 1 , wherein the one or more EEG sensors include four EEG sensors.

3. The method of claim 1 , wherein the horizontal-vertical montage comprises 10 channels.

4. 10. The method of claim 1, wherein each of the one or more EEG sensors includes two electrodes that form a single bipolar channel.

5. The method of claim 1 , further comprising providing one or more channels of the portrait-landscape montage to be displayed.

6. providing said portrait-landscape montage to be displayed; The method of claim 1.

7. 10. The method of claim 1, wherein the one or more EEG sensors include four EEG sensors configured to be positioned on the patient's scalp at the left forehead, the right forehead, behind the left ear, and behind the right ear, respectively.

8. 2. The method of claim 1, wherein the horizontal and vertical montage includes channels F7, F8, TP9, TP10, F7-F8, F7-TP9, F7-TP10, F8-TP9, F8-TP10, and TP9-TP10.

9. the one or more EEG sensors are spaced apart such that they are above each eye and behind each ear; The method of claim 1.

10. one or more EEG sensors configured to be placed around the patient's scalp and to collect EEG signals; automatically creating a vertical-horizontal montage from the EEG signals by subtracting, for each of the sensors, each corresponding EEG signal from corresponding EEG signals of the remaining sensors; detecting seizure events based on the longitudinal and lateral montage; providing information associated with the seizure event; a non-transitory memory storing instructions executable by at least one processor to perform the A system including:

11. 11. The system of claim 10, wherein the one or more wireless EEG sensors include four wireless EEG sensors.

12. The system of claim 10 , wherein the vertical and horizontal montage includes 10 channels.

13. 11. The system of claim 10, wherein each of the one or more wireless EEG sensors includes two electrodes that form a single bipolar channel.

14. 11. The system of claim 10, wherein the instructions are further configured to cause the at least one processor to: provide one or more channels of the portrait-landscape montage to be displayed.

15. The system of claim 10 , wherein the instructions are further configured to cause the at least one processor to: provide the portrait-landscape montage to be displayed.

16. 11. The system of claim 10, wherein the one or more wireless EEG sensors include four wireless EEG sensors, each configured to be positioned on the patient's scalp at a left forehead, a right forehead, behind a left ear, and behind a right ear, respectively.

17. 11. The system of claim 10, wherein the vertical and horizontal montage includes channels F7, F8, TP9, TP10, F7-F8, F7-TP9, F7-TP10, F8-TP9, F8-TP10, and TP9-TP10.

18. 11. The system of claim 10, wherein the one or more wireless EEG sensors are spaced apart from one another such that the sensors are above each eye and behind each ear.

Images