Elimination of artifacts in cranial acceleration signals

Abstract

A computer-aided method for processing data acquired using accelerometers in contact with a patient's anatomical body part is disclosed. The method involves acquiring acceleration measurement data via accelerometers and identifying potential artifacts within the acceleration signals. Sections of the acceleration signals containing identified artifacts are then ignored or the artifacts are removed in subsequent processing steps, allowing for a comparison between artifact-free acceleration signals acquired from different subjects with similar and / or differing physiological status.

Description

TECHNICAL AREA

[0001] The present invention relates to a computer-implemented method for processing data acquired using accelerometers touching an anatomical part of a patient's body, a corresponding computer program, a computer-readable storage medium storing such a program, a computer executing the program, and a medical system comprising an electronic data storage device and the aforementioned computer. STATE OF THE ART

[0002] The detection of brain abnormalities such as large vessel occlusions, aneurysms, or vasospasms is typically performed using imaging techniques (e.g., CT, MRI, etc.) and other technical methods such as EEG, Doppler sonography, cerebral ultrasound, or auscultation via microphones. Most of these procedures can only be performed in a clinical setting, creating a need for a mobile application that enables diagnostic support directly at the patient's bedside. Cranial acceleration measurement has been described as a potential solution but is highly susceptible to interference. The approach described here addresses this problem by providing a user-friendly and reliable system for generating data suitable for analysis.

[0003] Accelerometers are very sensitive and therefore prone to artifacts. For meaningful conclusions to be drawn from recordings of cranial acceleration, high-quality and largely artifact-free data are required for analysis.

[0004] The aim of the present invention is to detect and remove artifacts in acceleration signals so that artifact-free data are available for the analysis of these signals in order to reliably identify indicative features which can then be used to classify a subject / patient on the basis of this indication.

[0005] The following discloses aspects of the present invention, examples and exemplary steps and their embodiments. Various exemplary embodiments of the invention can be combined according to the invention, insofar as this is technically expedient and feasible. EXAMPLE-STYLE SHORT DESCRIPTION OF THE INVENTION

[0006] The disclosed computer-implemented method for processing data acquired by accelerometers touching a patient's anatomical body part comprises acquiring acceleration measurement data via accelerometers and identifying potential artifacts within the acceleration signals. The segments of the acceleration signals containing identified artifacts are then disregarded or the artifacts are removed in subsequent processing steps, enabling a comparison between artifact-free acceleration signals acquired from different subjects with similar and / or different physiological conditions. GENERAL DESCRIPTION OF THE INVENTION

[0007] This section describes the general features of the present invention by way of example with reference to possible embodiments of the invention.

[0008] In general, the invention achieves the aforementioned purpose by providing, in a first aspect, a computer-implemented medical method for processing data acquired by means of accelerometers touching an anatomical part of a patient's body. The anatomical part may be the patient's head, but also any other anatomical part. The method comprises carrying out the following exemplary steps on at least one processor of at least one computer (for example, at least one computer that is part of a portable device, such as a mobile device like a mobile phone or a tablet computer).

[0009] In an exemplary (for example, first) step, acceleration measurement data is acquired, describing the acceleration signals detected by the accelerometers. The accelerometer(s) can be fixed using any device suitable for maintaining contact between the sensors and the patient's anatomical body part. In the case of cranial acceleration measurement, a headband-like device equipped with multiple accelerometers can be used. When the headband is worn by the patient, the multiple sensors are positioned at various desired locations on the patient's head, where they detect the acceleration of the skull and send corresponding signals to a computer for further processing.

[0010] In a (for example, second) exemplary step, artifact data is determined based on the acceleration measurement data. This data describes the presence of one or more artifacts in the acceleration signals. These artifacts are unwanted deviations or errors in the acceleration signals that originate from sources other than the accelerations being measured in the area of ​​interest, which is an anatomical body part and could, for example, be the skull. Artifacts can arise, for example, from ambient noise such as speech, the patient's breathing sounds, noise caused by the physical contact of the accelerometers with objects such as the patient's pillow, operating noises from people or machines, or even acoustic signals from machines.Since these artifacts have a different origin than the cranial accelerations of interest, they must be identified and removed from the acceleration signals. Artifacts can be identified, for example, using sensors suitable for measuring ambient noise, such as sound pressure level sensors, which can be provided together with the acceleration sensors, e.g., as part of a sensor carrier structure like the headband mentioned above. Similarities in the measurements acquired by the one or more sound sensors and the one or more acceleration sensors can thus indicate artifacts caused by ambient noise that must be eliminated. This can be done by excluding signal segments containing the identified artifacts from further processing.Alternatively, artifacts can be eliminated by determining and removing their waveform from the acceleration signals in the time domain. Another alternative involves performing a spectral analysis of the measurement signals followed by spectral subtraction of the identified artifacts to remove them. In the latter case, the artifact-containing signal segments can be retained for further processing. For example, to identify artifact data, the acquired acceleration signals are decomposed into a multitude of frequency ranges, and the decomposed signals are then analyzed for artifacts.

[0011] Artifacts can also be caused by structure-borne noise, for example, when accelerometers come into contact with objects such as the patient's pillow. These artifacts can be identified by a time-domain analysis of the acceleration signals. Sharp, spike-like artifacts can indicate such contact and are therefore easily identified by a Fourier analysis of the acceleration signals in the frequency domain, where the signal was transferred from the time domain. Once artifacts have been identified, they are either removed from the acceleration signals or the affected sections of the acceleration signals are ignored in subsequent processing steps.

[0012] In an exemplary (for example, third) step, usable signal data is determined based on the acceleration measurement data and the artifact data. The usable signal data describes continuous segments of the acceleration signals that are free of identified artifacts; that is, no acceleration signals containing artifacts are processed in subsequent steps. In other words, all acceleration signals described by the usable signal data are generally available for further processing, where features in the acceleration signals are identified and ultimately compared with features from acceleration signals acquired from other patients / subjects with a specific physiological status. In one example, this involves acquiring additional sensor data of a different type than the acceleration measurement data, specifically describing at least one of the following features. - an acoustic noise from the environment, including structure-borne sound and / or airborne sound; and - a patient's heart rate; whereby, at least partially based on the additional sensor data, the artifacts are identified and / or the continuous sections are divided into segments.

[0013] In an exemplary step (for example, a fourth step), feature data is determined based on the usable signal data, describing at least one predefined feature in the acceleration signals. In other words, unique features are identified in the acceleration signals, whereby the identification can be based on known properties of features that may be stored in a database. Such features can be identified within the time domain of the acceleration signals, within time-frequency representations, and / or within the frequency domain of the acceleration signals, and / or within a statistical range of the acceleration signals. Features can be identified in the acceleration signals received by at least one acceleration sensor, and in particular by a plurality of acceleration sensors.Acceleration signals received from a multitude of sensors can even be subjected to combined analysis to identify different features among these signals. For example, determining feature data involves determining signal component data based on the usable signal data and / or the segmented signal data that describe at least one of the continuous sections and / or at least one of the segments thereof, which are decomposed into a multitude of frequency ranges.

[0014] In an exemplary step (for example, a fifth step), reference feature data are determined that describe at least one predefined feature in acceleration signals recorded for at least one reference individual with a specific condition. In other words, acceleration signals, which may have been previously recorded and stored in a database from one or more, preferably multiple, reference individuals, are compared to identify similarities or differences with respect to unique features identified in the respective acceleration signals. It is important to note that the measurements taken by the respective subjects must be sufficiently similar in their recording to allow for a comparison between them and a meaningful derivation of similarities and differences regarding the features contained in the compared signals.

[0015] In an exemplary (for example, sixth) step, classification data are determined based on the feature data and the reference feature data. These data describe a degree of similarity between the at least one predefined feature in the acceleration signals recorded for the patient and the at least one predefined feature in the acceleration signals recorded for the at least one reference subject. For example, the features identified in the acceleration signals recorded from the patient and the at least one reference subject are compared to find similarities or differences in the respective acceleration signals.Based on such a comparison, the current patient can be compared with a selection of reference subjects, with any approach using thresholds, confidence intervals, clustering, feature registration, or any possible combination thereof being considered to classify the current patient in relation to the reference subjects.

[0016] For example, determining classification data involves defining one or more classes, each class comprising at least one reference subject, and defining at least one threshold for the at least one predefined feature in the acceleration signals acquired for the patient. Assigning the patient to one or more classes is based on whether the at least one predefined feature in the acceleration signals acquired for the patient exceeds the at least one threshold. A threshold-based approach may involve determining a class for the patient by defining a predetermined threshold based on one or more reference subjects assigned to that particular class.Depending on whether at least one characteristic of the subject exceeds the predetermined threshold or not, the subject is assigned to this particular class or not.

[0017] In another example, confidence intervals can be used for classification. For the one or more characteristics extracted from the patient's acceleration measurements, it is determined whether they lie within or outside a predefined confidence interval. If the one or more characteristics lie outside a confidence interval for a particular class, the characteristic, or even the patient, is assigned to a different class.

[0018] In another example, determining classification data involves defining one or more clusters in a space with two or more dimensions. The one or more clusters are defined by collections of predefined features in the acceleration signals recorded for a large number of reference individuals. Patient assignment to the one or more clusters is based on the distance between the at least one predefined feature and the one or more clusters in the space with two or more dimensions. Cluster analysis or cluster formation can be used for classification, where cluster collections of predefined features are assigned in a multidimensional space, such as 2D or 3D. Within each cluster, a cluster center is determined, for example, using K-means clustering.Consequently, the classification can be based on determining the distance between the one or more characteristics designated for the patient and the respective cluster centers, so that the characteristic or even the patient is assigned to the nearest cluster.

[0019] In another example, determining classification data involves performing a regression analysis, such as linear or sigmoidal regression. The regression analysis may also include the use of thresholds to perform binary classifications.

[0020] For example, regression analysis is used for predictive modeling, where an algorithm is used to predict continuous outcomes such as medical assessments, infarct sizes, etc. To do this, the algorithm is trained to understand the relationship between independent variables (e.g., medical assessments, infarct sizes, etc.) of known data points and their respective outcomes (e.g., stroke type, patient outcome, etc.). The trained model is then used to predict the outcome of the actual regression data features.

[0021] As will be described in more detail below, the feature space containing the features extracted from the acceleration signals can be extended to include further features extracted from other possible measurements on the patient or other data sources.

[0022] Regarding the acceleration signals, as well as signals / data received from other data sources, all features can be extracted from the original signals (e.g., from the time domain of the signals), from decomposed signals (e.g., from the frequency domain of the signals), and from any relationships or comparisons between these signals. Such a comparison can include differences, ratios, or any other possible mathematical comparison between these signals. In this context, it should be noted that the acceleration signals can be influenced by the patient's respiration. Therefore, it may be desirable to determine the correlation between the respiratory cycle on the one hand and the acceleration signals on the other. For example, modulation of cardiac-related signals or information can be used to approximate the patient's respiratory cycle.Such cardiac signals or information can be obtained from one or more PPG sensors, a phonocardiogram, a seismocardiogram, or similar sources. Based on this, the phase of the respiratory cycle and the acceleration signal, in particular the continuous sections or the segments derived therefrom, can be correlated. Specifically, the segmented signal data can be enriched with corresponding information about the associated respiratory cycle phase for subsequent processing. This ultimately enables the identification and selection of segments with the same or different respiratory phases for or during subsequent procedural steps described herein.

[0023] In addition to or as an alternative to the statistical approaches mentioned above, machine learning approaches can also be used to classify the patient in relation to a number of reference subjects. For example, features extracted from acceleration signals or obtained from other measurements or data sources can be fed into a pre-trained machine learning algorithm. This algorithm is trained on features derived from a sufficiently large number of reference subjects who either have or do not have a specific physiological condition. For instance, determining classification data involves using a pre-trained machine learning algorithm, where a training set for the machine learning algorithm comprises a large number of reference subjects with the specific condition and a large number of reference subjects without the specific condition.

[0024] In an example of the procedure according to the first aspect, determining usable signal data involves determining segmented signal data based on the usable signal data, which describes the continuous sections subdivided into a multitude of segments. For example, the sections free of identified artifacts are subdivided or divided into smaller, i.e., shorter, (sub)segments. The subdivision can be based on a heartbeat signal, acquired, for example, by one or more heartbeat sensors, such as photoplethysmography (PPG) sensors. Alternatively, the heartbeat can also be detected within the acceleration signals by identifying characteristic, repeating waveforms within the acceleration signals.In this case, the heartbeat can be detected directly via the signals captured by the one or more accelerometers, as described, for example, in US 2018 / 0296107, or, for example, by filtering for repeating waveforms in the time series signal that describes the heartbeat.

[0025] In this way, the continuous, artifact-free sections can be divided into segments that can have a length of at least one cardiac cycle.

[0026] It is also possible that each cardiac cycle is divided into a multitude of segments, covering, for example, a systolic and / or diastolic phase of the heartbeat. Furthermore, the segments can cover essentially the same period of time. In a specific example, each segment essentially covers one cardiac cycle.

[0027] In another example, the segments covering one or more cardiac cycles are searched for additional artifacts not described in the artifact data and therefore not removed, or for signal segments containing these artifacts that have not yet been considered. If further artifacts are identified within the heartbeat-segmented signals, the corresponding segments can be disregarded for further processing. Artifact identification can be based on the same and / or different principles implemented for the artifact identification described above with respect to the non-segmented acceleration signals. It should be noted that arrhythmic and / or ectopic heartbeats can occur under both normal and pathophysiological conditions in the patient.Since cranial acceleration signals are sensitive to arrhythmic and / or ectopic heartbeats, this affects the signal properties. In another example, these signal properties can therefore be identified via cranial acceleration features such as signal amplitude. Furthermore, ratios and statistical features, such as the heartbeat-by-heartbeat distortion of the aforementioned features, allow for the identification of arrhythmic and / or ectopic heartbeats. It is also possible to implement threshold values, which can be determined empirically, for example, to identify arrhythmic and / or ectopic heartbeats. Finally, arrhythmic and / or ectopic heartbeats can be identified and, in particular, annotated for or during subsequent procedural steps described herein.

[0028] In another example, the determination of usable signal data includes the generation of an indicator based on the quantity and / or quality of identified artifacts, which describes the usability of the captured acceleration signals, in particular the usability of the continuous sections and / or segments thereof, for processing to determine feature data, wherein acceleration signals, in particular continuous sections and / or segments thereof, are disregarded for the determination of feature data if the captured acceleration signals, in particular the usability of the continuous sections and / or segments thereof, are marked as unusable.

[0029] In other words, a tolerance range can be defined in which continuous sections of the acceleration signals and / or subdivided segments thereof are considered usable for further processing, even if artifacts have been discovered in them, as long as these artifacts are considered harmless for further processing, i.e., for the detection, identification, and evaluation of features within the acceleration signals.

[0030] As mentioned above, additional features can be considered that do not originate from acceleration signals but from other measurements taken on the patient or from other available data sources. For example, feature data acquisition can include the collection of condition data, particularly condition data manually entered by a user and relating to a patient's condition, especially a medical and / or pathological condition.Such characteristics can be captured, for example, via user input describing age, gender, height, weight, head circumference, pre-existing medical conditions, assessments according to clinically recognized assessment methods such as the National Institutes of Health Stroke Scale (NIHSS), Face-Arms-Speech-Time (FAST) tests, the Cincinnati Prehospital Stroke Scale test, or other officially recognized assessments for evaluating the severity of a suspected stroke.

[0031] In another example, the state data can describe at least one of the following characteristics: - a motor impairment of the patient, which is detected in particular via a handheld device held by the patient; - an impairment of the patient's vision or sight, which is detected in particular by a device that tracks at least one of the patient's eyes; - a characteristic of the patient's peripheral vascular system; - activity of the patient's brain or heart; - the patient's blood flow.

[0032] For example, each of the above features can be provided using additional equipment, such as: - a handheld device for measuring finger pressure / finger strength to determine parameters for characteristics related to motor deficits; - an eye-tracking device for monitoring gaze or vision disorders; - a device for measuring vascular characteristics, such as blood flow, blood velocity, or similar; - Devices for measuring physiological signals to monitor brain activity (e.g. electroencephalography (EEG) sensors), to monitor heart activity (e.g. electrocardiography (ECG) sensors) or to monitor blood flow (e.g. Doppler ultrasound sensors).

[0033] In another example, a statistical analysis of cranial acceleration signals can reveal correlations between features extracted from segmented acceleration signals and other factors such as the patient's age and heart rate. Based on this information, which can be obtained from one or more previously acquired datasets, the assignment and weighting of cranial acceleration features can be performed for or during subsequent procedural steps described herein.

[0034] In a second aspect, the invention relates to a computer program comprising instructions which, when executed by at least one computer, cause that computer to execute the method according to the first aspect. The invention may alternatively or additionally relate to a (physical, for example, electrical, for example, technically generated) signal wave, such as a digital signal wave or an electromagnetic carrier wave, carrying information that represents the program, for example, the aforementioned program, which, for example, includes code means designed to execute one or all steps of the method according to the first aspect. In one example, the signal wave is a data carrier signal carrying the aforementioned computer program.A computer program stored on a disc is a data file, and when the file is read and transmitted, it becomes a data stream, for example, in the form of a (physical, for example, electrical, for example, technically generated) signal. The signal can be implemented as a signal wave, for example, as the electromagnetic carrier wave described here. For example, the signal, for example, the signal wave, is structured such that it is transmitted via a computer network, for example, a LAN, WLAN, WAN, mobile network, for example, the Internet. For example, the signal, for example, the signal wave, is structured such that it is transmitted by optical or acoustic data transmission. The invention according to the second aspect can therefore alternatively or additionally relate to a data stream that is representative of the aforementioned program, i.e., that comprises the program.

[0035] In a third aspect, the invention relates to a computer-readable storage medium on which the program according to the second aspect is stored. The program storage medium is, for example, non-perishable.

[0036] In a fourth aspect, the invention relates to at least one computer (e.g., a computer) comprising at least one processor (e.g., a processor), wherein the program according to the second aspect is executed by the processor, or wherein the at least one computer comprises the computer-readable storage medium according to the third aspect.

[0037] In a fifth aspect, the invention relates to a medical system (for example, a system for measuring cranial acceleration) comprising the following: a) the at least one computer according to the fourth aspect; b) at least one electronic data storage device that stores at least the acceleration measurement data; and c) at least one accelerometer for receiving acceleration signals from an anatomical body part; d) for example, at least one heartbeat detector for receiving the patient's heartbeat signals; wherein the at least one computer is functionally coupled with - the at least one electronic data storage device to record at least the acceleration measurement data from the at least one data storage device, and - the accelerometer, in order to receive the patient's acceleration signals from the accelerometer and to generate acceleration measurement data from the acceleration signals, and - for example, the heartbeat detector, in order to receive the patient's heartbeat signals from the heartbeat detector and to generate heartbeat signal data from the heartbeat signals.

[0038] Alternatively or additionally, according to the fifth aspect, the invention relates to, for example, a non-transitory, computer-readable program storage medium that stores a program to cause the computer, according to the fourth aspect, to execute the data processing steps of the method according to the first aspect.

[0039] For example, the invention does not include an invasive step that would represent a significant physical impairment of the body, that would require professional medical expertise to perform, and that would pose a significant health risk even if performed with the necessary professional care and expertise. DEFINITIONS

[0040] This section provides definitions for certain terms used in this revelation, which are also part of the present revelation.

[0041] The method according to the invention is, for example, a computer-implemented method. For example, all steps or only some of the steps (i.e., fewer than the total number of steps) of the method according to the invention can be performed by a computer (for example, at least one computer). One embodiment of the computer-implemented method is the use of the computer to perform a data processing procedure. Another embodiment of the computer-implemented method is a method relating to the operation of the computer such that the computer is operated to perform one, several, or all steps of the method.

[0042] The computer comprises, for example, at least one processor and at least one memory component for processing the data (technically), for example, electronically and / or optically. The processor consists, for example, of a material or composition that is a semiconductor, for example, at least partially n- and / or p-doped semiconductors, for example, at least one of type II, III, IV, V, VI semiconductor material, for example, (doped) silicon and / or gallium arsenide. The described calculation or determination steps are performed, for example, by a computer. Determination or calculation steps are, for example, steps for determining data within the framework of the technical process, for example, within a program.

[0043] A computer is, for example, any type of data processing device, such as an electronic data processing device. A computer can be a device generally considered as such, such as desktop PCs, notebooks, netbooks, etc., but also any programmable device, such as a mobile phone or an embedded processor. A computer can, for example, comprise a system (network) of "subcomputers," with each subcomputer being an independent computer. The term "computer" includes a cloud computer, such as a cloud server. The term "computer" includes a server resource. The term "cloud computer" includes a cloud computer system, which, for example, comprises a system consisting of at least one cloud computer and, for example, a multitude of operationally interconnected cloud computers, such as a server farm.Such a cloud computer is preferably connected to a wide area network such as the World Wide Web (WWW) and is located in a so-called cloud of computers, all of which are connected to the World Wide Web. Such infrastructure is used for "cloud computing," which describes computing, software, data access, and storage services where the end user does not need to know the physical location and / or configuration of the computer providing a particular service. The term "cloud" is used in this context, for example, as a metaphor for the internet (World Wide Web). The cloud provides computing infrastructure as a service (IaaS). The cloud computer can act as a virtual host for an operating system and / or a data processing application used to execute the method of the invention.The cloud computer is, for example, an elastic computing cloud (EC2) such as that provided by Amazon Web Services™. A computer includes, for example, interfaces for receiving or outputting data and / or performing analog-to-digital conversion. The data can be, for example, data representing physical properties and / or generated from technical signals. These technical signals are generated, for example, by (technical) detection devices (such as devices for detecting marking devices) and / or (technical) analysis devices (such as devices for performing (medical) imaging procedures), where the technical signals are, for example, electrical or optical signals. The technical signals represent, for example, the data received or output by the computer.The computer is preferably functionally coupled with a display device that allows information output by the computer to be shown to a user, for example. An example of a display device is a virtual reality device or an augmented reality device (also called virtual reality glasses or augmented reality glasses), which can be used as "glasses" for navigation. A specific example of such augmented reality glasses is Google Glass (a trademark of Google, Inc.). An augmented reality device or a virtual reality device can be used both for inputting information into the computer through user interaction and for displaying information output by the computer.Another example of a display device would be a standard computer monitor, which includes, for example, a liquid crystal display (LCD) that is functionally connected to the computer to receive display control data from the computer in order to generate signals used to display image information on the display device. A specific embodiment of such a computer monitor is a digital lightbox. An example of such a digital lightbox is Buzz®, a product of Brainlab AG. The monitor can also be the display of a portable device, such as a handheld device like a smartphone, a personal digital assistant, or a digital media player.

[0044] The invention also relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to execute the method or methods, for example, the steps of the method or methods described herein, and / or a computer-readable storage medium (for example, a non-transitory computer-readable storage medium) on which the program is stored, and / or to a computer comprising the program storage medium, and / or to a (physical, for example, electrical, for example, technically generated) signal wave, for example, a digital signal wave, such as an electromagnetic carrier wave, which carries information representing the program, for example, the aforementioned program, which for example, comprises code means designed to execute one or all of the method steps described herein.In one example, the signal wave is a data carrier signal that carries the aforementioned computer program. The invention also relates to a computer comprising at least one processor and / or the aforementioned computer-readable storage medium and, for example, a memory, wherein the program is executed by the processor.

[0045] Within the scope of the invention, computer program elements can be embodied by hardware and / or software (including firmware, resident software, microcode, etc.). Within the scope of the invention, computer program elements can take the form of a computer program product, which can be embodied by a computer-usable, for example, computer-readable data storage medium comprising computer-usable, for example, computer-readable program instructions, "code," or a "computer program" embodied in the data storage medium for use on or in conjunction with the instruction execution system.Such a system can be a computer; a computer can be a data processing device comprising means for executing the computer program elements and / or the program according to the invention, for example, a data processing device comprising a digital processor (central processing unit or CPU) that executes the computer program elements and optionally volatile memory (for example, random access memory or RAM) for storing data used for and / or generated by the execution of the computer program elements. Within the scope of the present invention, a computer-usable, for example, computer-readable data storage medium can be any data storage medium that can contain, store, transmit, distribute, or transport the program for use on or in conjunction with the instruction execution system, the device, or the apparatus.The computer-usable, for example, computer-readable data storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or a distribution medium such as the internet. The computer-usable or computer-readable data storage medium could even be, for example, paper or another suitable medium on which the program is printed, since the program could be captured electronically, for example, by optically scanning the paper or other suitable medium, and then appropriately compiled, interpreted, or otherwise processed. The data storage medium is preferably a non-volatile data storage medium.The computer program product and each software and / or hardware described herein constitute the various means for performing the functions of the invention in the exemplary embodiments. The computer and / or data processing device may, for example, include a guidance information device comprising means for outputting guidance information. The guidance information may, for example, be output to a user visually by means of a visual display (e.g., a monitor and / or a lamp) and / or audibly by means of an audible display (e.g., a loudspeaker and / or a digital speech output device) and / or tactilely by means of a tactile display (e.g., a vibration element or a vibration element integrated into an instrument).For the purposes of this document, a computer is a technical computer that includes, for example, technical components, such as physical components, such as mechanical and / or electronic components. Any device mentioned as such in this document is a technical device and includes, for example, physical components.

[0046] The term "data acquisition" includes, for example (within the context of a computer-implemented method), the scenario in which the data is determined by the computer-implemented method or program. Determining data includes, for example, measuring physical quantities and converting the measured values ​​into data, such as digital data, and / or calculating (and, for example, outputting) the data using a computer, for example, within the context of the method according to the invention. A step of "determining," as described herein, includes or consists, for example, of issuing a command to execute the determination described herein. For example, the step includes or consists of issuing a command to cause a computer, such as a remote computer or a remote server, such as one in the cloud, to perform the determination.Alternatively or additionally, a step of "determining" as described herein includes, for example, receiving the data resulting from the determination described herein, for example, receiving the resulting data from the remote computer, such as the remote computer that was caused to perform the determination. The meaning of "acquiring data" also includes, for example, the scenario in which the data is received or retrieved (e.g., inputted) by the computer-implemented method or program, for example, from another program, a previous method step, or a data storage medium, for example, for further processing by the computer-implemented method or program. The generation of the data to be acquired may, but need not, be part of the method according to the invention.The term "data acquisition" can therefore also mean, for example, waiting to receive data and / or receiving the data. The received data can, for example, be entered via an interface. The term "data acquisition" can also mean that the computer-implemented method or program performs steps to (actively) receive or retrieve the data from a data source, such as a data storage medium (like ROM, RAM, a database, a hard disk, etc.), or via the interface (for example, from another computer or a network). The data acquired by the disclosed method or device can be acquired from a database located in a data storage device that is operational with a computer for data transfer between the database and the computer, for example, from the database to the computer.The computer collects the data for use as input in steps for determining data. The determined data can be output back to the same or a different database for later storage. The database or databases used to carry out the disclosed process can be located on a network data storage device or server (e.g., a cloud data storage device or server) or on a local data storage device (e.g., a mass storage device connected to at least one computer on which the disclosed process is executed). The data can be made "ready for use" by performing an additional step prior to the collection step. According to this additional step, the data is generated for the purpose of collection. For example, the data is collected or recorded (e.g., by an analysis device).Alternatively or additionally, the data is entered according to the additional step, for example via interfaces. The generated data can be entered, for example, into a computer. According to the additional step (which precedes the acquisition step), the data can also be made available by performing the additional step of storing the data in a data storage medium (such as a ROM, RAM, CD, and / or hard drive) so that it is ready for use within the method or program according to the invention. The step of "acquiring data" can therefore also include instructing a device to obtain and / or make available the data to be acquired.In particular, the data acquisition step does not include any invasive step that would involve significant physical impairment of the body, require professional medical expertise, and, even if performed with the necessary professional care and expertise, would pose a significant health risk. Specifically, the data acquisition step, such as determining data, does not include any surgical step, and in particular, does not include any step involving the treatment of a human or animal body by means of surgery or therapy. To distinguish the various data used in the present procedure, the data will be referred to as "XY data" and the like, and defined by the information that describes them, which will then preferably be referred to as "XY information" and the like. FIGURE DESCRIPTION

[0047] The invention is described below with reference to the accompanying figures, which contain background information and illustrate specific embodiments of the invention. However, the scope of the invention is not limited to the specific features disclosed in connection with the figures. Fig. 1. The basic steps of the procedure are illustrated according to the first aspect; Fig. 2 shows an embodiment of the present invention, in particular the method according to the first aspect; Fig. 3 and Fig. 4. Show a flowchart illustrating the process of removing artifacts from acceleration signals; Fig. 5 provides an overview of a data processing flow in which the method according to the present invention can be used; Fig. 6. A schematic representation of the system according to the fifth aspect; and Fig. Figure 7 shows an example of recorded acceleration measurements. DESCRIPTION OF SPECIAL EXECUTION FORMS

[0048] Fig. Figure 1 illustrates the basic steps of the procedure according to the first aspect, where step S11 includes the acquisition of the acceleration measurement data, step S12 includes the determination of the artifact data, step S13 includes the determination of the usable signal data, step 14 includes the determination of the feature data, step 15 includes the determination of the reference feature data, and step 16 includes the determination of the classification data.

[0049] Fig. Figure 2 illustrates an embodiment of the present invention which includes all the essential features of the invention. In this embodiment, all the data processing, which is part of the method according to the first aspect, is carried out by a computer 2. Reference numeral 1 denotes the input of data acquired by the method according to the first aspect into the computer 2, and reference numeral 3 denotes the output of data determined by the method according to the first aspect.

[0050] Fig. This section illustrates the process of performing a cranial acceleration measurement on a patient, the removal of artifacts from the acceleration signals, and the extraction and evaluation of features from the artifact-free acceleration signals. Determining the patient's physiological state begins with offline processing of the acceleration signals, i.e., processing previously acquired acceleration signals obtained from at least one, and preferably several, accelerometers in contact with the area of ​​interest, such as the skull. Each accelerometer is assigned an individual signal channel. The acquired acceleration signals are segmented into continuous segments of sufficient length—that is, of at least one predefined length—to provide enough data for the subsequent processing steps.Optionally, the continuous sections can be correlated and / or evaluated with corresponding respiratory cycle phases derived from heartbeat-related signals. Artifacts identified in the acceleration signals are then either removed from the acceleration signals or signal segments containing these artifacts are "cut out" and discarded, i.e., not considered in subsequent processing steps. The now artifact-free acceleration signal segments are then segmented / divided into a multitude of segments of a predefined length. In the example shown, the segment length essentially corresponds to a cardiac cycle. The segmentation can thus be based either on signal data describing the heartbeat or "directly" on the acceleration signals by detecting characteristic repeating waveforms within them.Segments of essentially equal length are then subjected to an additional search for artifacts. Segments containing detected artifacts are excluded from subsequent processing steps. As an optional step, segments containing one or more arrhythmic heartbeats, ectopic heartbeats, and premature ventricular contractions can also be removed. To search the remaining artifact-free portions of the acceleration signals for unique features, these segments are decomposed, i.e., converted from the time domain to the frequency domain, specifically into a) low frequencies (e.g., 0.5–20 Hz) and b) higher frequencies (e.g., 20 Hz to the Nyquist frequency). The decomposed signal segments are then subjected to a quality check to determine whether they are of sufficient quality for feature extraction and evaluation.If it is determined that a sufficient quantity of segmented and decomposed acceleration signal segments of sufficient quality is available, feature recognition and evaluation can begin (see . Fig. 4) During recording, the patient's positioning can be estimated, and parameters such as arterial stiffness, intracranial pressure (ICP), blood pressure, etc., can be extracted for feature expansion and feature evaluation / selection. Features are identified within the segmented and decomposed acceleration signals. Based on confounding factors such as the patient's age and heart rate, the cranial acceleration features can be identified and / or weighted as an optional step. Knowledge of certain properties of the features of interest aids in identifying such features within the (segmented and decomposed) acceleration signals. This can be achieved either through a statistical approach or a machine learning approach, particularly if a list of predictive features is provided for each channel. The extracted features are then classified by implementing one of the classification approaches described above.The features extracted from the acceleration signals, along with additional features obtained from other measurement sources and / or data sources, are enhanced before the patient is classified within the reference group. A report on the classification results is then generated.

[0051] Fig. This illustrates a complete workflow within the scope of the disclosed invention, encompassing steps performed before and after the data processing described above. After applying the headband / headset to the patient and performing a preliminary check and online preprocessing as described in PCT / EP2021 / 078408, artifacts are detected and removed from the acceleration measurement data. The acceleration signals can be segmented into a portion indicating a heartbeat signal and a portion indicating another acceleration-induced signal component, followed by signal decomposition by frequency and optionally feature extension. Subsequently, a suitably configured machine learning classifier or a statistical approach is used to determine the patient's physiological state based on the results of the preceding data processing.Optionally, a report on the results and, for example, the meta-information characterizing the measurement can then be generated.

[0052] Fig. Figure 6 is a schematic representation of the medical system 4 according to the fifth aspect. The system as a whole is identified by reference numeral 4 and comprises a computer 5, an electronic data storage device (such as a hard disk) 6 for storing at least the patient data, and a medical device 7 (such as a radiotherapy device). The components of the medical system 4 have the functionalities and properties described above in relation to the fifth aspect of this disclosure.

[0053] Fig.Figure 7 shows a raw acceleration signal acquired via an accelerometer in contact with the patient's skull. Occasionally, spike-like artifacts are observed in the acceleration signal, which can be detected using artifact detection techniques.

[0054] An additional explanation of the system within the scope of the present invention is given below.

[0055] The system comprises various subsystem components, including: - a headset that is in contact with the patient and includes the following: ◯ one or more photoplethysmography sensors (PPG) for recording the heartbeat, heart rate and time; o one or more sound pressure level (SPL) sensors for detecting ambient noise; and a large number of acceleration sensors to detect acceleration at specific points around the patient's head. - a data collector that digitizes the analog sensor signals (either as a separate component or integrated into the sensor elements); - a computer unit containing the device software and storage space for the recording data; and - Device software that provides a user interface, hardware control, software libraries and algorithms for signal conditioning, signal processing, signal separation and classification for a variety of clinical indications.

[0056] The system captures and stores sensor data, such as that generated by the pulsating blood flow from the cardiac cycle, which leads to a slight acceleration of the skull. The system uses, for example, piezoelectric accelerometers that measure various signal components resulting from the head's / brain's response to the blood flow.

[0057] The numerous accelerometers detect movement, and the data logger digitizes the signal. The computer unit provides the user interface, stores the data, performs signal separation, and classifies the recorded data according to clinical indication.

[0058] A user places the device's headset on the patient and configures the user interface to begin recording. Typically, the user performs a recording lasting approximately one to two minutes; however, in some cases, the recording may be longer if the patient moves or the headset shifts. The device software analyzes the recording and separates the recorded signal into its component parts. Predictive features are calculated for each component and then classified using statistical and / or machine learning methods based on known thresholds, data from known states, and other factors. The device software then displays the classification results to the user, categorizing them into defined clinical indications.

[0059] It is preferable that the recording be largely free of artifacts that could, for example, impair classification. Procedures for quality analysis of the signal before the start of recording and for continuous quality control during recording are disclosed.

[0060] The headset is attached to the test subject's head and the device is switched on. A recording is made, which is then evaluated using the following preprocessing approach: Artifact detection and / or removal After recording, the recorded data is preprocessed to remove sections containing artifacts from the recording or to remove artifacts from specific sections, thus restoring the original waveforms without artifacts. The artifacts can have the following causes: Acceleration signal quality checkBackground noise and sounds. Background noise and sounds (speech, beeps from vital signs monitors, the test subject's breathing, etc.) are detected by determining a similarity measure between the SPL sensor and the multitude of accelerometers (e.g., coherence analysis). The similarity measure with the SPL channel is determined individually for each accelerometer channel. Furthermore, the degree of similarity can also be achieved through signal decomposition using various decomposition methods (e.g., principal component analysis, independent component analysis, blind source separation, spectral subtraction, adaptive noise reduction, noise modeling via convolutional neural networks, etc.).If background noise is detected, the corresponding segments are either excluded from the analysis or the artifact-free signal waveform is restored by identifying and removing the artifact waveform in the time domain or the artifact energy through spectral analysis and subsequent spectral subtraction. In the latter case, the artifacts are removed and the segments are used for analysis. - Contact of the headset sensors with objects during measurement (e.g., pillows used to stabilize the test subject's head, headrests, etc.). Time-domain analysis can be used to detect short, spike-like artifacts. For this, a signal is decomposed using decomposition methods (e.g., Fourier decomposition). A description of a sampling range (e.g., a moving median determined by a window size) is then defined on the resulting signal. Both the decomposed signal and the description of the sampling range are compared sample by sample (e.g., by subtraction) and flagged as artifacts if a defined threshold is exceeded. Creating a subset of good segments - The above-mentioned preprocessing methods, alone or in combination with the results of the online preprocessing methods as described in PCT / EP2021 / 078408, which is hereby incorporated by reference, can be used to create a subset of the recorded sensor data that contains data that meets all quality-related criteria. Heartbeat segmentation - The segments without artifacts are then processed and further subdivided by identifying individual heartbeats. - The heartbeat signal is detected, for example, with a heartbeat sensor (e.g., a PPG sensor) or multiple heartbeat sensors. - The heartbeat can also be detected based on characteristic repeating waveforms within the acceleration signal. In some cases, heartbeats can be detected directly via the accelerometer. Frequency decomposition - The individual heartbeats are decomposed to isolate physiological components within the acceleration signal that reflect low, medium and high frequencies, using methods such as Fourier decomposition, forward-backward filtering and spectral subtraction. Feature extraction - Features from the time domain, the frequency domain and the statistical domain are extracted for each acceleration channel, for each frequency domain and for sensor combinations (to derive differential features). Feature augmentation - Additional features are derived from sensors other than the accelerometers, other devices, or user input, for example: Features are extracted, for example, from the heartbeat sensor or multiple heartbeat sensors. These features include risk factors for the respective disease, such as arterial stiffness, intracranial pressure, blood pressure, and atrial fibrillation. Characteristics can be provided to the software via user input (e.g., age, gender, pre-existing conditions, scores obtained through clinically recognized assessment methods such as NIHSS, FAST, Cincinnati Prehospital Stroke Scale, etc.). These scores are officially recognized values ​​used to assess the severity of a suspected stroke and are well-documented in the literature. They do not represent an actual measurement but are based on the subjective assessment of the patient by a medically trained individual. Features can be provided using additional devices, for example: Eye tracking, for example, can be used to check for gaze or vision disorders on one or both sides, providing an additional parameter that indicates the location of the patient's physiological impairment. • A device can be used to determine characteristics of the peripheral vascular system (e.g., blood flow, velocity). Measuring the vascular system can also allow for the normalization of blood flow to the brain of individual subjects, leading to a reduction in inter-subject variability. ◯ Devices for recording the patient's physiological signals in order to add further features for evaluation, e.g. brain activity (EEG), heart activity (ECG), blood flow (Doppler sonography). Classification using statistical methods - After extracting the characteristics, the characteristics of a particular person can be compared with a sample of control subjects and persons with the respective disease, e.g. using thresholds, confidence intervals, clustering or regression of the characteristics. Threshold-based methods determine a class (i.e., in a binary classification: class 0 or class 1) using a predefined threshold based on samples of the classes. Depending on whether the subject exceeds the threshold or not, they are assigned to the respective class. Confidence intervals are a statistical measure that describes the accuracy range of a given characteristic. In classification, for example, a test subject's characteristic might lie outside a predefined confidence interval for class 0 and therefore be assigned to class 1. Clusters are assigned to groups of specific features in 2D / 3D space. Within each cluster, a centroid ("center") is determined (e.g., using K-means clustering). Clusters can be used for classification by calculating which centroid is closer to a test object and assigning the object accordingly. Features can be described by regression (e.g., linear, sigmoidal). For binary classifications, a regression threshold can be used to assign each class. - Feature extension can be used to expand the feature space for classification. - Comparison of features extracted from the original, decomposed signal or signal relationships (e.g. ratios, differences), and comparison of these features using statistical methods. Classification using machine learning methods - After feature extraction, a pre-trained machine learning algorithm can be used for a specific recording. - The training set for the machine learning algorithm is a sufficiently large dataset that includes subjects with and without the respective disease. - To expand the feature space for classification, a feature extension is used.

[0061] The system includes, for example, the following devices: - a headset with: ◯ one or more PPG sensors; ◯ one or more SPL sensors; ◯ multiple acceleration sensors; ◯ a data collector; ◯ a computer unit; ◯ of device software; ◯ a device for determining the vascular characteristics of the test subject; ◯ a device for eye tracking of the test subject. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] US 2018 / 0296107

[0024] EP 2021 / 078408 [0051, 0060]

Claims

[1] Computer-implemented medical procedure for processing data acquired using accelerometers in contact with an anatomical part of a patient’s body, the procedure comprising the following steps: a) Acceleration measurement data are recorded (S11), describing acceleration signals that were recorded using the acceleration sensors; b) Artifact data are determined on the basis of the acceleration measurement data (S12), which describe the presence of one or more artifacts in the acceleration signals; c) Usable signal data are determined on the basis of the acceleration measurement data and the artifact data (S13), which describe continuous sections of the acceleration signals that are free of identified artifacts; d) Feature data are determined on the basis of the usable signal data (S14) which describe at least one predefined feature in the acceleration signals; e) Reference feature data are determined (S15) which describe at least one predefined feature in acceleration signals recorded for at least one reference person with a specific disease; f) Classification data are determined on the basis of the feature data and the reference feature data (S16) which describe a degree of similarity between the at least one predefined feature in the acceleration signals recorded for the patient and the at least one predefined feature in the acceleration signals recorded for the at least one reference patient. [2] Method according to claim 1, wherein the determination of usable signal data comprises the determination of segmented signal data based on the usable signal data which describe the continuous sections divided into a plurality of segments. [3] Method according to claim 2, wherein the continuous sections are divided into segments with a length of at least one cardiac cycle, in particular wherein the segments cover a substantially equal period of time. [4] Method according to one of claims 2 and 3, wherein the segments are searched for additional artifacts that are not described in the artifact data. [5] Method according to any one of claims 1 to 4, wherein the determination of usable signal data comprises generating an indicator that describes the usability of the detected acceleration signals, in particular the usability of the continuous sections and / or segments thereof, for processing to determine feature data, based on the quantity and / or quality of the identified artifacts, wherein acceleration signals, in particular continuous sections and / or segments thereof, are disregarded for the determination of feature data if the detected acceleration signals, in particular the usability of the continuous sections and / or segments thereof, are indicated as unusable. [6] Method according to any one of claims 1 to 5, wherein the determination of usable signal data comprises the acquisition of additional sensor data of a different type than the acceleration measurement data, which in particular describe at least one of the following elements: - an acoustic noise from the environment, including structure-borne sound and / or airborne sound; and - a patient's heart rate; whereby, at least partially based on the additional sensor data, the artifacts are identified and / or the continuous sections are divided into segments. [7] Method according to any one of claims 1 to 6, wherein the determination of feature data comprises the acquisition of condition data, in particular condition data manually entered by a user, relating to the patient and / or a condition of the patient, in particular a medical and / or pathological condition of the patient. [8] Method according to claim 7, wherein determining state data describes at least one of the following elements: - a motor impairment of the patient, in particular detected via a handheld device held by the patient; - a gaze or visual impairment of the patient, in particular detected by a device that tracks at least one of the patient's eyes; - a characteristic of the patient's peripheral vascular system; - activity of the patient's brain or heart; - the patient's blood flow. [9] Method according to any one of claims 1 to 8, wherein the determination of artifact data comprises decomposing the detected acceleration signals into several frequency ranges and analyzing the decomposed signals to identify artifacts. [10] Method according to any one of claims 1 to 9, wherein the determination of feature data comprises the determination of signal component data based on the usable signal data and / or the segmented signal data describing at least one of the continuous sections and / or at least one of the segments thereof, which are decomposed into a plurality of frequency ranges. [11] Method according to any one of claims 1 to 10, wherein determining classification data comprises defining one or more classes, each class comprising at least one reference subject, and defining at least one threshold value for the at least one predefined feature in the acceleration signals recorded for the patient, wherein an assignment of the patient to the one or more classes is based on whether the at least one predefined feature in the acceleration signals recorded for the patient exceeds the at least one threshold value or not. [12] Method according to any one of claims 1 to 11, wherein determining classification data comprises defining one or more clusters in a space with two or more dimensions, wherein the one or more clusters are defined by collections of predefined features in the acceleration signals recorded for a plurality of reference persons, wherein the assignment of the patient to the one or more clusters is based on the distance between the at least one predefined feature and the one or more clusters in the space with two or more dimensions. [13] Method according to any one of claims 1 to 12, wherein the determination of classification data comprises the use of a pre-trained machine learning algorithm, wherein a training set for the machine learning algorithm comprises a plurality of reference subjects with the specific disease and a plurality of reference subjects without the specific disease. [14] Computer program comprising instructions which, when the program is executed by a computer, cause the computer to execute the method according to any one of claims 1 to 13; and / or a computer-readable storage medium on which the program is stored; and / or a computer comprising at least one processor and / or the program storage medium, wherein the program is executed by the processor; and / or a data carrier signal that carries the program; and / or a data stream that includes the program. [15] Medical system (4), comprising: a) the at least one computer (5) according to claim 14; b) at least one electronic data storage device (6) which stores at least the reference feature data; and (c) a medical device (7) for recording the acceleration measurement data of a patient, including acceleration sensors that are in contact with an anatomical part of the patient's body, wherein at least one computer is functionally coupled with - the at least one electronic data storage device to capture at least the reference feature data from the at least one data storage device, and - the medical device, in order to capture at least the acceleration measurement data from the medical device.

Images