WO2017045976A1

WO2017045976A1 - Device and method for migraine monitoring

Info

Publication number: WO2017045976A1
Application number: PCT/EP2016/071001
Authority: WO
Inventors: Maria Estrella Mena Benito; Ihor Olehovych Kirenko
Original assignee: Koninklijke Philips N.V.
Priority date: 2015-09-18
Filing date: 2016-09-07
Publication date: 2017-03-23

Abstract

The present disclosure relates to a migraine monitoring device (10) and to a migraine monitoring method. The device (10)comprises an imaging unit (16) arranged to observe a subject(34) and to generate image data representing blood perfusion at a skin area of the subject (34), a detector unit (22) arranged to detect a characteristic spatial blood perfusion discrepancy between at least two regions (44, 46) of the skin area, a data conditioning unit (20) arranged to condition the image data, taking into account reference image data obtained from at least one reference region (48, 50) of the subject (34),wherein the detected blood perfusion discrepancy is indicative of a migraine-related symptom, and wherein the reference image data is used as a reference for compensating non-migraine- related influences.The present disclosure further relates to a corresponding computer program.

Description

Device and method for migraine monitoring

FIELD OF THE INVENTION

The present disclosure relates to a migraine monitoring device and to a migraine monitoring method. More particularly, the present disclosure relates to refinements in such devices and methods. In a general context, the present disclosure relates to unobtrusive monitoring methods, devices and systems for migraine monitoring, migraine detection, migraine assessment, migraine documentation, and migraine prediction.

The present disclosure further relates to a corresponding computer program implementing the above migraine monitoring method. BACKGROUND OF THE INVENTION

US 5,840,018 A discloses a method and a system for real-time determination of variations in effective diameter of cranial blood vessels, to provide an indication of migraine activity, wherein the method comprises determining the blood flow rate to the brain of a subject; determining the intracranial blood flow rate in selected blood vessels; and comparing said intracranial blood flow rate with said determined blood flow rate to the brain thereby to determine a change in said intracranial blood flow rate relative to said blood flow rate to the brain, indicating a corresponding change in the effective diameter of the preselected blood vessel.

Teplov V. et al. (2014-08-22) "Ambiguity of mapping the relative phase of blood pulsations" in BIOMEDICAL OPTICS EXPRESS, vol. 5, no. 9, page 3123, ISSN: 2156-7085, doi: 10.1364/boe.5.003123 relates to blood pulsation imaging (BPI) methods. BPI is a non- invasive optical method based on photoplethysmography (PPG) which may be used for the visualization of changes in the spatial distribution of blood in the microvascular bed. BPI specifically allows measurements of the relative phase of blood pulsations and using it we detected a novel type of PPG fast waveforms, which were observable in limited areas with asynchronous regional blood supply. Migraine detection is mentioned as a potential field of application.

Migraine is a medical disorder which usually causes a pounding, throbbing headache on one side of the head. The pain may be very bad and hurt so much that a person may have a hard time doing anything. Most migraines cause a headache and nausea and might make the person dizzy or very sensitive to bright lights or loud noises. Some migraines are preceded or accompanied by sensory warning symptoms (aura), such as flashes of light, blind spots, or tingling in an arm or leg. Other senses can change before or during a migraine attack, and the migraineur may sense strange smells or tastes. Migraines can last a long time. Most migraines only last about four hours. Some can last up to 72 hours.

According to the World Health Organization (WHO), migraine headache is a costly neurological disease in well-developed countries such as in the European Union and the United States. Migraine may be classified, e.g. based on how often attacks happen in a month: If a person suffers from migraine symptoms for less than fifteen days, the migraine is called episodic migraine. If it happens more than fifteen days, it is called chronic migraine. Chronic means it happens over a long amount of time. Some people who start their migraine experiences with getting episodic migraines may start to get chronic migraines later. Even nowadays, exact causes for migraine cannot be fully explained.

In everyday life of people experiencing migraine (so-called migraineurs), migraine diagnosis, particularly the prediction and/or detection of migraine attacks is a huge and crucial challenge. For migraine diagnosis, so-called migraine diaries are commonly used, wherein the migraineur basically has to keep a diary so as to trace the course of migraine attacks and non-migraine periods over a sufficiently long time so as to allow conclusions as to the nature and possible patterns of the recurrent migraine attacks.

However, some people may suffer from migraine which is characterized by symptoms that include more than just the typically to-be-expected headache. This may particularly apply to younger people, children, etc.

In recent times, several approaches to migraine detection have been proposed. For instance, Zaproudina N. et al. (2013) "Asynchronicity of Facial Blood Perfusion in

Migraine". PLoS ONE 8(12): e80189. doi: 10.1371/journal.pone.0080189, relates to insights in migraine detection. Accordingly, asymmetrical changes in blood perfusion and

asynchronous blood supply to head tissues likely contribute to migraine pathophysiology. In this context, mapping of blood pulsations in the face of migraineurs is proposed, using an imaging techniques so as to establish 2D-imaging of blood pulsation parameters. Further, a characteristic phenomenon of transverse waves of facial blood perfusion in migraineurs in contrast to healthy subjects who showed synchronous blood delivery to both sides of the face is described. Moreover, in accordance with this reference, amplitude of blood pulsations was observed to be symmetrically distributed over the face of healthy subjects, but asymmetrically in migraineurs and subjects with a family history of migraine. In migraine patients, a remarkable correlation between the side of unilateral headache and the direction of the blood perfusion wave may be detected. Accordingly, it is suggested that migraine is associated with lateralization of blood perfusion and asynchronous blood pulsations in the facial area, which could be due to essential dysfunction of the autonomic vascular control in the face.

However, also these new approaches are prone to failures and therefore often lack reliability. One reason for this is that optical imaging - as a matter of fact - also captures potentially disturbing and distorting artifacts which are somewhat inherent in the detected signals. One approach to this problem is to provide stable conditions which may be also referred to as laboratory conditions. This may involve a well-defined illumination, a defined and fixed position of the monitored subject, etc. However, these laboratory conditions cannot be ensured in everyday life.

Hence, there is still a need for improvement in migraine monitoring and detection involving alternative devices, systems and methods for migraine monitoring and diagnosis. Preferably, these devices and methods are applicable to everyday life

environments and particularly suited for outpatient monitoring or home monitoring.

SUMMARY OF THE INVENTION

In view of the above, it is an object of the present disclosure to seek for improved or alternative approaches to migraine monitoring and migraine diagnosis which may be implemented in respective methods and devices. More particularly, it is an object of the present disclosure to provide a migraine monitoring device and a corresponding method that enable increased migraine detection reliability and that shows improved performance in everyday life environments. Further, it would be advantageous to provide an enhanced migraine monitoring device and method that enable to utilize standard (off-the-shelf) monitoring equipment, for instance standard illumination sources, common cameras, including webcams, consumer digital cameras, and mobile phone/tablet cameras, etc.

Consequently, it is an object of the present disclosure to present robust methods and devices for migraine monitoring which facilitate migraine detection and prediction even in disturbance-affected environments that suffer from non- laboratory conditions. It is a further object of the present disclosure to present devices and methods that facilitate long-term migraine monitoring and that may simplify keeping a migraine diary for tracing the course of the migraine disease.

In a first aspect of the present invention a migraine monitoring device is presented, the device comprising:

an imaging unit arranged to observe a subject, and to generate image data representing blood perfusion at a skin area of the subject,

a detector unit arranged to detect a characteristic spatial blood perfusion discrepancy between at least two regions of the skin area, and

- a data conditioning unit arranged to condition the image data, taking into account reference image data obtained from at least one reference region of the subject,

wherein the data conditioning unit is arranged to process a plurality of regions, wherein the regions comprise a first indicative region, at least one second indicative region, and at least one of a ballistocardiographic motion reference region and a blood perfusion reference region,

wherein the detected blood perfusion discrepancy is indicative of a migraine- related symptom, and

wherein the reference image data is used as a reference for compensating non- migraine-related influences.

This aspect is based on the insight that migraine detection based on blood perfusion monitoring, as described in the above Zaproudina N. et al, is still prone to disturbances and distortions which are inherent to everyday life environments. Therefore, it is proposed in accordance with the above aspect to utilize reference data obtained from at least one reference region may provide a baseline which enables to attenuate or even discard non- migraine related signal components, at least to a great extent. Accordingly, several regions may be defined and monitored, at least over a certain amount of time. At least some of these regions are in a (skin) portion of the subject of interest which is typically affected by migraine. However, at least one additional reference region is provided which may represent a (skin) portion of the subject which is generally not affected by migraine or which may be otherwise used as a reference marker, for instance for motion detection which may involve also the detection of vascular motion attributable to circulatory processes.

In a more general context, a main aspect of the present disclosure relates to a beneficial combination of non-obtrusive monitoring and detected characteristics, particularly spatial signal discrepancies in the blood perfusion of migraine patients, and to refinements thereof. As used herein, non-obtrusive or unobtrusive monitoring relates to non-contact monitoring, particularly to remote unobtrusive monitoring of patients and further subjects of interest. Accordingly, as used herein, unobtrusive non-contact monitoring approaches may dispense with the need of applying contact sensors and similar equipment which may be experienced by the monitored subject as being unpleasant.

In this respect, reference is made to a method to measure skin color variations (indicative of changes in volume of blood vessels), called photoplethysmographic imaging (PPG) which is described in Wim Verkruysse et al., "Remote plethysmographic imaging using ambient light", Optics Express, Vol. 16, No. 26, December 2008. It is based on the principle that temporal variations in blood volume in the skin lead to variations in light absorptions by the skin. Such variations can be registered by a video camera that takes images of a skin area, e.g. the face, further processing includes calculating the pixel average over a manually selected region (typically part of the cheek in this system).

By looking at periodic variations of this averaged signal, the heart beat rate and respiratory rate can be extracted. The pulsation of arterial blood causes changes in light absorption. Those changes are observed with a photodetector (or an array of photodetectors) form a PPG (photoplethysmography) signal. Pulsation of the blood is caused by the beating heart, i.e. peaks in the PPG signal correspond to the individual beats of the heart.

Further, reference in this context is also made to WO 2012/098289 Al disclosing a method for visualization of cardiovascular pulsation waves, in which method the living body is illuminated with light penetrating through the skin of the body for interacting via absorption and/or scattering with a vascular system of the living body, wherein the light reflected from the living body is collected in the form of a focused frame into an image capturing device, and a series of frames is captured by the image capturing unit.

As used herein, the spatial blood perfusion discrepancy may relate to an at least slight discrepancy between the blood pulsation behavior at different skin regions of the subject. Typically, blood pulsation is reflected in at least minute fluctuations of the skin color of the subject of interest and may be therefore detected by appropriate imaging units which may, as an example, include CCD-sensors, CMOS-sensors, etc. Hence, by capturing and tracing these slight color fluctuations/variations, an at least partially periodic phenomenon may be detected, namely the blood pulsation waveform. In migraine patients, this waveform somewhat varies at different regions, when a migraine attack occurs, or is at least impending. By contrast, in healthy subjects, basically the same "comparable" regions would exhibit a basically equal and congruent blood pulsation waveform. Needless to say, blood pulsation is propelled by the activity of the heart. Therefore, some discrepancy may be already present between vessels that are close to the heart and vessels that are further away from the heart. However, this discrepancy is generally not indicative of migraine. Further, this effect may be corrected for by simply choosing similar regions that are arranged at a more or less equal distance from the heart. Further, also a corrective calculation or analysis may be applied when the regions of interest are not arranged at the same "effective" distance from the heart.

In one exemplary embodiment, the data conditioning unit is arranged to discard at least one of ballistocardiographic motion artifacts and vasoconstriction- indicative artifacts in the image data, the artifacts being non-indicative of the migraine-related symptom. Hence, the conditioning unit may be operated or trained to detect non- indicative patterns by comparing data obtained from at least one reference region and data obtained from at least one indicative (migraine affected) regions. For instance, ballistocardiographic motion may be a main reason for a reduced performance of migraine monitoring devices. Further,

vasoconstriction may impede the quality and accuracy of migraine detection. Typically, vasoconstriction may be reflected by local drops in an amplitude of the waveform detected signal. It has been observed that vasoconstriction, particularly local differences of

vasoconstriction, may be indicative of migraine. However, on the other hand, it has been observed as well that vasoconstriction may be also caused by other reasons or conditions than migraine-related ones. Hence, also a reference region for vasoconstriction may be used so as to discriminate non-indicative components.

Ballistocardiographic motion is caused by the heartbeat as such and reflected in minute pulsations of tissue portions that are provided with blood vessels.

Cardioballistographic motion may cause a non-indicative temporal discrepancy between signals of indicative regions which is, however, not indicative of migraine. Hence, a respective reference region may provide a "baseline" or reference signal so as to attenuate or discard the non- indicative ballistocardiography-related signal components.

The data conditioning unit is arranged to process a plurality of regions, wherein the regions comprise a first indicative region, at least one second indicative region, and at least one of a motion reference region and a blood perfusion reference region. The regions may be separate from one another. However, at least some of the regions may at least partially overlap one another. Preferably, at least the two indicative regions are separate and offset from one another, such that a respective comparison reveals the desired signal discrepancy. In another exemplary embodiment, the device further comprises a region selecting unit arranged to define at least three regions comprising the at least one reference region and the at least two indicative regions in the image data obtained from the imaging unit. Needless to say, the region selecting unit may be also arranged for (overall) motion compensation, i.e. for macroscopic motion compensation. For instance, the region selecting unit may be arranged for skin detection, head detection, nose detection, ear detection, eye detection, mouth detection, chin detection, etc. Preferably, the region selecting unit is arranged to automatically select and track the at least three regions without the need of human intervention, provided that the subject of interest is still in sight and in an appropriate orientation.

However, in alternative embodiments, the region selecting unit may be arranged for manual intervention, for instance for an initial manual selection of the at least three regions.

In another exemplary embodiment, the region selecting unit is arragned to define a first region of interest that is arranged as a first indicative region in a facial portion of the subject, a second region of interest that is arranged as a second indicative region in a facial portion of the subject, wherein the second region of interest is spaced apart from the first indicative region, a third region of interest as a motion reference region that is arranged as a ballistocardiographic motion reference region, and a fourth region of interest that is arranged as a blood perfusion reference region in a non-facial skin portion of the subject.

In still another exemplary embodiment, the region selecting unit may be arranged to discriminate a facial (skin) portion and a non-facial skin portion of the subject of interest, and a transitional portion, wherein the transitional portion comprises a skin-to- background transition. For instance, the transitional portion may comprise a face-to- background transition or a cheek-to-background transition. Generally, a present or impending migraine attack may be reflected in a characteristic blood perfusion/pulsation behavior at the head or face of the subject of interest. Further, also the characteristic discrepancy that is highly- indicative of migraine is typically present between distinct portions of the face or head of the subject of interest. Further, a non-facial or non-head skin portion of the subject is therefore generally not afflicted with migraine-related blood perfusion variations. Hence, the non- facial skin portion is a proper reference region for detecting non-migraine related artifacts. The transitional portion may be used to detect (overall) macroscopic motion of the subject of interest, but also ballistocardiographic (microscopic) motion at the skin of the subject as such which may involve minute recurring (periodic) extensions and contractions. Hence, motion of the skin with respect to the background may be detected in the transitional portion and may be utilized as a further reference for migraine detection.

In yet another embodiment of the device, the imaging unit is a remote monitoring imaging minute, particularly a remote photoplethysmographic monitoring imaging unit or a remote laser speckle imaging unit, wherein the detector unit is arranged to detect color fluctuations in the obtained image data that are indicative of blood perfusion. The color fluctuations may be attributed to circulatory processes in the subject of interest.

Laser speckle imaging (LSI), also referred to as Laser Speckle Contrast Imaging (LSCI), is an alternative unobtrusive method for monitoring blood perfusion in a subject of interest. In this connection, reference is made to Lisa M. Richards et al., "Low-cost laser speckle contrast imaging of blood flow using a webcam", Biomed Opt Express, 4(10): 2269-2283 (2013), describing new approaches to laser speckle contrast imaging utilizing standard (off-the-shelf) monitoring equipment.

Preferably, the imaging unit is arranged to detect visible radiation. As used herein, visible radiation may be referred to as the portion of the electromagnetic spectrum that is visible to a human eye (e.g., a wavelength portion between 380 nm and 780 nm (nanometers)). Hence, standard imaging units may be utilized. Further, in alternative embodiments, the imaging unit may be arranged to detect non- visible radiation, for instance infrared radiation, near- infrared radiation, or ultraviolet radiation. To this end, the imaging unit may be combined with appropriate radiation sources. Non-visible radiation may be utilized when the device is at least temporarily used for monitoring resting or sleeping patients, for instance.

As used herein, the remote monitoring imaging unit may be arranged as a non- contact and a non-obtrusive unit. Generally, the remote monitoring imaging unit may be arranged at a distance from the to-be-observed subject. This may on the one hand reduce any discomfort experienced by the user. On the other hand, remote imaging is susceptible to external influences which may cause relatively huge distortions which need to be

compensated.

In a refinement of the above embodiment, the detector unit is arranged to detect at least one of a spatial discrepancy between amplitudes of photoplethysmographic signals obtained from the observed region and a spatial discrepancy between phases of photoplethysmographic signals obtained from the observed regions. Both characteristics may be utilized for migraine detection. Preferably, the detector unit is arranged to detect both amplitude discrepancy and phase discrepancy so as to further improve the reliability and accuracy.

In yet another refinement of the above embodiment, the detector unit is arranged to detect a migraine-related vasoconstriction-indicative signal in an observed region, particularly in a nasal region. This may even further improve the validity and reliability of the migraine detection. In a photoplethysmographic waveform, vasoconstriction may be reflected by local drops in the amplitudes.

In yet another embodiment of the device, the detector unit is arranged to detect a characteristic spatial blood perfusion discrepancy between a right facial side of the subject, and a left facial side of the subject. In a healthy subject, typically no migraine-indicative discrepancy between the right facial side and the left facial side of the subject can be detected. In a subject suffering from migraine, typically also no discrepancy can be detected when no migraine attack is present or, at least, impending. However, in case of a forthcoming, imminent or even a present migraine attack, a characteristic signal discrepancy between the right facial side and the left facial side may be detected. An evolvement of the discrepancy may be monitored and traced over time which may be used for drawing further conclusions therefrom. For instance, the effect of medication against migraine may be observed in this way. Further, it has been found out that even when no (macroscopic) migraine symptoms (headache, etc.) are present, already at least a slight signal discrepancy between the right side and the left side of the subject can be detected which may be used to reliably predict an imminent migraine attack.

In accordance with the above embodiment, the first reference region may be arranged in the right or the left facial side of the subject, and the second referenced region may be arranged in the opposite one.

In a refinement of the above embodiment, the detector unit is further arranged to detect at least one of an amplitude asymmetry and an asynchronicity of facial blood perfusion between different observed facial regions of the subject. This may involve the detection of an amplitude asymmetry (non-symmetry) and an asynchronicity (non- synchronicity) of the PPG waveform (photoplethysmographic waveform) obtained from the different observed regions.

In another exemplary embodiment of the device, an analyzing unit is provided which is arranged to assess an actual migraine status based on at least one of model-based data and empirically obtained data. In this way, the device may alert the subject of an imminent or an actual migraine attack. Further, the device may be capable of monitoring, assessing and evaluating the effect of migraine medication. Needless to say, the actual migraine status may be assessed based on a combination of model-based data and empirically obtained data. Empirically obtained data may be obtained from the subject itself. Further, empirically obtained data may be obtained from reference subjects, e.g. representative healthy and/or migraine-suffering subjects.

Further, the analyzing unit may be arranged to record or log the migraine status of the subject over a long period. In this way, the device may basically automatically keep the subject's migraine diary. This is particularly beneficial for children and further persons which are not capable of handling the diary by themselves.

An automatic or semi-automatic (computer-aided or device-aided) assessment of the actual migraine status may have the further advantage that also in cases where no typical migraine symptom is present (e.g., no typical migraine-related headache), an assessment as to whether or not a migraine state is present can be made.

In another aspect of the present disclosure, a migraine monitoring method is presented, the method comprising the following steps:

observing a subject, including generating image data representing blood perfusion at a skin area of the subject,

conditioning the image data, taking into account reference image data obtained from at least one reference region of the subject,

wherein the step of conditioning involves processing a plurality of regions, wherein the regions comprise a first indicative region, at least one second indicative region, and at least one of a ballistocardiographic motion reference region and a blood perfusion reference region, and

detecting a characteristic spatial blood perfusion discrepancy between at least two regions of the skin area,

Preferably, a migraine monitoring device in accordance with the present disclosure is operated in accordance with the above method.

In an exemplary embodiment of the above method, the step of conditioning the image data comprises: detecting at least one of ballistocardiographic motion artifacts and

vasoconstriction-indicative artifacts in the image data, wherein the detected artifacts are non- indicative of the migraine-related symptom, and

attenuating the non- indicative artifacts in the image data.

As used herein, the non-indicative artifacts would have an influence on the detected signal (waveform), but are not attributable to migraine symptoms.

In another exemplary embodiment of the migraine monitoring method, the method further comprises:

defining a first region of interest, particularly a first indicative region in a facial portion of the subject,

defining a second region of interest, particularly a second indicative region in a facial portion of the subject, that is spaced apart from the first indicative region,

defining a third region of interest as a motion reference region, particularly a ballistocardiographic motion reference region,

- defining a fourth region of interest, particularly a blood perfusion reference region in a non-facial skin portion of the subject,

extracting a first photoplethysmographic signal from the first region of interest, extracting a second photoplethysmographic signal from the second region of interest,

- detecting a migraine-indicative phase discrepancy, particularly a spatial phase asymmetry between the first photoplethysmographic signal and the second

photoplethysmographic signal, under consideration of reference data obtained from the third region of interest, and

detecting a migraine-indicative signal shape or amplitude discrepancy, particularly a spatial shape or amplitude asymmetry based on the first photoplethysmographic signal and the second photoplethysmographic signal, under consideration of a reference photoplethysmographic signal obtained from the fourth region of interest.

In yet another aspect of the present invention there is provided a computer program which comprises program code means for causing a computer to perform the steps of the method as discussed herein when said computer program is carried out on that computer. Preferably, the computer is coupled with an arrangement in accordance with at least one embodiment of the monitoring device as discussed herein when the program is executed. The program code (or: logic) can be encoded in one or more non-transitory, tangible media for execution by a computing machine, such as a computer. In some exemplary embodiments, the program code may be downloaded over a network to a persistent memory unit or storage from another device or data processing system through computer readable signal media for use within the system. For instance, program code stored in a computer readable memory unit or storage medium in a server data processing system may be downloaded over a network from the server to the system. The data processing device providing program code may be a server computer, a client computer, or some other device capable of storing and transmitting program code.

As used herein, the term "computer" may stand for a large variety of processing devices. In other words, also mobile devices having a considerable computing capacity can be referred to as computing devices, even though they provide less processing power resources than standard "computers". Needless to say, such a "computer" can be part of a medical device and/or system. Furthermore, the term "computer" may also refer to a distributed computing device which may involve or make use of computing capacity provided in a cloud environment. The term "computer" may also relate to medical technology devices, fitness equipment devices, and monitoring devices in general, that are capable of processing data.

Preferred embodiments of the disclosure are defined in the dependent claims. It should be understood that the claimed method and the claimed computer program can have similar preferred embodiments as the claimed device and as defined in the dependent device claims.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. In the following drawings

Fig. 1 shows a schematic block diagram of an exemplary embodiment of a migraine monitoring device in accordance with the present disclosure;

Fig. 2 shows a simplified schematic representation of an image frame representing an observed subject of interest;

Fig. 3 shows an illustrative signal chart representing an exemplary photoplethysmographic signal (PPG signal); Fig. 4 shows an illustrative signal chart representing a first and a second photoplethysmographic signal, wherein an amplitude discrepancy between the first and the second signal is present;

Fig. 5 shows an illustrative signal chart representing a first and a second photoplethysmographic signal, wherein a phase discrepancy between the first and the second signal is present;

Fig. 6 shows an illustrative block diagram representing several steps of an exemplary embodiment of a migraine monitoring method in accordance with the present disclosure; and

Fig. 7 shows another illustrative block diagram representing several steps of another exemplary embodiment of a migraine monitoring method in accordance with the present disclosure.

DETAILED DESCRIPTION OF THE INVENTION

In the following, main aspects and insights of the disclosure will be illustrated and further elucidated with reference to exemplary embodiments of migraine

monitoring/migraine detection methods and devices. It should be noted that the following exemplary embodiments and descriptions shall not be interpreted in a limiting sense. Rather, the skilled person may readily transfer and broaden respective specific embodiments and the components as well as process steps disclosed herein so as to arrive at the general concept of the present disclosure.

As simplified block representation of a migraine monitoring or migraine detecting device 10 in accordance with at least some embodiments of the present disclosure is shown in Fig. 1. Hereinafter, the device 10 shall be understood as a migraine monitoring device. However, the device 10 may be also arranged as a migraine tracking, migraine detecting, migraine predicting device, or a combination thereof. Generally, the device 10 may comprise a distributed structure, i.e. at least some components thereof are arranged remote or separate from other components. In alternative embodiments, the device 10 may comprise an integral or integrated arrangement. In this context, mobile devices, such as mobile phones, tablet computers, mobile computers, etc. are referred to. The device 10 comprises an observing module 12 and a processing module 14. The observing module 12 comprises at least one imaging unit 16 which may be arranged as a camera unit comprising at least one imaging sensor, such as a CCD-sensor or a CMOS-sensor. Hence, also on-board cameras provided in mobile computing devices, mobile phones, etc. may be utilized. However, at least in some embodiments, separate imaging units 16 may be utilized, for instance separate cameras such as so-called webcams, action cams, etc.

Further, the imaging unit 16 may be arranged as a visible light imaging unit, a non- visible light imaging unit, or a combination thereof. As used herein, the term visible light shall refer to the range of electromagnetic radiation that is visible to the human eye. The term non- visible radiation or light shall refer to portions or ranges of the electromagnetic spectrum that are not visible to the human eye (i.e. infrared radiation, ultraviolet radiation, etc.). In any such case, the observing module 12 may be provided with or coupled to appropriate radiation sources, such as visible light sources and/or non-visible light sources. Needless to say, also already present ambient light sources may be utilized, since the device 10 is generally arranged as a non-obtrusive, non-contact remote detection device.

The processing modules 14 may arranged as an integrated processing module or a distributed processing module. I.e., at least some of the components thereof may be embodied in separate hardware. However, also at least some of the components of the processing module 14 may be embodied by or may share the same common hardware units. Further, one or more components of the processing module 14 may be implemented in discrete hardware. However, also off-the-shelf computing and processing devices may be utilized, such as standard computing environments (including desktop computers, mobile computers, mobile phones, server-client environments, distributed computing environments, cloud computing environments, etc.). Hence, components of the processing module 14 may be arranged as virtual components defined by appropriate software code.

The processing module 14 in accordance with the embodiment shown in Fig. 1 comprises a region selecting unit 18, a data conditioning unit 20, a detector unit 22, and an analyzing unit 24. Further, an input interface 26 and an output interface 28 may be provided. Via the input 26, image data obtained by the observing module 12 may be transmitted to the processing module 14. Typically, the image data comprises an image stream representing at least part of a subject of interest.

Further, the processing module 14 may be provided with or coupled to a memory unit 30. As with the embodiment shown in Fig. 1, the memory unit 30 is arranged as a separate memory unit which is coupled to at least the analyzing unit 24. Needless to say, also the region selecting unit 18, the data conditioning unit 20, and the detector unit 22 may be provided with or coupled to the same or a separate memory unit 30.

Additional reference is made to Fig. 2., showing an exemplary image frame 32 provided in the image data captured by the observing module 12, particularly the at least one imaging unit 16 thereof. In the exemplary representation of Fig. 2, the image frame 32 represents a forehead portion of an observed subject 34. Particularly, the image frame 32 may contain a head portion 36 including a facial portion 38 of the subject 34. Further, for reference purposes, as will be further explained hereinafter, also at least one non- facial portion 40 may be provided in the image frame 32. Preferably, also a so-called transition portion 42 is provided. The transition portion 42 includes a skin-to-background or face-to- background transition, and is preferably a high-contrast portion (i.e., a relatively sharp edge between the subject 34 and the background portion may be detected).

Further, several regions of interest that may be processed by the processing module 14 are indicated in Fig. 2 by dashed blocks. The regions of interest may cover a basically rectangular portion, but may also cover non-rectangular or even freeform portions, depending on the actually utilized region selecting algorithm or procedure.

In Fig. 2, a first region of interest is indicated by reference numeral 44. A second region of interest is indicated by reference numeral 46. A third region of interest is indicated by reference numeral 48. A fourth region of interest is indicated by reference numeral 50. A nasal region of interest is indicated by reference numeral 52.

The first region 44 may be also referred to as first indicative region. The second region 46 may be also referred to as second indicative region. The first region 44 and the second region 46 are arranged on respective sides of the face of the subject 44. In other words, one of the first and second region 44, 46 is arranged the right side (with respect to the vertical axis of symmetry of the face), whereas the other one is arranged in the left side. The first region 44 and the second region 46 are selected since in migraine-experiencing patients a characteristic discrepancy between blood-perfusion signals of the respective regions 44, 46 may be detected.

For illustrative reasons, in one embodiment, the first region 44 may be referred to as right side region, and the second region 46 may be referred to as left side region (as seen from the perspective of the subject 34).

The supplemental or auxiliary regions 48, 50, 42 may be also referred to as reference regions. Region 48 covers at least part of the transition portion 42. Hence, motion reference data may be obtained since macroscopic and/or microscopic motion of the subject 34 with respect to a steady motion reference (background, camera-based reference, etc.) can be detected. Particularly, so-called ballistocardiographic motion which is attributable to blood pulsation in blood tissues may be detected in this way and, so-to-say, removed from the signals obtained from the indicative regions 44, 46 so as to attenuate or level off these non- migraine-related influences.

The fourth region 50 basically comprises a non-facial skin portion of the subject 34, for instance a neck portion. From the fourth region 50, a signal reference, particularly a photoplethysmographic reference signal may be obtained. Typically, the respective blood-perfusion or blood-pulsation indicative signals obtained from the fourth region 50 are not affected by migraine. Hence, a proper signal reference is provided which enables to detect and attenuate or level off non-migraine related signal patterns in blood- pulsation indicative data.

Further, at least in some embodiments, also the nasal region 52 may be defined and utilized which covers at least a portion of the nose of the subject 34. It has been observed that at the nose or adjacent to the nose, so-called vasoconstriction-related signal patterns may be detected which may be caused by non-migraine -related reasons and which may therefore distort the obtained signals and mitigate the detection accuracy. Hence, proper correction may further improve the reliability of migraine detection.

Reference is again made to Fig. 1. The region selecting unit 18 of the processing module 14 may be arranged for automatically defining and selecting appropriate regions 44, 46, 48, 50, 52. To this end, the region selecting unit 18 may utilize and implement skin-detection algorithms, face-detection algorithms, edge-detection algorithms, and further pattern-detection algorithms. At least in some embodiments, the region selecting unit 18 may be operated or controlled by manual/human intervention, wherein for instance an operator of the device may initially define at least some of the respective regions, 44, 46, 48, 50, 52. The operator may be the subject 34 itself, or may be an auxiliary person, medical staff, etc. The region selecting unit 18 may be arranged for (overall) macroscopic motion detection and compensation. That is, macroscopic movements of the subject 34 with respect to the background may be tracked and compensated in the obtained signals. However, needless to say, preferably microscopic movements, particularly ballistocardiographic movements may pass a respective motion-compensation algorithm and procedure for being further processed by the processing module 14.

The data conditioning unit 20 is arranged to condition the signals, particularly photoplethysmographic signals obtained from the indicative regions 44, 46. To this end, the reference regions 48, 50, 52 may be utilized as a source for reference signals. Hence, a distortion-compensated signal baseline may be obtained based on which the signals obtained from the indicative regions may be conditioned or rectified. Preferably, particularly migraine- indicative characteristic signal patterns remain in the signals obtained from the indicative regions 44, 46.

Corrected and conditioned data may be transmitted to the detector unit 22 which is arranged to detect characteristic signal discrepancies at or between the indicative regions 44, 46. As already explained above, migraine attacks and even impending or imminent migraine attacks may be reflected in a so-to-say asymmetric blood pulsation or blood perfusion between the right side and the left side of the face of the subject 34. Hence, the presence and level of the detected discrepancy is indicative of migraine or migraine- related symptoms.

In a particular non- limiting embodiment, the device 10 is arranged as a mobile device, and comprises a camera integrally arranged observing module 12, for measuring PPG signals, wherein the camera is arranged to observe a facial portion 38 and neck portion 40 of the subject of interest 34. In one example, the device 10 is arranged to measure PPG signals at various wavelengths (e.g. RGB channels, R and IR channels, or even a combination thereof). Further, a computing or processing module 14 is provided which may form an integral part of the mobile device 10. The device 10 may therefore also utilize screens, touchscreens, and wireless and wire communications for interaction with users and further devices (e.g. fitness trackers, computing networks, Sp02 sensors, etc.). Hence, the device 10 may provide feedback to the migraine sufferers and to further people.

PPG detection may involve defining a pattern of sub-regions (covering one or more pixels) of the skin of the subject 34 which may be processed accordingly. As a result, a PPG imaging map of the spatial distribution of the PPG waveform (including shape, amplitude, and phase) may be provided and further processed. Accordingly, a so-called blood perfusion imaging baseline may be obtained. Preferably, an initial version of the blood perfusion imaging baseline is generated in a basically migraine-free period wherein no symptoms are present in the subject 34. A corresponding record may be referred to as blood perfusion imaging baseline. The blood perfusion imaging baseline may be used as reference value for calculation of a migraine probability.

In one embodiment, the device 10 is operated to monitors changes in the PPG pulsatility map over a defined period of time, by way of a spot-check monitoring or in a basically continuous fashion. Accordingly, the obtained PPG imaging maps at various wavelengths may be compared against blood perfusion imaging baselines acquired at the same wavelengths during the first, initial monitoring procedure. Then, differences in pulsatility of the PPG signals at the defined wavelengths may be calculated, dependent on the respective spatial location of the sub-regions of the skin portion. Consequently, a two-dimensional map of changes in the blood perfusion may be detected and registered. Hence, the two-dimensional map may be further processed and analyzed, involving a comparison with the previously obtained blood perfusion imaging baseline, so as to estimate a pre-disposition or probability of having or becoming a migraine attack. Further on, mid-term and long-term measurement may be used to record the temporal and dynamic course of the migraine.

Further reference is made to Fig. 3, Fig. 4, and to Fig. 5. Figs. 3 to 5 represent so-called photoplethysmographic charts which may be obtained by remote

photoplethysmographic monitoring from skin portions of the subject 34. In Figs. 3, 4 and 5, an axis of abscissas indicates time. An ordinate axis indicates a photoplethysmographic (PPG) signal which may be aggregated over a respective region of interest (which typically covers a plurality of image pixels). Hence, the charts shown in Figs. 3 to 5 illustrate a temporal record of (spatially averaged and accumulated) skin color fluctuations that are attributable to blood pulsation or blood perfusion.

Fig. 3 illustrates a reference chart 58 representing a reference PPG signal. The reference PPG signal may be obtained from the PPG reference region 50, and may be therefore not affected by migraine.

By contrast, Fig. 4 and Fig. 5 exemplarily illustrate PPG charts 60, 62 and 68,

70 respectively which are obtained from the indicative regions 44, 46 at the skin of the subject 34.

As can be directly seen in Fig. 4 and in Fig. 5, a characteristic signal discrepancy between the regions of interest 44, 46 may be present in the PPG-signals when a migraine attack is present or at least impending. Fig. 4 illustrates a first PPG signal 60 having a first amplitude 64. Further, a second PPG signal 62 having a second amplitude 66 is illustrated. As with the embodiment of Fig. 4, the signals 60, 62 are basically synchronized (phase- synchronized) but exhibit different amplitudes. As both signals 60, 62 are obtained from the same subject, and from basically similar and comparable portions thereof, the discrepancy between the amplitudes 64, 66 is a measure for migraine-related symptoms, or migraine as such.

Similarly, Fig. 5 illustrates a first PPG signal curve 68 and a second PPG signal curve 70. By way of example, the PPG signals 68, 70 have similar or basically equal amplitudes, but are phase-shifted, refer to the phase shift indicated by reference numeral 72 in Fig. 5. Since similar and comparable regions of the subject 34 are utilized for obtaining the PPG signals 68, 70, basically no or only a small phase shift may be expected in healthy subject that do not suffer from migraine. Basically the same applies to migraine patients which are in a basically migraine-free state. Therefore, the presence of a significant phase shift 72 is highly indicative of an impending, imminent or an actual migraine attack.

Generally, also a combination of the discrepancies illustrated in Figs. 4 and 5 may be present. Therefore, both an amplitude difference and a phase shift may be present between PPG signals obtained from the first indicative region 44 and the second indicative region 46. In a more general context, a spatial signal shape discrepancy between the indicative regions 44, 46 may be detected, particularly a characteristic spatial PPG signal discrepancy. The detector unit 22 is basically arranged to detect migraine indicative signal discrepancies. Since the PPG data is conditioned by the conditioning unit 20, migraine detecting and migraine predicting reliability can be further improved.

It has been observed that also an impending or imminent migraine attack may be reflected in respective signal discrepancies between the PPG signals obtained from the first indicative region 44 and the second indicative region 46. This may even be the case when common migraine-symptoms or migraine-indicators (severe headache, etc.) are not yet present. Therefore, also migraine forecast or migraine prediction may be envisaged.

While main aspects of the disclosure are described and detailed herein with particular reference to photoplethysmographic (PPG) imaging, further alternative remote imaging methods may be utilized, such as Laser speckle imaging (LSI) Laser speckle contrast imaging (LSCI), for instance. Therefore, the description of particular embodiments provided herein shall not be construed in a limiting sense. Therefore, photoplethysmographic signals which are frequently mentioned herein as a representative signal type may be replaced or supplemented by Laser speckle imaging signals in respective embodiments.

So as to further process data provided by the detector unit 22, the analyzing unit 24 is present in the processing module 14. The analyzing unit may be arranged to assess an actual migraine status based on the detected signal discrepancy. To this end, the analyzing unit 24 may utilize model-based approaches and approaches based on empirically obtained data. Respective algorithms and reference data may be provided by the storage unit or memory 30. Further, the analyzing unit 24 may be arranged to trace and track the course of migraine attacks and non-migraine periods and may therefore generate (or assist in generating) a migraine record or keep (or assist in keeping) a migraine diary. A long-term migraine record may be created in this way which may be used for improved and enhanced migraine treatment, and for assessing medication and therapeutic effects.

Via the output interface 28, the processing module 14 may communicate with, or may be coupled with, appropriate output units, such as screens, speakers, control lights, etc. or with external data processing/exchanging/storing facilities.

Reference is made to Fig. 6, illustrating a simplified block diagram exemplifying an embodiment of a migraine monitoring or migraine detection method in accordance with at least some aspects of the present disclosure. Initially, in a step S10, a subject of interest is observed, wherein image data is generated. Particularly, image data representing blood pulsation or blood perfusion at a skin area of the subject of interest is generated. Preferably, non-contact monitoring or remote monitoring is utilized. The step S10 preferably implements remote photoplethysmographic monitoring.

In a further step S12, the image data obtained in the step S10 is conditioned. Conditioning the image data may include deriving photoplethysmographic (PPG) signals from more than one region in the image data. For instance, so-called indicative regions which are indicative of migraine-related symptoms and so-called non-indicative reference regions may be utilized to this end. Indicative regions may be selected in the facial skin portion of the subject of interest. Reference regions may be selected in a non-facial skin portion of the subject of interest. Further reference regions may involve transitional regions including a skin-to-background transition for motion detection, and further reference regions, for instance a nasal region.

Thanks to the reference regions, a signal baseline may be obtained which may be utilized to condition the signals obtained from the indicative regions so as to attenuate or discard non-migraine-related signal patterns.

In a further step SI 4, a characteristic spatial blood perfusion discrepancy between at least two regions of the skin area is detected. The characteristic discrepancy is preferably detected by comparing at least two indicative regions at different face portions of the subject of interest. A significant discrepancy between the right face side and the left face side may be highly indicative of an imminent or an actual migraine attack. The signal discrepancy may be reflected in PPG signals obtained from the indicative regions, and may for instance involve an amplitude discrepancy, a phase discrepancy, or, more generally, a signal shape discrepancy. The detected blood perfusion discrepancy is indicative of at least one migraine-related symptom or, more generally, of an imminent or an actual migraine attack.

A further step S16 may follow which includes the assessment of an actual migraine status based on at least one model-based data and empirically obtained data.

Assessing the migraine status may also involve keeping a migraine diary or, more generally, generating a migraine record over a long period. Further, assessing the actual migraine status may also involve predicting an impending or imminent migraine attack since the

characteristic signal discrepancy, at least in some cases, may be already detected when no clearly perceivable migraine symptom (headache, etc.) is present yet.

Further reference is made to Fig. 7 illustrating an exemplary embodiment of a migraine detection and monitoring procedure utilizing reference data for improving accuracy and reliability. Hence, a further embodiment of the present disclosure relates to a processing module which is arranged to carry out additional steps so as to further enhance the migraine detection. Accordingly, virtual (software) or discrete (hardware) processing blocks may be provided which, as a result, allow more reliable diagnosis of migraine in an uncontrolled (non- laboratory) environment. Preferably, the processing module is arranged to process image data captured by off-the-shelf (remotely-detecting) camera systems.

In accordance with this embodiment, the processing module, particularly the data conditioning unit and the detector unit thereof, comprise(s) the following blocks:

a) a spatial PPG reference extraction block arranged to measure a shape, amplitude and time of a PPG signal from non-face skin area (invariant to migraine impacts), wherein the data is used as a reference to estimate the amplitude discrepancy (e.g.

asymmetry) and temporal discrepancy (e.g. phase asymmetry),

b) a temporal PPG reference extraction block arranged to measure PPG signals from at least two distinct regions of a face, affected by migraine, and

c) a motion analysis block arranged to differentiate an impact of ballistocardiography from the PPG phase discrepancy.

Fig. 7 shows a corresponding flowchart of an exemplary embodiment of the disclosure. At the first step S50, an image stream (video stream) of a subject of interest is provided. The step S50 may involve an automatic or manual selection of several regions of interest in the image of a face and torso of the subject of interest (refer also to Fig. 2). For instance, a PPG reference region contains a skin area outside a face (e.g. on a neck), where the shape and amplitude of a PPG signal is not affected by migraine. A step S52 indicates corresponding signal processing action. The respective region may be referred to as blood perfusion reference region.

Further, a first indicative region and a second indicative region are selected on opposite sides of a face and contain skin areas with PPG signals affected by migraine.

Preferably, first indicative region and second indicative region are selected on opposite cheeks of a person, since those may be regarded as sources of a strong PPG signal. A step S54 indicates corresponding signal processing action for the first indicative region. A step S56 indicates corresponding signal processing action for the second indicative region.

Further, a motion reference region is selected, e.g. on the edge of the head (or another portion including a basically high-contrast edge), and may be used for detection and estimation of ballistocardiography motion, which might affect the analysis of phase difference between PPG signals of the first indicative region and the second indicative region. A step S58 indicates corresponding signal processing action. The respective region may be referred to as ballistocardiographic motion reference region.

A further step S60 relates to ballistocardiographic motion compensation based on motion reference data obtained from the motion reference region. Further, in case the amplitude of ballistocardiographic motion is within the range of amplitudes detected during previous measurements, the phase difference between PPG signals extracted from first indicative region and second indicative region may be calculated. If the phase difference between PPG signals from first indicative region and second indicative region is higher than the difference measured in previous sessions (preferably during periods of time without migraine pain), then quite probably an actual symptom (headache, for instance) is caused by migraine.

The amplitude and shape of PPG signals extracted from first indicative region and second indicative region are compared with a PPG signal extracted from the PPG reference region, step S62. In an exemplary refinement, the step S62 may further include a comparison of the differences in shape and amplitude between the first indicative region/PPG reference region and the second indicative region/PPG reference region with corresponding differences calculated in previous periods of time, which may be referred to as baseline (preferably detected during periods of time without migraine symptoms). Large differences in shape and amplitude may basically point to a high probability of a migraine attack.

At a further step S64, the results of the shape, amplitude, and phase asymmetry analysis are estimated to define an (overall) likelihood of an actual or imminent migraine attack, as a source of an actual symptom (headache, for instance). In case no migraine-related differences (asymmetries) may be detected, probably another reason for the actual symptom is present. Also migraine prediction based on the processed signals may be envisaged.

In the current embodiment both the steps S60 and S62 are executed for monitoring migraine. In an alternative embbodiment, one of the steps, i.e. S60 or S62, may be executed to monitor migraine.

A device in accordance with the current disclosure may be suited for being used for diagnosis of migraine and management of migraine episodes over longer time periods.

In a further embodiment, the device may also process additional information, e.g. estimated hormonal changes (e.g. fluctuations of estrogen hormone in women) as an increasing factor of having a migraine attack. In another embodiment, the device may also automatically keep a migraine diary. Hence, migraine management can be greatly facilitated, which may involve detecting and tracking of particular patterns of the (recurrent) migraine attacks. These insights may be linked with potential triggers that might cause migraine attacks. Hence, the device may also generate and provide data for enhancing post-processing analysis which may include the determination of further derived information and signals.

While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. A migraine monitoring device (10) comprising:

an imaging unit (16) arranged to observe a subject (34), and to generate image data representing blood perfusion at a skin area of the subject (34),

a detector unit (22) arranged to detect a characteristic spatial blood perfusion discrepancy between at least two regions (44, 46) of the skin area, and

a data conditioning unit (20) arranged to condition the image data, taking into account reference image data obtained from at least one reference region (48, 50) of the subject (34),

wherein the data conditioning unit (20) is arranged to process a plurality of regions (44, 46, 48, 50, 52), wherein the regions comprise a first indicative region (44), at least one second indicative region (46), and at least one of a ballistocardiographic motion reference region (48) and a blood perfusion reference region (50),

2. The device (10) as claimed in claim 1, wherein the data conditioning unit (20) is arranged to discard at least one of ballistocardiographic motion artefacts and

vasoconstriction-indicative artefacts in the image data, the artefacts being non-indicative of the migraine-related symptom.

3. The device (10) as claimed in claim 1, further comprising a region selecting unit (18) arranged to define at least three regions comprising the at least one reference region (48, 50) and the at least two indicative regions (44, 46) in the image data obtained from the imaging unit (16).

4. The device (10) as claimed in claim 3, wherein the region selecting unit (18) is arragned to define: a first region of interest (44) that is arranged as a first indicative region in a facial portion of the subject (34),

a second region of interest (46) that is arranged as a second indicative region in a facial portion of the subject (34), wherein the second region of interest (46) is spaced apart from the first indicative region,

a third region of interest (48) as a motion reference region that is arranged as a ballistocardiographic motion reference region, and

a fourth region of interest (50) that is arranged as a blood perfusion reference region in a non-facial skin portion of the subject (34).

5. The device (10) as claimed in claim 3, wherein the region selecting unit (18) is arranged to discriminate a facial portion (38) and a non-facial skin portion (40) of the subject of interest (34), and a transitional portion (42), wherein the transitional portion (42) comprises a skin-to-background transition.

6. The device (10) as claimed in claim 1, wherein the imaging unit (16) is a remote monitoring imaging unit (16), particularly a remote photoplethysmographic monitoring imaging unit or a remote laser speckle imaging unit, wherein the detector unit (22) is arranged to detect color fluctuations in the obtained image data that are indicative of blood perfusion.

7. The device (10) as claimed in claim 6, wherein the detector unit (22) is arranged to detect at least one of a spatial discrepancy between amplitudes (64, 66) of photoplethysmographic signals (60, 62) obtained from the observed regions (44, 46) and a spatial discrepancy (72) between phases of photoplethysmographic signals (68, 70) obtained from the observed regions (44, 46).

8. The device (10) as claimed in claim 6, wherein the detector unit (22) is arranged to detect a migraine-related vasoconstriction-indicative signal in an observed region, particularly in a nasal region.

9. The device (10) as claimed in claim 1, wherein the detector unit (22) is arranged to detect a characteristic spatial blood perfusion discrepancy between a right facial side of the subject (34), and a left facial side of the subject (34).

10. The device (10) as claimed in claim 9, wherein the detector unit (22) is further arranged to detect at least one of an amplitude asymmetry and an asynchronicity of facial blood perfusion between different observed facial regions (44, 46) of the subject (34).

11. The device (10) as claimed in claim 1, further comprising an analyzing unit (24) arranged to assess an actual migraine status based on at least one of model-based data and empirically obtained data.

12. A migraine monitoring method comprising the following steps:

observing a subject (34), including generating image data representing blood perfusion at a skin area of the subject (34),

conditioning the image data, taking into account reference image data obtained from at least one reference region (48, 50) of the subject (34),

wherein the step of conditioning involves processing a plurality of regions (44,

46, 48, 50, 52), wherein the regions comprise a first indicative region (44), at least one second indicative region (46), and at least one of a ballistocardiographic motion reference region (48) and a blood perfusion reference region (50), and

detecting a characteristic spatial blood perfusion discrepancy between at least two regions (44, 46) of the skin area,

13. The method as claimed in claim 12, wherein the step of conditioning the image data comprises:

detecting at least one of ballistocardiographic motion artefacts and vasoconstriction- indicative artefacts in the image data, wherein the detected artefacts are non- indicative of the migraine-related symptom, and

attenuating the non- indicative artefacts in the image data.

14. The method as claimed in claim 12, further comprising: defining a first region of interest (44), particularly a first indicative region in a facial portion of the subject (34),

defining a second region of interest (46), particularly a second indicative region in a facial portion of the subject (34), that is spaced apart from the first indicative r ion,

extracting a second photoplethysmographic signal from the second region of interest (46),

detecting a migraine-indicative phase discrepancy, particularly a spatial phase asymmetry between the first photoplethysmographic signal and the second

photoplethysmographic signal, under consideration of reference data obtained from the third region of interest (48), and

detecting a migraine-indicative signal shape or amplitude discrepancy, particularly a spatial shape or amplitude asymmetry based on the first photoplethysmographic signal and the second photoplethysmographic signal, under consideration of a reference photoplethysmographic signal obtained from the fourth region of interest (50).

15. Computer program comprising program code means for causing a computer to carry out the steps of the method as claimed in claim 12 when said computer program is carried out on a computer that is coupled with a device as claimed in claim 1.