Systems and methods for an electroencephalographic (EEG) biomarker for migraine

Abstract

A system can identify a quantitative biomarker for migraine in real-time from electrophysiological measurements and can be made using current wearable devices. The system uses the quantitative biomarker to track the onset and offset of migraine and also the degree of migraine pain which can be incorporated to guide and control occipital nerve stimulation parameters.

Description

SYSTEMS AND METHODS FOR AN ELECTROENCEPHALOGRAPHIC (EEG) BIOMARKER FOR MIGRAINECROSS REFERENCE TO RELATED APPLICATIONS

[0001] This is a PCT patent application that claims benefit to U.S. Provisional Application Serial No. 63 / 540,613 filed on September 26, 2023, which is herein incorporated by reference in its entirety.FIELD

[0002] The present disclosure generally relates to treatments, therapies, and devices for migraines and, in particular, to systems and methods for the detection of migraine using Electroencephalography (EEG) that would allow for closed-loop control of neurostimulators used to reduce migraine symptoms. This disclosure includes: (I) a biomarker for migraine identified through real-time EEG measurements in an experimental model of migraine that involves nitroglycerin (NTG) administration; and (II) occipital nerve stimulation (ONS) using injectable wireless neurostimulators to restore migraine induced changes in cortical or EEG activity.BACKGROUND

[0003] Chronic intractable migraine is a severe neurological disorder that significantly impacts the lives of millions of people around the world. Migraines in general impact millions worldwide. Yet, present technology does not sufficiently meet the needs of migraine detection and treatment. For example, current methods of physiological-based migraine detection methods involving tissue (i.e. , blood) sampling which is a cumbersome process.

[0004] It is with these observations in mind, among others, that various aspects of the present disclosure were conceived and developed.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The present patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

[0006] FIG. 1 is a simplified diagram showing a system for migraine detection and mitigation based on quantitative biomarkers outlined herein.

[0007] FIG. 2 is a simplified block diagram showing a computing device that can be used to implement aspects of the system of FIG. 1.

[0008] FIGS. 3A-3D are a series of graphical representations showing mean power spectrograms for cortical activity captured for control rats and for rats with induced migraine and varying dosages of migraine-inducing substance (nitroglycerin), where FIG. 3A shows cortical activity for rats without migraine (administered saline), FIG. 3B shows cortical activity for rats with migraine (administered 15mg / kg nitroglycerin), FIG. 3C shows cortical activity for rats with migraine (administered 13mg / kg nitroglycerin), and FIG. 3D shows cortical activity for rats with migraine (administered 10mg / kg nitroglycerin).

[0009] FIGS. 4A-4D are a series of graphical representations showing total power over time across five frequency bands (Delta 0.5-4Hz, Theta 4-8Hz, Alpha 8- 14Hz. Beta 14-30Hz, and Gamma 30-55Hz) that correlate with the graphical representations in FIGS. 3A-3D, where FIG. 4A shows total power for rats without migraine (administered saline) and without significant suppression of power in any band, FIG. 4B shows total power for rats with migraine (administered 15mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands, FIG. 4C shows total power for rats with migraine (administered 13mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands, and FIG. 4D shows total power for rats with migraine (administered 10mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands.

[0010] FIG. 5 is a graphical representation showing change in power over time after injection across five frequency bands (Delta 0.5-4Hz, Theta 4-8Hz, Alpha 8- 14Hz. Beta 14-30Hz, and Gamma 30-55Hz) that correlate with the graphical representations in FIGS. 3A-3D and 4A-4D.

[0011] FIGS. 6A-6E are a series of graphical representations showing mean power spectrograms for cortical activity captured for control rats, rats with induced migraine, and rats with induced migraine and various regimes of occipital nerve stimulation (ONS), where FIG. 6A shows a mean power spectrogram for cortical activity for control rats without migraine (administered saline), FIG. 6B shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) but without ONS, FIG. 6C shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with prophylactic ONS (administered prior to injection of nitroglycerin), FIG. 6D shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with short-duration acute intermittent ONS (administered intermittently and after injection of nitroglycerin with a duration of about 100 seconds), and FIG. 6E shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with long- duration acute intermittent ONS (administered intermittently and after injection of nitroglycerin with a duration of about 9 minutes per occipital nerve, for 18 minutes total).

[0012] FIG. 7 is a process flow diagram showing a process for implementation of the system of FIG. 1 for migraine detection and mitigation based on quantitative biomarkers.

[0013] FIGS. 8A-8D are a series of photographs showing an experimental setup for development of the system of FIG. 1 and acquisition of the example electrophysiological information shown in FIGS. 3A-6E.

[0014] FIGS. 9A and 9B are a pair of graphical representations showing effect of different doses of nitroglycerin (15 mg / kg, 13 mg / kg, 10 mg / kg) — used to induce migraine — and saline (control) on the duration of power suppression (FIG. 9A) and enhancement (FIG. 9B) in the five EEG bands for a first example outlined herein.

[0015] FIGS. 10A-10C are a series of graphical representations showing behavioral assessments on n=6 animals using the orofacial pain assessment device (OPAD) used to quantify degree of headache pain following a drug injection

[0016] FIGS. 11A and 11 B respectively show an injectable neurostimulator and an example of the pseudo-monophasic output of the neurostimulator when powered by an ultrasound transducer showing a 20 ps burst of sinusoidal pulses running at 2 MHz for a second example outlined herein.

[0017] FIGS. 12A-12F are a series of graphical representations showing cortical evoked responses due to occipital nerve stimulation

[0018] FIG. 13 shows a timeline of events in an experiment day for each animal and a representative spectrogram generated from an EEG electrode from one animal after intraperitoneally (i.p.) injecting 13 mg / kg nitroglycerin (NTG) or Saline.

[0019] FIGS. 14A-14C are a series of graphical representations showing mean power spectrogram of all animals in the (FIG. 14A) control condition (saline) and (FIG. 14B) nitroglycerin condition (NTG 13 mg / kg), where FIG. 14C shows a difference spectrogram created by subtracting the mean spectrogram of the Nitroglycerin condition from that of the control condition with black outlines revealing regions of significant differences (p < 0.05) that was identified using the nonparametric cluster-based permutation test.

[0020] FIGS. 15A-15C are a series of graphical representations showing change in EEG power as a function of time in different frequency bands including (FIG. 15A) low frequency band (0.5-30 Hz), (FIG. 15B) low-gamma Band (30-60 Hz) and (FIG. 15C) high gamma band (60-100 Hz) following i.p. injection of saline (n=10) (green trace), and NTG 13mg / kg (n=10) (red trace). Graphs represent data from the left (left column) and right (right column) hemispheric HLS1 electrodes.

[0021] FIGS. 16A and 16B are a pair of graphical representations showing time and frequency averaged changes in EEG power with respect to baseline recordings in (FIG. 16A) the left hemisphere HLS1 electrode and (FIG. 16B) the right hemisphere electrode. Power changes in three frequency bands at six time segments (18 bars) are displayed. Data are shown as mean standard error of the mean (SEM). Individual mean bars were created after computing an average power in a given time segment and frequency band window in the spectrogram from each animal in the control (saline) condition (green bars) and the NTG condition (red bars). Two-way repeated measures ANOVA followed by the false-discovery rate (FDR) multiple comparison correction, nd p> 0.05, *FDR-adjusted p < 0.05.

[0022] FIGS. 17A and 17B are a pair of graphical representations showing time and frequency averaged changes in EEG coherence with respect to baseline recordings in (FIG. 17A) in different frequency bands as a function of time and (FIG. 17B) the right hemisphere electrode. Data are shown as meant standard error of the mean (SEM), from control (n=10) animals (green bars) and NTG (n= 10) animals (redbars). Two-way repeated measures ANOVA followed by the false-discovery rate (FDR) multiple comparison correction, *FDR-adjusted p < 0.05.

[0023] FIGS. 18A-18H are a series of graphical representations showing spontaneous behavioral measurements during a 10-min open field session performed at 30 min, 2 hrs and 4 hrs following injection of either saline or NTG. (FIG. 18A) distance (m) travelled in cage, (FIG. 18B) average speed (m / s) of the animal in the cage, (FIG. 18C) number of freezing episodes, (FIG. 18D) time the animal was immobile, (FIG. 18E) time spent in the inner square zone, (FIG. 18F) distance travelled in the inner square zone, (FIG. 18G) total time the animal spent rearing and (FIG. 18H) total time the animal spent grooming. Data in the red bar is from the NTG cohort (n=10), green bar is from the control cohort (n=10). Data are shown as mean standard error of the mean (SEM). Two-way repeated measures ANOVA followed by the Holm-Sidak multiple comparison test, *p < 0.05, **p < 0.01 , ***p < 0.001 .

[0024] FIGS. 19A and 19B show: (FIG. 19A) Experiment timeline for all three ONS regimens tested, including long-duration acute intermittent ONS, short duration ONS intermittent ONS, and prophylactic ONS. Black arrows point to time segments an ONS treatment was applied. (FIG. 19B) A representative spectrogram generated from the EEG collected from an animal receiving a 13 mg / kg dose of NTG and a prophylactic dose of ONS treatment.

[0025] FIGS. 20A-20C are a series of graphical representations showing temporal changes in the power of three EEG bands following saline / NTG 13 mg / kg administration, including, (A) low frequency band (0.5-30 Hz), (B) low-gamma band (30-60 Hz) and (C) high-gamma band. Data from five different experimental conditions are shown including saline (green) (n=10), NTG 13 mg / kg (red) (n=10), NTG13 mg / kg + prophylactic ONS treatment (blue) (n=10), NTG13 mg / kg + acute (short) intermittent ONS treatment (light blue) (n=5), and NTG13 mg / kg + acute (long) intermittent ONS treatment (black) (n=5). Data is presented as mean dB change in power from baseline recording (darker color line) and 95% confidence interval (lighter color area around the line).

[0026] FIG. 21 is a graphical representation showing power of four spectrogram regions / windows in the 0.5-20 Hz frequency range from (A) the left hemisphere HLS1 electrode and (B) the right hemisphere HLS1 electrode. The time window for the first region (R1 ) is 55-72 min, the second region (R2) is 100-116 min, the third region (R3) is 137-157 min, and the fourth region (R4) is 182-234 min. Barsare presented as mean db change from baseline ± standard error of the mean (SEM) for five experimental conditions including saline (green) (n=10), NTG 13 mg / kg (red) (n=10), NTG13 mg / kg + prophylactic ONS treatment (blue) (n=10), NTG13 mg / kg + acute (short) intermittent ONS treatment (light blue) (n=5), and NTG13 mg / kg + acute (long) intermittent ONS treatment (grey) (n=5). Two-way repeated measures ANOVA followed by the false-discovery rate (FDR) multiple comparison correction, nd p> 0.05, *FDR-adjusted p < 0.05.

[0027] FIGS. 22A and 22B are a pair of graphical representations showing time and frequency averaged EEG power changes following injection in (FIG. 22A) the left hemisphere HLS1 electrode and (FIG. 22B) the right hemisphere HLS1 electrode. Data is divided into three frequency bands, each displaying average power at six-time segments. Bars represent mean db change in power from baseline ± standard error of the mean (SEM) for five experimental conditions including saline (green) (n=10), NTG 13 mg / kg (red) (n=10), NTG13 mg / kg + prophylactic ONS treatment (blue) (n=10), NTG13 mg / kg + acute (short) intermittent ONS treatment (light blue) (n=5), and NTG13 mg / kg + acute (long) intermittent ONS treatment (grey) (n=5). Two-way repeated measures ANOVA followed by the false-discovery rate (FDR) multiple comparison correction, nd p> 0.05, *FDR-adjusted p < 0.05.

[0028] FIGS. 23A and 23B are a pair of graphical representations showing changes in coherence with respect to baseline following injection for three different frequency bands. (FIG. 23A) Coherence changes as a function of time (each point represents a 2-min average). (FIG. 23B) Coherence changes in six-time segments following injection are displayed. Bars represent the mean change in coherence from baseline ± standard error of the mean (SEM) for five experimental conditions, including saline (green) (n=10), NTG 13 mg / kg (red) (n=10), NTG13 mg / kg + prophylactic ONS treatment (blue) (n=10), NTG13 mg / kg + acute (short) intermittent ONS treatment (light blue) (n=5), and NTG13 mg / kg + acute (long) intermittent ONS treatment (grey) (n=5). Two-way repeated measures ANOVA followed by the false- discovery rate (FDR) multiple comparison correction, nd p> 0.05, *FDR-adjusted p < 0.05.

[0029] FIGS. 24A-24H are Spontaneous behavioral measurements during a 10-minute open field session performed at 30 minutes, 2 hours and 4 hours following injection of either saline or NTG. (FIG. 24A) distance (m) travelled in cage, (FIG. 24B) average speed (m / s) of the animal in the cage, (FIG. 24C) number of freezingepisodes, (FIG. 24D) time the animal was immobile, (FIG. 24E) time spent in the inner square zone, (FIG. 24F) distance travelled in the inner square zone, (FIG. 24G) total time the animal spent rearing and (FIG. 24H) total time the animal spent grooming. Data from different experimental conditions are shown including saline (green) (n=10), NTG 13 mg / kg (red) (n=10), NTG13 mg / kg + prophylactic ONS treatment (blue) (n=10), NTG13 mg / kg + acute (short) intermittent ONS treatment (light blue) (n=5), and NTG13 mg / kg + acute (long) intermittent ONS treatment (grey) (n=5). Data is shown as mean standard error of the mean (SEM). Two-way repeated measures ANOVA followed by the Tukey multiple comparison post-hoc test, ns p> 0.05, *p < 0.05, **p < 0.01 , ***p < 0.001 .

[0030] Corresponding reference characters indicate corresponding elements among the view of the drawings. The headings used in the figures do not limit the scope of the claims.DETAILED DESCRIPTION

[0031] The present disclosure relates to detecting a quantitative biomarker of migraine onset and offset, and also systems and methods for: (I) differential modulation of gamma band activity in EEG during migraine in rodents; and (II) occipital nerve stimulation (ONS) using injectable, wireless neurostimulators which can be modulated based on the biomarker to cause electrophysiological changes in rodent models of migraine as further described herein.

[0032] In general, the present inventive concept includes technology related to a quantitative biomarker of migraine onset and offset. The biomarker is an EEG profile that can be detected within 30 minutes of the onset of migraine. This is an improvement on current methods of physiological-based migraine detection methods involving tissue (i.e. , blood) sampling which is a cumbersome process. This biomarker can be used to detect migraines in real time and measured with current wearable devices.

[0033] Validation and testing were conducted in a rodent model by testing the relationship of increasing dose of nitroglycerin (to induce migraines of increasing severity) on EEG output. It was discovered that EEG output signature is immediate after nitroglycerin dose, and that EEG signature includes a suppression of power of all EEG frequencies below 30 Hz and in the high gamma band (60-100 Hz). This is a quantitative biomarker for migraine that is real-time and derived from electrophysiological recordings. It was further discovered that duration of measured EEG biomarker profile changes is dependent on dose of nitroglycerin.Introduction

[0034] Migraine is one of the top three most debilitating neurological disorders, measured in terms of years lived with disability (YLD). Therapeutic approaches have not been fully successful leaving a significant number of people resistant to treatment. A quantitative biomarker of migraine will be useful both for diagnosis and treatment of migraine. However, there are no real-time, quantitative biomarkers of migraine currently. Molecular biomarkers are episodic because they require sampling of blood and cumbersome assays that are not real-time.

[0035] Accordingly, a need exists for a real-time biomarker for pain, in general and migraine in particular. Described here is an electrophysiological biomarker for migraine demonstrated in a rodent model of chronic migraine.Consistent changes in energy in specific frequency bands of the power spectrum of the electroencephalogram (EEG) is manifested in the first 30 min of the onset of migraine, followed by a rapid restoration back to normal values.System Overview

[0036] The present disclosure outlines systems and methods for identifying migraine based on quantitative biomarkers present within electrophysiological information captured from the primary somatosensory cortex (“S1” region of the brain). Further, the systems and methods outlined herein include restoring changes in cortical activity as a sign of treating or otherwise mitigating migraine by application of occipital nerve stimulation which can be controlled in a closed-loop manner based on the electrophysiological information.

[0037] Note that while the term “migraine” as used herein applies to a moderate to severe headache caused by a neurological disorder, the systems outlined herein may be similarly applied to other chronic headache conditions such as cluster headaches.

[0038] Experimental results outlined herein indicate a correlation between the onset of migraine and suppression of low-frequency electroencephalogram (EEG) activity (below 30Hz) in the S1 region of the brain. This type of EEG profile can be used as a quantitative biomarker for evaluating onset, offset, and intensity of migraine or other headache disorders. Rats with induced migraine showed a significant reduction or suppression of low-frequency activity in the S1 region as captured using EEG, while control rats without induced migraine showed no significant changes in low-frequency activity in the S1 region. In some rats, enhancement of power in the gamma frequency band (30-55Hz) can be observed. However, after performing more experiments to validate this observation, we found that the early enhancement in gamma power observed within the first 30 minutes after migraine induction was not significantly different from control animals. On the contrary, a long-lasting suppression in gamma band power that started 90 mins after migraine induction and lasted for at least 4 hours was observed and found to be significantly different compared to control animals. This delayed long-lasting suppression in gamma band power, can also be used as a biomarker that indicates the presence of migraine.

[0039] Further, experiments outlined herein indicate that occipital nerve stimulation (ONS) appears to restore cortical activity associated with frequencies below the gamma band (low-frequency, <30Hz). Current headache disorder treatment regimens employing ONS for migraine pain reduction have not been able to incorporate real-time feedback measurements for modulating or controlling ONS dose, as little was known about migraine-indicating biomarkers. Current ONS systems that do not incorporate feedback risk overstimulation of the occipital nerve, which can cause or further exacerbate pain for the individual.

[0040] As such, the present disclosure outlines a closed-loop system that incorporates EEG information to identify an onset, offset, and / or severity of a migraine (or cluster headache) and adjust ONS parameters accordingly for effective treatment. For example, if EEG data indicates a moderate migraine, rather than a severe migraine, then the system can initiate ONS and adjust ONS parameters to address the moderate migraine while avoiding overstimulation of the occipital nerve. In another example, if EEG data indicates that cortical activity has been restored during application of ONS (i.e., indicating offset or reduction in severity of migraine), then the system can cease or reduce ONS stimulation based on the EEG data.

[0041] Further, ONS is known to be an effective prophylactic treatment for chronic migraine patients, where patients often use it throughout the day prior and / or during the attack to reduce the frequency and intensity of migraines. Interestingly, our experiments outlined herein indicate that short-term, acute application of ONS (e.g., applied over a 2-minute interval, prior to and during migraine) can also be effective for restoring cortical activity in lieu of applying a long 30-minute prophylactic dose just prior, but not during a migraine attack.

[0042] FIG. 1 shows a system 100 including EEG measuring electrodes 110 that capture electrophysiological information for identification of a presence (or absence) of migraine. The EEG measuring electrodes 110 can include two or more measuring electrodes associated with a primary somatosensory cortex (S1 region) of an individual, where a first measuring electrode is positioned along a left hemisphere within the S1 region and a second measuring electrode is positioned along a right hemisphere within the S1 region. Further, in some examples, the EEG measuring electrodes 110 can also include a third measuring electrode as a reference electrode which can be placed in another region, such as the right secondary motor cortex region (M2). The EEG measuring electrodes 110 can also include a fourthmeasuring electrode as a ground electrode which can also be placed in another region, such as a few millimeters posterior to lambda.

[0043] The system 100 can further include a computing device 200 in communication with the EEG measuring electrodes 110 that accesses the electrophysiological information for the individual. The computing device 200 can identify a presence of a quantitative biomarker for migraine of the individual based on the electrophysiological information, which can be an electroencephalogram (EEG) profile that can be detected within 30 minutes of onset of migraine. The quantitative biomarker can indicate onset of migraine and a degree of migraine pain of the individual, which can be identified based on reduction or suppression of power for frequencies present within the electrophysiological information that are below a gamma frequency band (30Hz). Conversely, the absence, reduction, or recession of the quantitative biomarker (i.e. , showing an increase in power for frequencies present within the electrophysiological information that are below a gamma frequency band (30Hz)) can indicate offset of migraine or a reduction in severity of migraine.

[0044] In a further aspect, the system 100 can include a neurostimulator device 120 that is configured to stimulate an occipital nerve (ON) of an individual based on the electroencephalogram profile (i.e., the quantitative biomarker). The computing device 200 can incorporate the electrophysiological information as feedback in real-time and apply one or more control signals to the neurostimulator device based on the electroencephalogram profile for prophylactic (pre-onset) or acute (post-onset) intervention. The neurostimulator device 120 can be injected into the individual or can be implanted by a skin incision or other minimally invasive procedures.

[0045] In a further aspect, the computing device 200 can include or otherwise communicate with a display device 230 that can provide migraine-related information to an individual, including information about the biomarker for migraine. Upon detection of migraine based on the electrophysiological information, the computing device 200 can generate an alert for display at the display device 230 that informs the individual. In some examples, the computing device 200 can display information about a duration and / or intensity / severity of the migraine to the individual. The computing device 200 can also generate an alert for display at the display device 230 that informs the individual about parameters being used for ONSbased on the electrophysiological information. The computing device 200 can access inputs from the individual received though a user interface that can further be used to control aspects of ONS being applied to the individual. In examples where the computing device 200 is operable for predicting an impending migraine based on electrophysiological information and / or information about user behaviors that correlate with pre-migraine biomarkers, the computing device 200 can also be operable for displaying information about the prediction of impending migraine to the individual.Computing Device

[0046] FIG. 2 is a schematic block diagram of an example computing device that may be used with one or more embodiments described herein, e.g., as a component of system 100 and / or as computing device 200 shown in FIG. 1.

[0047] Computing device 200 comprises one or more network interfaces 210 (e.g., wired, wireless, PLC, etc.), at least one processor 220, and a memory 240 interconnected by a system bus 250, as well as a power supply 260 (e.g., battery, plug-in, etc.). Further, computing device 200 can include or otherwise communicate with display device 230 for providing alerts, control / user interfaces, and migraine-related information to an individual, including information about the biomarker for migraine.

[0048] Network interface(s) 210 include the mechanical, electrical, and signaling circuitry for communicating data over the communication links coupled to a communication network. Network interfaces 210 are configured to transmit and / or receive data using a variety of different communication protocols. As illustrated, the box representing network interfaces 210 is shown for simplicity, and it is appreciated that such interfaces may represent different types of network connections such as wireless and wired (physical) connections. Network interfaces 210 are shown separately from power supply 260, however it is appreciated that the interfaces that support PLC protocols may communicate through power supply 260 and / or may be an integral component coupled to power supply 260.

[0049] Memory 240 includes a plurality of storage locations that are addressable by processor 220 and network interfaces 210 for storing software programs and data structures associated with the embodiments described herein. In some embodiments, device 200 may have limited memory or no memory (e.g., nomemory for storage other than for programs / processes operating on the device and associated caches). Memory 240 can include instructions executable by the processor 220 that, when executed by the processor 220, cause the processor 220 to implement aspects of the system 100 and associated methods for migraine detection, prediction, and intervention outlined herein.

[0050] Processor 220 comprises hardware elements or logic adapted to execute the software programs (e.g., instructions) and manipulate data structures 245. An operating system 242, portions of which are typically resident in memory 240 and executed by the processor, functionally organizes device 200 by, inter alia, invoking operations in support of software processes and / or services executing on the device. These software processes and / or services may include Migraine Detection and Mitigation processes / services 290, which can include aspects of the methods and / or implementations of various modules described herein. Note that while Migraine Detection and Mitigation processes / services 290 is illustrated in centralized memory 240, alternative embodiments provide for the process to be operated within the network interfaces 210, such as a component of a MAC layer, and / or as part of a distributed computing network environment.

[0051] It will be apparent to those skilled in the art that other processor and memory types, including various computer-readable media, may be used to store and execute program instructions pertaining to the techniques described herein. Also, while the description illustrates various processes, it is expressly contemplated that various processes may be embodied as modules or engines configured to operate in accordance with the techniques herein (e.g., according to the functionality of a similar process). In this context, the term module and engine may be interchangeable. In general, the term module or engine refers to model or an organization of interrelated software components / functions. Further, while the Migraine Detection and Mitigation processes / services 290 is shown as a standalone process, those skilled in the art will appreciate that this process may be executed as a routine or module within other processes.Identifying Biomarkers for Migraine

[0052] The computing device 200 can apply a Fourier Transform operation to the electrophysiological information resulting in a frequency-domain representation of the electrophysiological information. Using the frequency-domainrepresentation of the electrophysiological information, the computing device 200 can evaluate a change in power for frequencies present within the electrophysiological information relative to frequencies present within previous electrophysiological information for a previous time interval. In other words, the computing device 200 can examine whether an onset or offset of migraine is occurring based on changes in power across a frequency spectrum between a current time interval and a previous time interval. The previous time interval can correlate with baseline or preonset EEG information, or can correlate with previously acquired post-onset EEG information (i.e. , where the individual was previously or is currently experiencing a migraine, and the system 100 is monitoring for improvement at the time of the current time interval).

[0053] The computing device 200 can identify an onset of the quantitative biomarker for migraine of the individual based on a decrease in power for frequencies present within the electrophysiological information that are below the gamma frequency band, relative to frequencies that are below the gamma frequency band present within the previous electrophysiological information.

[0054] To clarify, consider the “previous electrophysiological information” as representing a pre-onset baseline taken at a previous time interval. The previous electrophysiological information can include a frequency-domain representation of activity in the S1 region for the previous time interval. The system 100 can capture the electrophysiological information for a current time interval to evaluate if a migraine is currently occurring. The electrophysiological information for the current time interval can likewise include a frequency-domain representation of activity in the S1 region for the current time interval. If the individual is experiencing or beginning to experience a migraine at the current time interval, then this can be identified or confirmed by comparing the electrophysiological information for the current time interval with the previous electrophysiological information from the previous time interval. If the electrophysiological information for the current time interval indicates a significant decrease, suppression, or otherwise reduction in power for frequencies below the gamma frequency band, then the system 100 can determine that the electrophysiological information for the current time interval is consistent with the quantitative biomarker for migraine. The system 100 can identify or otherwise estimate an onset of migraine as correlating with a time at which thepower for frequencies below the gamma frequency band starts to reduce significantly.

[0055] Further, the system 100 can evaluate or otherwise estimate a relative severity of migraine based on a degree of reduction of power for frequencies below the gamma frequency band. The degree of reduction of power used to evaluate severity can be relative to a baseline (i.e. , “normal” or pre-onset), relative to previous migraine events for the individual, and / or relative to a scale that may be consistent across a plurality of individuals.

[0056] The system 100 can evaluate migraine severity based on a length of time that the reduction in total power for frequency bands below the Gamma band is observed, and / or a magnitude of change in power relative to baseline (non-migraine). In the graphical representations shown in the figures, the duration of the reduction in total power for frequency bands below the Gamma band appears to be dependent upon the dosage of nitroglycerin used to induce migraine.

[0057] In some examples, the quantitative biomarker can further indicate the onset of migraine and the degree of migraine pain of the individual by correlating it with the measured changes in the power of frequencies present within the electrophysiological information. In our experiments, the early immediate suppression observed in low-frequency bands after nitroglycerin administration indicates the onset of the migraine attack, while the consistent suppression in the gamma band lasting at least 4 hours is indicative of migraine pain lasting throughout this period.

[0058] Results of experiments reported herein indicate that, interestingly, within the gamma band (30-100 Hz), changes in power have two distinct profiles. In the low gamma band activity (30-60 Hz) of EEG a delayed significant suppression was observed that was initiated 90 minutes following migraine induction lasting at least for 4 hours from migraine induction (end of EEG recordings in the experiment). On the other hand, EEG power in the high gamma band (60-100 Hz) was significantly suppressed throughout the time-course of experiments following migraine induction (by injection of NTG). The long lasting suppression in power of the low and high gamma bands can be correlated with the degree of migraine attack, while the immediate suppression in the high gamma band indicates an immediate migraine pain onset.

[0059] Likewise, the system 100 can identify an offset of the quantitative biomarker for migraine of the individual based on an increase in power for frequencies present within the electrophysiological information that are below the gamma frequency band relative to frequencies that are below the gamma frequency band present within the previous electrophysiological information (the previous electrophysiological information correlating with presence of the quantitative biomarker for migraine for the previous time interval).

[0060] In some aspects, the system 100 can continuously collect electrophysiological information and / or information about user behaviors for identification of pre-migraine biomarkers which can ultimately enable the computing device 200 to predict an impending migraine. For example, the computing device 200 may be operable to identify aspects of an EEG profile for a user that is likely to correlate with a migraine event prior to onset, based on a collection of EEG profiles for the user that precede migraine events. The system 100 can also incorporate information about other biomarkers (such as but not limited to cardiovascular, respiratory, sleep, exercise, and blood glucose data reported by the individual or obtained from wearable devices worn by the individual) in addition to user behaviors that can be ascertained from devices associated with or otherwise reported by the individual (such as but not limited to geolocation, screen time, app usage, activities, ambient noise or light, plane or car travel, food / drink intake).

[0061] FIGS. 3A-6 show example electrophysiological information including those obtained from control rats as well as rats with induced migraine of varying severity. In particular, these figures show example EEG profiles that indicate the quantitative biomarker for migraine especially when contrasted with control or non-migraine EEG profiles.

[0062] FIGS. 3A-3D show mean power spectrograms for cortical activity captured for control rats and for rats with induced migraine and varying dosages of migraine-inducing substance (nitroglycerin, abbreviated as “NTG”), where FIG. 3A shows cortical activity for rats without migraine (administered saline), FIG. 3B shows cortical activity for rats with migraine (administered 15mg / kg nitroglycerin), FIG. 3C shows cortical activity for rats with migraine (administered 13mg / kg nitroglycerin), and FIG. 3D shows cortical activity for rats with migraine (administered 10mg / kg nitroglycerin). In FIGS. 3B-3D corresponding to induced migraine, one can observe a clear reduction in total power for frequency bandsbelow the Gamma band following injection of nitroglycerin when compared to FIG. 3A corresponding to no induced migraine. Further, in the graphical representations shown in FIGS. 3B-3D, the duration of the reduction in total power for frequency bands below the Gamma band appears to be dependent upon the dosage of nitroglycerin used to induce migraine.

[0063] FIGS. 4A-4D are a series of graphical representations showing total power over time across five frequency bands (Delta 0.5-4Hz, Theta 4-8Hz, Alpha 8-14Hz. Beta 14-30Hz, and Gamma 30-55Hz) that correlate with the graphical representations in FIGS. 3A-3D, where FIG. 4A shows total power for rats without migraine (administered saline) and without significant suppression of power in any band, FIG. 4B shows total power for rats with migraine (administered 15mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands, FIG. 4C shows total power for rats with migraine (administered 13mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands, and FIG. 4D shows total power for rats with migraine (administered 10mg / kg nitroglycerin) and with significant suppression in power for Delta, Theta, Alpha and Beta bands. Note that the solid bars in each of FIGS. 4A-4D indicate intervals where no EEG data was recorded as the rats underwent behavioral testing.

[0064] FIG. 5 is a graphical representation showing change in power over time after injection across five frequency bands (Delta 0.5-4Hz, Theta 4-8Hz, Alpha 8-14Hz. Beta 14-30Hz, and Gamma 30-55Hz) that correlate with the graphical representations in FIGS. 3A-3D and 4A-4D. Notice the significant negative change in power for Delta, Theta, Alpha and Beta bands within the first two hours for NTG dosages of 10mg / kg, 13mg / kg and 15mg / kg. In this study (corresponding to Example 1 outlined herein), NTG dosages of 10mg / kg, 13mg / kg and 15mg / kg all show an early increase in power for Gamma band within the first two hours. However, this enhancement in gamma was reported only in one rat. After performing more experiments on cohort of rats, we found that a statistically significant suppression of power in EEG frequencies in the high gamma band (60-100 Hz) lasting 4 hrs post induction of migraine. Each treatment was given in a separate day. Each bar in the graph represents the mean power in a 30 min window.

[0065] FIGS. 6A-6E correlate with Example 4 outlined herein, and show mean power spectrograms for cortical activity captured for control rats, rats with induced migraine, and rats with induced migraine and various regimes ofoccipital nerve stimulation (ONS). FIG. 6A shows a mean power spectrogram for cortical activity for control rats without migraine (administered saline), FIG. 6B shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) but without ONS, FIG. 6C shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with prophylactic ONS (administered prior to injection of nitroglycerin), FIG. 6D shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with short-duration acute intermittent ONS (administered intermittently and after injection of nitroglycerin with a duration of about 100 seconds), and FIG. 6E shows a mean power spectrogram for cortical activity for rats with migraine (administered 13mg / kg nitroglycerin) and with long- duration acute intermittent ONS (administered intermittently and after injection of nitroglycerin with a duration of about 9 minutes per occipital nerve, for 18 minutes total). Note that the solid bars in each of FIGS. 6A-6E indicate intervals where no EEG data was recorded as the rats underwent behavioral testing.Process

[0066] FIGS. 7 A and 7B show an example process for migraine detection and mitigation that can be implemented using the system 100 of FIG. 1 based on the biomarkers illustrated in FIGS. 3A-6E. Process 300 of FIGS. 7A and 7B can correlate with Migraine Detection and Mitigation processes / services 290 of FIG. 2.

[0067] Referring to FIG. 7A, step 302 of process 300 can include collecting electrophysiological information including electroencephalogram data over a predetermined time period to assess brain state changes in an individual. Step 304 of process 300 can include accessing (e.g., by processor 220 of FIG. 2) electrophysiological information acquired by two or more measuring electrodes associated with a primary somatosensory cortex of the individual. Step 306 of process 300 can include identifying a presence of a quantitative biomarker for migraine of the individual based on suppression of power for frequencies present within the electrophysiological information that are below a gamma frequency band. FIG. 7B elaborates on aspects of step 306.

[0068] Referring to FIG. 7B, step 306 can encompass steps 308-314. Step 308 of process 300 can include applying a Fourier Transform operation to theelectrophysiological information resulting in a frequency-domain representation of the electrophysiological information (where the frequency-domain representation is used to examine changes in power across different frequency bands). Step 310 of process 300 can include evaluating a change in power for frequencies present within the electrophysiological information relative to frequencies present within previous electrophysiological information for a previous time interval.

[0069] Step 312 of process 300 can include identifying an onset of the quantitative biomarker for migraine of the individual based on a decrease in power for frequencies below the gamma frequency band relative to the previous electrophysiological information. Conversely, step 314 of process 300 can include identifying an offset of the quantitative biomarker for migraine of the individual based on an increase in power for frequencies present within the electrophysiological information that are below the gamma frequency band. Process 300 can also include quantifying a relative severity of the migraine based on decrease or increase in power for frequencies below the gamma frequency band relative to the previous electrophysiological information.

[0070] Referring back to FIG. 7A, following implementation of step 306, step 308 includes applying one or more control signals to a stimulating electrode positionable along an occipital nerve of the individual based on the quantitative biomarker. This can include starting, stopping, or adjusting stimulation parameters based on the electrophysiological information. For example, if a migraine is detected based on the electrophysiological information, then the system 100 can initiate acute short-term or long-term ONS at intermittent intervals until restoration of cortical activity (increase in power for frequencies present within the electrophysiological information that are below the gamma frequency band). Upon identification of increase in power for frequencies present within the electrophysiological information that are below the gamma frequency band, the system 100 can reduce or cease ONS stimulation entirely to avoid overstimulation.Description of ExamplesEXAMPLE / : Differential Modulation of Gamma Band Activity in EEG During Migraine in Rodents

[0071] Referring to FIGS. 3A-5 as well as FIGS. 8A-8D, 9A, 9B, and 10A-10C in a widely used rodent model of chronic migraine, in which migraine was induced in rats using an intraperitoneal injection of 10-15 mg / kg of nitroglycerine, we performed quantitative electrophysiological and behavioral assessments in awake, behaving animals. All experiments were performed on a male Sprague Dawley rat weighing 370 g. The animal underwent a craniotomy where four stainless steel screws were used for EEG recording. Two of those screws were placed over the S1 region of the barrel field (placed bilaterally, 5.5 mm lateral and 1.5-2 mm posterior to bregma) on either side of the brain and one screw was used as a reference, and the last was used as a ground. After giving sufficient time for the animal to heal, the animals underwent series of experiments over four days. On the first day the animal was injected saline. The following three days, 15mg / kg, 13 mg / kg and finally 10 mg / kg of NTG was injected. In all the four days, the animal was first acclimatized to a 20 cm by 13 cm clear cage and an hour of EEG baseline recording was performed, followed by injection of the test drug and continuation of recording for at least 4 hours following the injection. The implanted device which includes four 1 .6 mm stainless steel screws connected to a pedestal connector used to record EEG activity from the primary somatosensory cortex on both hemispheres. Another cohort of rats were trained in the orofacial pain assessment device (OPAD) to assess facial hypersensitivity (a widely used behavioral marker of migraine). After two-weeks of training in the OPAD, the rodents were injected with 10 mg / kg of nitroglycerine i.p. In the OPAD device, the rat’s face makes contact with the spiked bars that provides an aversive facial mechanical stimulus as they are receiving condensed milk. All experiments were approved by the Animal Care and Use Committee (IACUC) at Arizona State University. Spectral analysis of EEG was performed using multitapered power spectral analysis in Matlab. The EEG results show that immediately after injecting a high dose of NTG (>=10 mg / kg), there is a suppression in the power of all the EEG frequencies below 30 Hz, while an enhancement in the power of the gamma band (30-55 Hz). The duration of these changes are dependent on the dose of nitroglycerine. The OPAD data shows that when the animals are injected with nitroglycerin at a 10 mg / kg dose, there is a significant reduction in the number of licks at 30 min, but not at 2 hrs and 4 hrs. This suppression suggests that NTG causes reduced motor activity which could be used as a behavioral phenotype ofmigraine. This data also agrees with previously reported data in female rats when injected with nitroglycerin (10 mg / kg).

[0072] In other words, FIGS. 3A-5 and 8A-10C illustrate example aspects of a novel quantitative biomarker for migraine that is real-time from electrophysiological measurements that can be made using current wearable devices. The quantitative biomarker tracks the onset and offset of migraine and also the degree of migraine pain. This biomarker has been validated in rodent models of migraine and to our knowledge has not been reported in humans or awake, behaving animal models. Prior studies in rodents have been performed in the anesthetized state where anesthesia can confound pain assessments. Current approaches for measuring are either qualitative (self-reported) or involve molecular biomarkers (which need cumbersome instrumentation and are not real-time). This inventive concept can provide a quantitative assessment that will be useful for both diagnosis and development of therapeutic approaches.

[0073] Examples of the inventive concept can include one or more wearable devices such as a wireless, injectable neurostimulator to stimulate the occipital nerve as a treatment for migraine.

[0074] FIG. 8A shows a Rat in a 20 cm by 13 cm clear cage used for experiments involving ONS, EEG and behavioral measurements. FIG. 8B shows an implanted device on the left-hand side which included four 1 .6 mm stainless steel screws connected to a pedestal connector used to record EEG activity from the primary somatosensory cortex on both hemispheres. FIG. 8B also shows on the right-hand side platinum iridium electrodes for unilateral ONS (used for Example II). FIG. 8C shows a Rat in the orofacial pain assessment device (OPAD) used to assess facial hypersensitivity due to nitroglycerin. FIG. 8D shows an empty OPAD cage showing spiked bars and a feeding tube behind it. The rat’s face makes contact with the spiked bars that provides an aversive facial mechanical stimulus as they are receiving condensed milk. All experiments were approved by the Animal Care and Use Committee (IACUC) at Arizona State University.

[0075] FIG. 2 shows: multi-tapered power spectrograms obtained from EEG data collected over a span of several hours before and after i.p. injection of (A) 15 mg / kg nitroglycerine, (B) 13 mg / kg nitroglycerine, and (C) 10 mg / kg nitroglycerine, (D) Saline. The graphs below each spectrogram illustrates changes inpower in the different EEG bands including Delta (0.5-4 Hz), Theta (4-8 Hz), Alpha (8-14 Hz), Beta (14-30 Hz), Gamma (30-55 Hz).

[0076] FIGS. 9A and 9B show: Effect of different doses of nitroglycerin (15 mg / kg, 13 mg / kg, 10 mg / kg) — used to induce migraine — and saline (control) on the duration of power suppression (FIG. 9A) and power enhancement (FIG. 9B) in the five EEG bands.

[0077] FIGS. 10A-10C show: Behavioral assessments on n=6 animals using the orofacial pain assessment device (OPAD) used to quantify degree of headache pain following a drug injection. Assessments were performed over several days. Each day the animals were either not injected (light grey and black bars) or I.P. injected with saline (dark grey bars) or 10 mg / kg nitroglycerin (red bars). Following injection, three, 10-min OPAD sessions were then performed at 30 minutes (FIG. 10A), 2 hrs (FIG. 10B) and 4 hours (FIG. 10C) following injection.

[0078] The EEG results of the foregoing figures show that immediately after injecting a high dose of NTG (>=10 mg / kg), there is a suppression in the power of all the EEG bands below 30 Hz, while an enhancement in the power of the gamma band (30-55 Hz). The duration of these changes is dose dependent.

[0079] The OPAD data shows that when the animals are injected with nitroglycerin at a 10 mg / kg dose, there is a significant reduction in the number of licks at 30 min, but not at 2 hrs and 4 hrs. This suppression suggests that NTG causes reduced motor activity which could be used as a behavioral phenotype of migraine. This data also agrees with previously reported data in female rats when injected with nitroglycerin (10 mg / kg).

[0080] In short, the figures demonstrate a successfully ability to quantify brain state changes due to intraperitoneal injection of nitroglycerin at three different doses and understand the behavior correlate for those changes.EXAMPLE II: Occipital Nerve Stimulation (ONS) using injectable, wireless neurostimulators to cause electrophysiological changes in rodent models of migraine

[0081] A novel wireless, injectable neurostimulator was developed and tested to stimulate the occipital nerve as a treatment for migraine. Induced cutaneous facial allodynia was quantified using an orofacial pain assessment device (OPAD) to electronically assess facial pain. Additionally, electrophysiologicalchanges (in EEG) in the primary somatosensory cortex were assessed due to nitroglycerine (NTG) administration. Finally, the effects of ONS using the novel miniature neurostimulator on cortical activity were studied.

[0082] Description of Examples: Referring again to FIGS. 8A-8D: (A) Rat in a 20 cm by 13 cm clear cage used for occipital nerve stimulation and EEG recording experiments. (B) Two devices that were implanted in the rat: the device on the left includes four 1 .6 mm stainless steel screws used to record EEG activity from the primary somatosensory cortex on both hemispheres. On the right, platinum iridium electrodes for unilateral ONS. (C) Rat in the orofacial pain assessment device (OPAD) used to assess facial hypersensitivity due to nitroglycerin. (D) Empty OPAD cage showing spiked bars and a feeding tube behind it. The rat’s face makes contact with the spiked bars that provides an aversive facial mechanical stimulus as they are receiving a condensed milk reward.

[0083] Referring to FIGS. 11 A and 11B, example aspects of the injectable (wireless) neurostimulators are shown. For example: (A) The miniature neurostimulator device (in black) is 5 mm in length and 1.2 mm in diameter. (B) Example of the pseudo-monophasic output of the neurostimulator when powered by an ultrasound transducer showing a 20 ps burst of sinusoidal pulses running at 2 MHz. The amplitude is dependent on the acoustic power intensity delivered by an external powering transducer. (The duration of the long black bar is 5 ps, and the shorter bar on the top is 0.5 ps).

[0084] FIGS. 12A-12F show: cortical evoked responses due to occipital nerve stimulation. Electrical stimulation was performed on the left occipital nerve using 300 psec cathodic pulses at 6 V running at 0.5 Hz for 5 min and EEG recordings were obtained from the primary somatosensory cortex ipsilateral (left column) and contralateral (right column) to stimulation site. Three different stimulation scenarios were tested: Stimulation using a commercial stimulator AM systems (FIGS. 12A and 12B). Stimulation using WINS (FIGS. 12C and 12D). Stimulation while the animal is under light isoflurane anesthesia (FIGS. 12E and 12F). Each graph is showing: 1- (grey) Snippets of EEG data aligned around the stimulus artifact for each stimulation trial. 2- (black / red) average of all the stimulation trials. 3- Multi-tapered power spectrogram of the average signal.

[0085] In general, the inventive disclosure of Example II, in addition to Example I, demonstrates an ability to successfully stimulate the ON and evoke cortical activity in the S1 using the novel miniature neurostimulator described; and OPAD data showed reduction in number of licks at 30 min after NTG injection and an increase in energy in the gamma band of EEG indicating migraine.EXAMPLE llh Changes in Cortical Oscillatory Activity Induced By Migraine in a Rodent Model

[0086] Background: The nitroglycerin (NTG) rodent model of migraine has been widely used in screening the effectiveness of migraine therapeutics. Electrophysiological recordings from trigeminocervical complex in these models have also been used to understand the mechanisms of pain transduction during migraine and to assess efficacy of treatments. However, such recordings require the animals to be anesthetized throughout the experiment which can be confounding. In this study, we investigate the effect of migraine induced by NTG on cortical oscillations in electroencephalogram (EEG) in awake, behaving animals

[0087] Methods: Two separate cohorts of rats underwent craniotomies where intracranial EEG screws were implanted to perform recordings from the right and left somatosensory cortices. Cohort 1 (n=10) was injected with 13 mg / kg NTG, and a control cohort (n=10) was injected with saline. In all animals, baseline EEG recordings for 30-min was followed by injection of NTG or saline and continuous EEG recordings up to 4 hrs after injection. EEG recordings were interrupted only at 30 min, 2 hrs, and 4 hrs for a 10-min session of spontaneous behavior assessments using the open field test.

[0088] A craniotomy was performed on all the animals where a total of four stainless steel screws (1 .5 mm in shaft diameter and 2 mm to 4 mm in shaft length) were implanted for intracranial EEG recordings. Two screws were implanted over the corresponding regions for the hindlimb (HL) primary somatosensory cortex (S1) (1 mm posterior and 3 mm lateral bregma), one screw on each hemisphere. A reference screw was placed over the right secondary motor cortex region (M2) (0.5 mm lateral and 2-3 mm anterior to bregma), and a ground screw was placed 2-3 mm posterior to lambda (FIG. 13). The short distance between the active and reference electrodes was mainly chosen to reduce movement artifacts and electrical noise during awake EEG activity measurements. All screws were connected to a 6-channelpedestal (MS363, Plastics One, USA). The animals were allowed to heal for a week before experiments were performed on them.

[0089] Animals were divided into two cohorts. The first cohort (n=10) received an i.p. injection of 13 mg / kg NTG, and the second cohort (n=10) received an i.p. injection of 0.9% saline for control. In both cohorts, EEG recording as well as behavioral assessment were performed on the same animal, and the experiments lasted ~5 hours for each animal. In the experiment, an animal was first brought from the vivarium and weighed. Then, the animal was placed in a custom-fabricated EEG cage for a 30-min baseline EEG recording. Before the initiation of baseline recordings, the animal was allowed to explore and acclimatize to the EEG recording cage for 10 min. Immediately following baseline recordings, the animals were removed from the cage and intraperitoneally injected with the drug (NTG or saline) assigned. Then the animals were returned to the EEG recording cage for continuous monitoring of EEG activity for up to 4 hrs after injection. Behavioral assessments using the open field test were conducted at 30 mins, 2 hrs, and 4 hrs after injection. EEG recordings were paused, and the recording cable was disconnected before behavioral tests were performed to allow for free exploration of an open-field cage. The timeline of events happening in an experiment day is illustrated in FIG. 13.

[0090] EEG recordings were performed using a headstage (Pico16, Ripple Neuro, USA) that was connected to a pedestal implanted on the animals. Two channels of raw EEG data were sampled at 1 kHz per channel, visualized, and saved using the Scout Grapevine processor (Ripple Neuro, USA).

[0091] To study spontaneous or ongoing headache pain behaviors in rodents, we utilized the open field assay following reports suggesting it might better predict the clinical efficacy of novel migraine drugs compared to cutaneous allodynia measurements since allodynia is not present in all migraine attacks, but the experience of headache pain is always present in migraine patients (Mogil, 2009; Mogil & Crager, 2004; Strassman & Burstein, 2013). Additionally, the von-Frey filament test of allodynia is subject to experimental bias as the experimenter decides whether the withdrawal is due to stress-induced due to the application of the filament or due to external factors such as stress induced by the experiment (Caudle et al., 2020). The open field test is almost fully automated, which significantly reduces experimenter bias in data collection.

[0092] For this test we utilized a custom-made box (L:90 cm x W: 90 cm * 1-1:60 cm) made of wood, the inside of which was covered with 1 .6 mm thick polyethylene (plastic) panels (FIG. 13). The cage was sprayed in black paint, similar to (Zimcikova et al., 2017), to create color contrast between the (white) animal and the (black) cage for easier tracking of the animal behavior using the ANY-Maze software (Version 7.4, Stoelting Co.). A night vision USB camera (SVPRO, Amazon, US) equipped with infrared light was positioned on top of the box at a height that allows the camera to clearly capture all the sides of the box (FIG. 13). Video recordings of the animal behavior were performed in the dark (<2 lux) following previous groups (Melo-Carrillo & Lopez-Avila, 2013) showing reduced exploratory behavior and locomotion during light compared to dark (Klejbor et al., 2013), suggesting light as an anxiety-provoking stimulus for rodents in this assay. The behavioral measurements were conducted in 10 min sessions at 30 mins, 2 hrs, and 4 hrs following drug injection.

[0093] Several measurements were quantified by the software automatically, including distance traveled, average speed of the animal, freezing episodes defined as freezing of the whole body for at least 2 sec, and immobility episodes defined as the animal not moving from place to place for at least 5 sec. Additionally, the software automatically quantified the distance traveled and the amount of time spent in the center zone of the cage, which was defined as a 70 cm x 70 cm square at the center of the cage. Reduction in the center zone activity was used consistently in several studies as a behavioral measure of anxiety in rodents (Vurall et al., 2019). Only rearing and grooming behaviors were quantified manually in the software, where the experimenter pressed a button every time a rearing or grooming behavior was observed.

[0094] EEG data was analyzed offline using MATLAB R2021A (The MathWorks, Inc.). Multitapered spectral analyses were performed on the EEG data using the Chronux toolbox (Bokil et al., 2010), which allowed for the generation of spectrograms spanning the total duration of experiments. In this analysis, we chose a multitapering moving window size (TV) of 120 sec with a step size of 20 sec and a frequency resolution (Af) of 0.5 Hz. Given those conditions, we computed the number of Slepian tapers by first calculating the time-half bandwidth (T W), which is the product of half the bandwidth of the main lobe (Af / 2) and the window size(N)(Prerau et al., 2017). This results in 59 Slepian tapers Tapers = [2 * TW] - 1), which were used in the analysis (Prerau et al., 2017).

[0095] All spectrograms generated were decibels (dB) normalized to an average power (calculated for each frequency bin) in a 30-minute baseline recording for each animal. We then performed an exploratory analysis to identify significantly different regions in the power spectrograms between the saline and NTG conditions. This analysis used non-parametric permutation testing followed by cluster-based multiple comparison correction (Maris & Oostenveld, 2007). We performed two rounds of permutation testing with 1000 iterations each. In the first round of permutations, the mapping between condition labels and individual animal power spectrogram was shuffled this was used to build a null-hypothesis distribution of test statistics for each pixel in the spectrogram. A z-score was computed, and a significance threshold of p<0.05 was applied to each pixel. The second round of permutations was used for the cluster-based multiple comparison correction, where a null distribution was created from the maximum cluster size identified from each permutation. Clusters in this analysis are defined as statistically significant pixels that are adjacent and grouped together. Cluster sizes that exceed the ninety-five percentile threshold of the distribution of largest clusters under the null hypothesis are considered statistically significant, while clusters that are below the threshold are deemed not significant and removed (Cohen, 2014).

[0096] Furthermore, upon visual inspection of the average power spectrogram in each condition (saline or NTG) and detection of significant clusters, we identified three frequency bands with distinct power variation profiles across time, which include a low-frequency band (0.5-30 Hz), low gamma band (30-60), and high gamma band (60-100). In the low-frequency band, the traditional delta (0.5-4 Hz), theta (4-8 Hz), alpha (8-14 Hz), and beta (14-30 Hz) bands were combined since their temporal profile of changes in power were identical. T o further highlight slow changes in power in those frequency bands, we further divided the spectrogram into six-time segments, including 1-29 min, 51-85 min, 85-119 min, 141-174 min, 174- 207 min, 207-239 min. Average power over frequency bands and time segments were computed for each animal, and differences in power between conditions were analyzed for statistical significance using a two-way repeated measures analysis of variance (two-way RM ANOVA) with the Geisser-Greenhouse correction followed bya two-stage step-up method of Benjamini, Krieger, and Yekutieli false discovery rate (FDR) multiple comparison correction with a q-value threshold of 0.05.

[0097] Since this analysis involves 18 comparisons (3 frequency bands x 6-time segments), the choice of FDR to correct for multiple comparisons is more appropriate compared to other methods like Bonferroni, which could be too stringent since the significance threshold of 0.05 is divided by the number of tests (Bonferoni- adjusted p = 0.05 / 18 = 0.0027) (Lage-Castellanos et al., 2010; Nolte et al., 2004). This makes it hard to detect true differences between conditions. Additionally, not correcting for multiple comparisons by using a post hoc test like Fisher's least significant difference (LSD) as implemented by (Wang et al., 2023), increases the likelihood of false positives and assumes that each time-frequency averaged power comparison is independent, which is not the case for EEG data since neighboring time points, and frequency bands in a time-frequency plot are correlated(Cohen, 2014).

[0098] Additionally, we investigated interhemispheric coherence changes between the signals collected from the right and left HL-S1 electrodes. Coherence (C,.() is defined as the cross-spectrum (Sr() of the two signals normalized by the product of the square root of the auto-spectrum from each signal (Srr) and The magnitude of the coherence ranges from 0 to 1 , where 1indicates a maximum correlation between the frequency components of the two signals and 0 indicates complete linear independence. We normalized changes in coherence across time after drug injection to the average coherence in each frequency bin during the 30-min baseline recording. Time and frequency averaged coherence data from the saline, and NTG conditions were compared using the same false discovery rate analysis described earlier for the power spectrum analysis. However, in this analysis, the low-frequency band was broken down back to four individual bands (delta, theta, alpha, and beta bands) since the coherence profiles in those bands were different. Therefore, the statistical comparisons on the coherence data were performed on six instead of three frequency bands at six time segments.

[0099] Other than the non-parametric cluster-based permutation test, all statistical analyses were performed using Prism GraphPad (GraphPad Software, Inc). The behavioral and normalized EEG (time and frequency averaged) data showed normal distribution when analyzed through the Kolmogorov-Smirnov (KS)normality test. For the behavioral data, a 2-way RM ANOVA followed by a Holm- Sidak multiple comparison posthoc test was used to study differences between saline and NTG conditions. In all statistical analyses, a p-value less than 0.05 was considered statistically significant.

[0100] Results: NTG induced significant suppression of power in the low frequencies (<30 Hz) of cortical EEG compared to saline controls immediately after injection that lasted for 90 min after injection. On the other hand, power in low gamma band of the EEG showed a delayed, but long-lasting significant suppression compared to saline controls starting at 90 minutes and lasting at least 4 hrs after injection. In the high gamma band, however, suppression was initiated within 5 mins from NTG injection and similarly for at least 4 hrs after injection. A significant suppression in locomotor / exploratory behaviors and significant enhancement in anxiety behaviors was also observed in cohort 1 compared to control cohorts with saline infusion.

[0101] In viewing and interpreting the presented data on NTG-induced changes on cortical oscillations, it is important to note the following:

[0102] (1 ) The recorded EEG signal from one hemisphere or channel(ex: HLS1 -right hemisphere), is the difference between the EEG activity in the active electrode placed over the HLS1 region and the reference electrode placed over the secondary motor cortex (M2).

[0103] (2) The EEG data in FIGS 14A-14C, 15A-15C, 16A, and 16B, present db change in (HLS1-M2) power with respect to a 30 min average baseline power recorded before drug injection.

[0104] NTG induced significant power changes in all the frequency bands (0-100 Hz) compared to saline. The data revealed three distinct temporal profiles of changes in power that are frequency band specific, including one for low- frequency bands (0-30 Hz), one for low gamma band (30-60 Hz), and another for high gamma band (60-100 Hz).

[0105] In the low-frequency band (0-30 Hz), NTG (n= 10) induced a transient but significant suppression in power for approximately 1.5 hrs immediately following NTG injection. This suppression was more pronounced in the first 30 min, followed by gradual restoration, and finally recovery towards baseline between 1 .5 and 2 hrs of the EEG recording (Fig. FIGS. 15A-15C). On the other hand, in control animals (n=10), power in the low-frequency band was marginally enhanced over thetime course of recording following injection, where the change in power from baseline recordings in the left hemisphere electrode (LHE) varied between 0.17 and1.4 db, and in the right hemisphere electrode (RHE), the mean varied between 0.05 dB and 1 .4 dB (FIGS. 16A and 16B). Analyzing time-averaged changes in EEG powerfor the low-frequency band revealed that the suppression induced by NTG was statistically significant compared to controls in the first two time segments, 1-29 min (FDR-adjusted p=0.009 for LHE; FDR-adjusted p=0.002 for RHE) and the 51-85 min (FDR-adjusted p=0.016 for LHE; FDR-adjusted p=0.003 for RHE), but not in other time segments (FIGS. 16A and 16B). The above finding is consistent with results from the cluster-based permutation test (FIGS. 14A-14C) of EEG from the RHE, where significant suppression (p<0.05) in EEG power was observed over time points throughout the first 1 .5 hrs in the low-frequency band (0.5-30 Hz band), and for approximately 3.5 hrs for frequencies above 40 Hz.

[0106] In contrast, NTG-induced changes in the low gamma band activity (30-60 Hz) of EEG appear to be delayed with a general trend of suppression in power in both hemispheres as shown in FIGS. 16A and 16B. In control animals (n=10), we observed a marginal suppression in the gamma band power following baseline recordings. The mean change in power varied between -0.64 dB and -0.17 dB (SEM - 0.16 dB to 0.23 dB) in the LHE, and between -0.5 dB and -0.16 dB (SEM - 0.16 dB to 0.23 dB) in the RHE. On the other hand, the NTG cohort showed a suppression in gamma band power that progressively increased in the first 1 .5 hrs following injection, after which the suppression plateaued. The suppression beyond1.5 hrs was found to be statistically significant compared to controls only for the RHE, and not for LHE, in the following time segments: 85-119 min (FDR-adjusted p=0.022), 141-174 min (FDR-adjusted p=0.022), and 207-239 min (FDR-adjusted p=0.045). The cluster-based-permutation test performed on the right hemisphere electrode was unable to detect large continuous clusters of significant suppression (p<0.05), indicating that the suppression was not consistently significant throughout time but happened sporadically between 1 .5-4 hrs after injection (FIGS. 14A-14C).

[0107] Similarly, EEG power in the high gamma band (60-100 Hz) was suppressed throughout the time-course of experiments following NTG and saline administration. However, NTG induced significantly more suppression in EEG power in this band compared to saline controls. In the control cohort, the mean EEG power in this band throughout all time segments in the time-frequency averaged analysis(FIGS. 16A and 16B) varied between -1.44 dB and -0.51 dB (SEM - 0.24 dB to 0.41 dB) in the LHE, and -1.1 dB and -0.39 dB (SEM - 0.24 dB to 0.37 dB) in the RHE. However, in the NTG cohort, mean EEG power in this band throughout all time segments varied from -3.3 dB and -2.4 dB (SEM - 0.46 dB to 0.67 dB) in the LHE, and -3.3 dB and -2.3 dB (SEM - 0.5 dB to 0.66 dB) in the RHE. Compared to the control cohort, the suppression induced by NTG on the high gamma band (60-100 Hz) was found to be significant in all time segments in the RHE including 1-29 min (FDR-adjusted p=0.010), 51-85 min (FDR-adjusted p=0.022), 85-119 min (FDR- adjusted p=0.005), 141-174 min (FDR-adjusted p=0.016), 174-207 min (FDR- adjusted p=0.035), 207-239 min (FDR-adjusted p=0.022), but only significant at onetime segment, 85-119 min (FDR-adjusted p=0.018) in the LHE. Additionally, the cluster-based permutation test performed on RHE data revealed several large clusters in the high gamma band, indicating that NTG induced significant suppression (p<0.05) of EEG power was continuous, covering almost the entire time course of recording following injection.

[0108] In the NTG cohort, a noticeable suppression in interhemispheric HLS1 coherence in the theta, alpha, beta, and high gamma (60-100 Hz) bands was observed. The suppression in EEG power in the above frequency bands was, however, only found to be statistically significant in the beta band at the 51-85 mins time segment (FDR-adjusted p<0.05) (FIGS. 17A and 17B).

[0109] Changes in interhemispheric HLS1 coherence in the low gamma band showed two distinct profiles when comparing the NTG cohort to the control cohort. For the NTG cohort, coherence was suppressed throughout the post-injection EEG recordings, where the suppression was significant during the first 1 .5 hours following injection compared to controls and in the 141-174 min time segment (FDR- adjusted p<0.05) (FIGS. 17A and 17B).

[0110] Statistically significant suppression in exploratory behaviors and enhancement in anxiety behaviors were observed in the animals from cohort 1 (NTG dose of 13 mg / kg) compared to animals in cohort 2 (saline controls) (FIGS. 18A- 18H). However, generally a trend of reduced exploratory behaviors (distance and speed measurements) going from 30 min to 4 hrs was observed in both the cohort 1 and 2 animals. This is likely due to increased habituation, as the rats are being exposed to the same cage in the second and third time point measurements at 2 hrs and 4 hrs, respectively. Similarly, behaviors indicative of anxiety including time anddistance in inner zone, as well as, resting behaviors including freezing episodes and immobility time, were significantly enhanced in the animals from cohort 1 compared to saline at all time points (p <0.05). While freezing episodes were enhanced at all time point measurements in cohort 1 , it was only significantly different compared to saline at 30 mins (p <0.05). Finally, grooming behaviors were only marginally enhanced in the animals from cohort 1 compared to saline, however this enhancement was not found to be statistically significant.

[0111] Discussion: Results demonstrate consistent alterations in cortical oscillations immediately after injecting a 13 mg / kg dose of NTG where it consistently induced suppression in EEG power recorded between HLS1 and M2 electrodes in the low frequency band (0.5-30 Hz) lasting around 90 min, and suppression in the high gamma band (60-100 Hz) lasting at least 240 min. During the first 90-minute period, suppression in power of the low gamma band of EEG was not significantly different compared to saline. However, after the 90 min period, increased suppression in the low gamma band power was observed (FIGS. 16A and 16B). These changes could perhaps be due to the known vasoactive effects of NTG on intra-cranial vasculature, resulting in changes in dural and cerebral blood flow that, through neurovascular coupling, can alter the metabolic and electrical activity of cortical neurons.

[0112] Indeed, intravenous administration of NTG has been shown to cause vasodilation of both middle cerebral and middle meningeal arteries, as observed in rats using laser Doppler flowmetry (Gozalov et al., 2008). Similarly, Pryazhnikov et al. used two-photon microscopy to distinguish the vasodynamic effects of NTG on meningeal vs cortical vessels and found transient vasoconstriction of meningeal arterioles, but vasodilation of the cortical ones during the first 15 min following i.p. injection of 10 mg / kg NTG (Pryazhnikov et al., 2014). Using imaging photoplethysmography, Sokolov et al. found that 2.5 hrs after i.p. injection of NTG, positive responder rats to NTG showed vasodilation, while negative responders showed vasoconstriction (Sokolov et al., 2023). To study the temporal changes in dural and cerebral blood flow, Greco et al., using laser Doppler flowmetry, tracked blood flow changes before and after NTG injection up to 150 min (Greco et al., 2011). They observed a significant suppression in dural blood flow at 30 min after NTG injection, followed by enhancement lasting up to 150 min. The enhancement was significantly different, starting at 90 min until the end of the experiment.Interestingly, this biphasic variation in dural blood flow is consistent with the temporal variations in EEG power change that we observed in the low-frequency band (0.5-30 Hz) (FIGS. 15A-16B) in the current study. In contrast, prior studies have shown that there were no transient suppressions in the cerebral blood flow but a gradual, continuous enhancement, where the enhancement was significant from 90 min until the end of the experiment compared to controls (Greco et al., 2011). Those temporal changes in cerebral blood flow closely mimic the temporal changes observed in the low gamma activity in this current study (FIGS. 15A-16B).

[0113] Furthermore, it is known that changes in cortical neuronal oscillations are correlated to changes in cerebral blood flow. Several studies have shown that as cerebral blood flow reduces, activity in higher-frequency EEG gets suppressed, while activity in lower EEG frequencies gets enhanced (Foreman & Claassen, 2012; Mabe et al., 1982). Additionally, studies that involved simultaneous EEG-fMRI measurements found a negative correlation between blood level oxygendependent (BOLD) fMRI signal and low-frequency EEG (<30 Hz), and a positive correlation between the same and high-frequency (>30) EEG (Haufe et al., 2018; Jacques et al., 2016; Yin et al., 2023). Taken together, the decrease in power in EEG in the frequency bands below the low gamma band and a relatively unchanged power in the EEG in the low gamma band in the first 90 min after NTG might be mediated by an increase in cerebral blood flow accompanied by an increase in neuronal metabolic activity.

[0114] Besides the neuro-vascular factor, alterations in the cortical activity in response to NTG could also be mediated by thalamocortical dysrhythmia triggered by activation of trigeminal afferents that innervate dural blood vessels leading to activation and sensitization of second-order trigeminocervical and third- order thalamic neurons which leads to alterations in cortical oscillatory activity. TCD has been implicated in several neurological and psychiatric disorders, including migraine, as a model to describe abnormal changes in resting-state oscillations (Coppola et al., 2007; R. Llinas et al., 2005; R. R. Llinas et al., 1999; Schulman et al., 2011). At its core, TCD is mediated by disordered signaling between the thalamus and the cortex, resulting in abnormal alterations in low and high-frequency cortical oscillatory activity (O'Hare et al., 2023). This is hypothesized to result from an imbalance in excitatory and inhibitory neurotransmitters observed in migraine (O'Hare et al., 2023). For instance, glutamate levels in the blood and in thecerebrospinal fluid in migraine patients were found to be high during migraine, and due to its excitatory effects, it has been implicated in sensitization of neurons in the trigeminovascular pain system in migraine (Hoffmann & Charles, 2018). Similarly, nitroglycerin in rats was shown to enhance plasma glutamate levels (Gao et al., 2014), and downregulate the kynurenic pathway (Nagy-Grocz et al., 2017; Sureda- Gibert et al., 2022). Additionally, NTG has been shown to induce activation and sensitization of meningeal nociceptors (first-order neurons) (X. Zhang et al., 2013), as well as trigem inocervical complex neurons (second-order neurons) (Akerman et al., 2019). The NTG-induced neuronal activation in the trigeminal nucleus caudalis (TNC) was attenuated by the administration of Kynurenic acid (KYNA — a glutamate receptor antagonist) (Knyihar-Csillik et al., 2008). KYNA analogue 1 (KYNA-A1) administration also reversed NTG-induced hyperalgesia in rats (Greco et al., 2017). Furthermore, Lally et al. found a positive correlation between glutamate levels in the human brain and the amplitude of the stimulus-evoked gamma band power (Lally et al. , 2014). All of the above evidence taken together suggests that the observed modulation in low and high gamma activity in EEG was mediated by thalamocortical dysrhythmia that was induced by abnormal enhancement in glutamate levels resulting from NTG administration. The presence of a thalamocortical dysrhythmia would also lead to behavioral changes in rats observed in this study, such as disturbances in motor behavior and cognition (i.e., migraine pain symptoms).

[0115] An EEG study performed in rats, found that enhancement of the delta and theta band oscillations 4 hrs following i.p. injection of 10 mg / kg NTG. A high dose of valproate (a known migraine therapeutic) reversed the above changes (Y. Li et al., 2016). However, no formal statistical analyses were presented in the above-referenced study beyond these anecdotal observations. In a different study using an NTG rat model of migraine (Z. Wang et al., 2023), a monotonic increase in the local field potential (LFP) power in all frequency bands, including delta, theta, alpha, beta, and gamma, was observed after NTG administration in several regions of the brain, including the trigeminal ganglion (TG), primary visual cortex (V1 M), posterior nucleus of the thalamus (Po) and anterior cingulate cortex (ACC). In our study, contrary to their findings, power in the above bands of EEG were significantly suppressed compared to those in saline controls during the first 90 min following NTG injection. Additionally, we observed a delayed but significant suppression in the low gamma band from 90-240 min and consistent suppression in the high gammaband that started within 5 mins after injection, lasting approximately 4 hrs. Thus, we observed three distinct profile changes in power that are band-specific which the previous two studies have not reported. One reason that could explain inconsistencies between our findings reported here and prior studies could be the different choices of ground and reference electrode locations and the different recording site locations.

[0116] More interestingly, the immediate suppression in the high gamma band power and the delayed and sustained suppression in low gamma activity observed in our study could be used as an indirect biomarker of trigem inovascular activation and headache pain. Such sustained suppression in gamma activity would correlate well with the previously reported sustained hindpaw and periorbital NTG-induced hyperalgesia (Di et al., 2016; Qin et al., 2018), as well as the suppression in motor / exploratory behaviors observed at all time points in this study. Additionally, changes observed in the low gamma band correlate well with the delayed activation and sensitization of neurons along the trigeminovascular pathway (Akerman et al., 2021 ; X. Zhang et al., 2013).

[0117] Headache pain in migraine patients is associated with several symptoms that affect their productivity, such as cognitive impairment (e.g. deficit in attention and memory, executive function, and processing speed), anxiety and depression (Begasse de Dhaem & Robbins, 2022; Oh et al., 2014; Peres et al., 2017). In our study with rat models, a single injection of 13 mg / kg NTG was sufficient to induce significant suppression in exploratory / motor behaviors, and enhancement in anxiety behaviors compared to saline controls (cohort 2) (FIGS. 18A-18H). Our findings agree with several prior rodent studies that showed similar changes in behavior in the open field test after a single injection of NTG (Demartini et al., 2022; Foudah et al., 2022; Liao, Li, Chen, et al., 2019).

[0118] Grooming behavior was the only readout that was enhanced in the NTG cohort but was not statistically significant compared to saline controls at all measurement time points in our study (FIGS. 18A-18H). This behavior was shown to be significantly enhanced after NTG injection in the previously mentioned studies. The presence of intracranial implants and skull-mounted electrical connectors for EEG recordings might have influenced the lack of statistically significant difference in grooming behavior due to NTG in our study. Furthermore, disconnecting the cable connecting the EEG electrodes with an external recording system just prior to theopen field test might have triggered grooming in both the control and migraine groups, making it hard to dissect NTG-induced effects on grooming.

[0119] Conclusion: This study demonstrated that a 13 mg / kg dose of NTG induced a transient suppression of energy in the low (<30 Hz) frequencies and sustained suppression of power in the gamma band in the EEG from somatosensory cortex which also correlated with a sustained behavioral dysfunctions observed in a popular rodent model of migraine.

[0120] Nitroglycerin (NTG) is being widely used to induce migraine in rodent models. The rodent model has been validated by the fact that several proven therapies have been shown to reverse NTG-induced trigeminocervical neuronal hypersensitivity and the reduction in cognitive and motor / exploratory behaviors. However, electrophysiological studies in this rodent migraine model have been largely confined to anesthetized experiments where the effectiveness of a migraine treatment was often evaluated through single neuronal recordings from a carefully targeted population of neurons in the trigeminocervical complex. The presence of anesthesia throughout the recordings can confound the complex neuronal interactions in migraine pain processing. In this study, we utilized intracranial EEG to investigate alterations in cortical oscillations induced by NTG in awake, behaving animals. We found that NTG induced a short-lasting (90 min post-injection of NTG) but significant suppression of power in EEG in the low-frequencies (<30Hz) compared to saline controls. Power in the low gamma band (30-60 Hz) of EEG was also decreasing progressively during the first 30 min post-injection of NTG, but was only significantly suppressed compared to saline controls starting at 90 min and lasting approximately 4 hrs following injection. Additionally, significant suppression was observed in the high gamma band within the first 5 minutes after NTG administration, lasting approximately 4 hours. During the first 90 min post-NTG administration, interhemispheric HLS1 coherence was significantly suppressed in the beta band and low gamma band. Behavioral data showed a significant reduction in all exploratory behaviors and enhancement in anxiety in NTG-administered animals compared to saline controls throughout all time point measurements. The delayed but significant suppression in power in the low gamma band and the consistent suppression in the high gamma band lasting at least 4 hours of EEG corresponds with a sustained reduction in exploratory behaviors observed. Given the evidence from prior studies showing that NTG-induced delayed headache can be a result ofactivation and sensitization of the trigem inocervical complex, we speculate that suppression in power in the gamma band of EEG is due to thalamocortical dysrhythmia induced by NTG. The reversal effect of known migraine therapeutics on NTG-induced delayed alterations in gamma band cortical oscillations would be necessary to validate this claim further.EXAMPLE IV: Effect of Occipital Nerve Stimulation on Cortical Oscillatory Activity in the Nitroglycerin Rodent Model of Migraine

[0121] Occipital nerve stimulation is being used to reduce headache intensity and frequency in medically refractory chronic headache disorder patients. The mechanistic effects of occipital nerve stimulation (ONS) have been investigated in a few rodent studies mainly focusing on its effects on neuronal activity in the trigem inocervical complex (TCC). However, ONS is known to modulate the activity of supraspinal structures involved in pain processing. In this study, we investigate the indirect modulation effects of ONS on cortical oscillations in the primary somatosensory cortex in the NTG rodent model of migraine.

[0122] Stimulation electrodes were implanted subcutaneously over the greater occipital nerves in (n=10) Sprague Dawley rats. In the same rats, four intracranial EEG screws were implanted where the active electrodes were placed over the regions corresponding to the right and left hind-limb primary somatosensory cortex (HL-S1 ). Three distinct ONS regimens were studied, including prophylactic and two dosages (short and long duration) of acute, intermittent treatments. EEGs were recorded before and after NTG injection for up to 4 hours.

[0123] All ONS regimens induced an early recovery from the suppression phase that is initiated immediately after NTG administration in EEG frequency bands below 30 Hz. The duration of the suppression phase was reduced from 90 minutes to 30 minutes in animals receiving an ONS treatment. Additionally, NTG-induced alterations in the gamma band power (30-100 Hz) were not reversed by any ONS treatment. However, prophylactic ONS without feedback-based modulation was able to reverse NTG-induced suppression in low-gamma band, interhemispheric HLS1 coherence. Furthermore, spontaneous behaviors indicative of headache pain were not improved by any ONS treatment. In the prophylactic ONS treatment, more suppression in exploratory behaviors and enhancement in anxiety behaviors were observed at all time points compared to animals receivingnitroglycerin injections. The results herein indicate that future research involving ONS should take caution in choosing appropriate electrical stimulation parameters that prevent noxious stimulation of ONS that could potentiate / trigger a headache attack instead of reducing it.

[0124] Methods: Briefly, all animals underwent a two-part surgery. The first part involved a craniotomy, where a total of four (1.5 mm) holes were created, including over the right secondary motor cortex (reference electrode), right and left hind-limb primary somatosensory (HL-S1) cortices (two active electrodes), and 2-3 mm posterior to lambda (ground electrode). Then, 1.5 mm diameter blunt stainless steel screws were placed over the holes while taking caution not to overturn the screws by stopping once the screw had a good first grip with the skull or around one full turn to avoid touching the dura matter. Once the screws were placed, a small layer of bone cement was added to hold the screws along with a 6-channel plastic pedestal (MS363, Plastics One, USA) that is electrically connected to the screws. This concludes the first part of the surgery. In the second part of the surgery, we extended the midline incision created during the craniotomy, followed by retracting the skin to expose the greater occipital nerves, which are superficial and can be found close to the ear. Two 127 micro-meter thick rectangular platinum-iridium electrodes (3 mm by 1 mm) were stitched over the right and left occipital nerves. Each pair of electrodes was separated by 5 mm and placed to cover the right and left greater occipital nerves. Once the electrodes are in place, the skin is stitched back together, and a 6-channel pedestal that is electrically connected with the four stimulation electrodes is bone cemented over the skull next to the EEG recording pedestal. After finishing the surgeries, the animals were returned to the vivarium and allowed to heal for a week before participating in any experiment,

[0125] The effect of three distinct ONS regimens was studied on NTG- induced alterations in cortical oscillatory activity and spontaneous behaviors indicative of headache pain. The first ONS regimen involves a 30-minute dose of bilateral ONS administered prior to intraperitoneal injection of a 13 mg / kg NTG to study the prophylactic effects of ONS. Previously, it has been shown that a 30- minute dose of ONS, 335 ps pulse duration, and amplitude of 4 mA at a stimulation frequency of 38 Hz was able to reverse several cephalic and extracephalic allodynia induced by a chronic migraine model in rats (De La Cruz et al., 2015). Additionally, in anesthetized animals, it has been shown that very short prophylactic ONS usingclinically relevant frequencies 50-100 Hz was able to reduce both ongoing and dural- evoked activity in the spinal trigeminal nucleus (STN) (Lyubashina et al., 2017). This suppression was maximally pronounced when a higher stimulation intensity of 6 V was administered. Given these promising data in rats and comparable average stimulation parameters of effective ONS in patients reported in several studies (Slotty et al., 2015; Trentman et al., 2009), the prophylactic treatment in this study was applied using a pulse duration of 300 ps, voltage intensities 2.5-3 V, and a stimulation frequency of 55 Hz. In this prophylactic treatment regimen, each animal would participate in a 5-hour experiment that first starts by weighing the animal, followed by placing the animal in a custom fabricated recording / stimulation cage, allowing the animals to explore the cage for 10 minutes. Then, a 30-minute baseline EEG recording is performed, which is followed by a 30-minute prophylactic ONS treatment. Immediately after ONS, the animals would be intraperitoneally injected with a 13 mg / kg NTG, followed by continuous EEG recording for up to 4 hours following injection. Behavioral assessments using the open field test were performed in three 10-minute sessions at 30 minutes, 2 hours, and 4 hours time points after injection (FIGS. 19A and 19B).

[0126] A subset of animals from the same cohort (n=5) were reused to study the effects of two acute ONS regimens. In both acute regimens, the stimulation frequency was reduced to 1 Hz while keeping the stimulation intensity and pulse duration the same as in the prophylactic treatment. Unlike the prophylactic treatment; acute ONS was applied before and after injection at 30 minutes and 2 hours. In the short-duration acute intermittent ONS regimen, a 100-second stimulation over each occipital nerve was administered to the rats, where we stimulated one occipital nerve at a time. This was followed by a 10-minute open field session at the 30 minute, 2 hour and 4 hour time points following injection (FIGS. 19A and 19B). In the long duration acute ONS regimen, the stimulation duration lasted 9 minutes applied on each occipital nerve, for a total stimulation duration of 18 minutes. No behavioral measurements were performed in this specific ONS regimen. Between the prophylactic ONS treatment and the short intermittent ONS treatment, a two-week break was given for the animals, and between the short and long intermittent ONS experiments, a one-week break was given to avoid carryover effects.

[0127] Stimulation and EEG recordings were all performed in one cage that had a commutator with two 6-channel pedestals, one for electrical stimulationand the other for EEG recording. The EEG recording pedestal of the commutator was connected to a 16-channel headstage, which was interfaced with the Scout Grapevine processor (Picol 6, Ripple Neuro, USA) for visualization and recording of raw EEG traces. For occipital nerve stimulation, the second pedestal of the commutator was connected to an isolated pulse stimulator AM-systems Model 2100 (A-M systems, Sequim, WA) that produced rectangular cathodic voltage pulses. The stimulation frequency and pulse duration, as well as the total stimulation duration, were all controlled by the stimulator.

[0128] EEG data were sampled at 1 k Hz. All the data were analyzed offline using MATLAB R2021A (The MathWorks, Inc.). Multitapered spectrograms (FIG. 19B) of the spontaneous EEG data were created using the Chronux MATLAB toolbox (Bokil et al., 2010). In this analysis, we applied 59 Slepian tapers, where the moving window of the multitaper FFT analysis was 120 seconds, overlapping with a step size of 20 seconds, and the spectrogram frequency resolution was 0.5 Hz. The spectrogram data post-drug injection was further dB-normalized to the average power in each frequency bin during the 30-minute baseline recording. An average spectrogram was then computed for each tested condition using the individual spectrograms generated from each animal (FIGS. 6A-6E). To track changes in individual EEG frequency bands across the time course of the experiment, we computed power averages over six frequency bands, including delta (0.5-4 Hz), theta (4-8 Hz), alpha (8-14 Hz), beta (14-30 Hz), low-gamma (30-60 Hz) and high- gamma (60-100 Hz) (Figure 3). Those frequency band-averaged powers were further averaged over six-time segments post-injection, including 1-29 min, 51-89 min, 89-119 min, 141-179 min, 179-209 min, 209-239 min, to statistically analyze differences in slow changing EEG power across the different tested conditions (FIGS. 22A and 22B). Changes in the power of the delta, theta, alpha, and beta bands across the time course of the experiment were found to have similar profiles. Therefore, they were grouped into one frequency band and named low-frequency band (0.5-30 Hz). Additionally, based on visual inspection of mean spectrograms and mean traces of power changes in EEG bands across experiment time, we identified four windows in the spectrogram within the 0.5-20 Hz range with interesting changes in power among the different tested conditions, including at 55-72 min, 100- 116 min, 137-157 min and 182-234 min. Finally, we studied changes in interhemispheric coherence across different conditions between the right and leftHLS1 regions by normalizing the cross-spectrum of the two signals to the product of the square root of the auto-spectrum from each electrode signal (right and left HLS1 electrodes).

[0129] Spontaneous behaviors indicative of pain were tracked in 10- minute sessions using the open field test. The test was performed at 30 minutes, 2 hours, and 4 hours following nitroglycerin injection and was conducted in a custom- made wooden box (L:90 cm x W: 90 cm * H:60 cm). This test was conducted in a dark room (<2 lux) to reduce anxiety triggers that could increase variance in the behavioral measurements (Klejbor et al., 2013). All spontaneous exploratory and locomotor behaviors were recorded using a night-vision camera and were analyzed automatically by the ANY-Maze software (Version 7.4, Stoelting Co.) as described in part 1 of this study. Only two behavioral measurements, grooming, and rearing, were recorded manually through the software by pressing keyboard buttons whenever the behavior was performed. Data generated by the ANY-Maze software were plotted using Prism GraphPad (GraphPad Software, Inc).

[0130] All statistical analyses were performed in Prism GraphPad (GraphPad Software, Inc). All statistical analyses were performed in Prism GraphPad (GraphPad Software, Inc). Statistical comparisons were performed between the ONS cohort and data from a control cohort that received saline injection (n=10) and a migraine cohort that received NTG (13 mg / kg) injection, which was described in a different paper. However, this study emphasizes the changes observed in the ONS stimulation cohort, not the differences between the migraine and control cohorts. All the EEG time and frequency averaged data were analyzed using 2-way repeated measures analysis of variance (ANOVA) followed by a false discovery rate (FDR) multiple comparison corrections with a q-value threshold of 0.05 to identify whether intergroup (or between treatment) differences are statistically significant. The behavioral data from the open field test was analyzed using a 2-way RM-ANOVA followed by Tukey’s post-hoc test. A p-value less than 0.05 was considered statistically significant.

[0131] Results: Generally, the temporal profile of changes in EEG power across frequency bands in all three occipital nerve stimulation regimens was similar to that observed in the nitroglycerin cohort, yet they exhibited notable differences. In the low-frequency band (0.5-30 Hz), all three ONS regimens showed a significant suppression in power observed immediately (within the first 5 mins) afterNTG administration, followed by a period of progressive enhancement and recovery afterward (FIGS. 20A-20C and FIGS. 22A and 22B). The early suppression phase during the 1-29 min time segment in both the treatment and NTG cohorts were significantly suppressed compared to saline (FDR-adjusted p<0.05). Unlike the NTG cohort, which showed a significant suppression phase lasting the first 90 mins following injection (FDR-adjusted p<0.05), all three ONS regimens showed a much shorter significant suppression phase, which lasted the first 30 mins (FDR-adjusted p<0.05), indicating that ONS facilitated an early recovery in frequency bands below 30 Hz. Following the suppression phase, a gradual enhancement in power in the low-frequency band (0.5-30 Hz) was observed, lasting around 180 mins, and recovery afterward until the end of recordings.

[0132] We further analyzed changes in power between 0.5-20 Hz at four different time segments, highlighting the early recovery period (55-72 mins) induced by ONS and the later enhancement in power observed in three instances following the recovery period at 100-116 min, 137-157 min and 182-234 min (FIG. 21). This analysis revealed significant differences between the tested conditions only during the recovery phase (55-72 mins) and not in the other time segments. The enhancement in power during the early recovery phase (55-72 min) for all three ONS treatments was found to be statistically significant compared to saline and NTG cohorts (FDR-adjusted p<0.05) in both right and left HLS1 electrodes. Additionally, in the same time segment, the enhancement in power observed in the prophylactic ONS treatment was found to be statistically significant compared to that of the long- duration acute intermittent ONS (FDR-adjusted p<0.05) only in the left hemisphere HLS1 electrode (FIG. 21).

[0133] Similarly, in the low-gamma band (30-60 Hz), trends observed in all three ONS regimens were similar to the profile observed in the nitroglycerin cohort, with differences mainly present in the first 30 minutes following injection (FIGS. 20A-20C and FIGS. 22A and 22B). Interestingly, during the first 30 minutes, only the prophylactic ONS treatment showed an overall enhancement in the low- gamma band power, which was insignificant compared to saline or NTG cohorts. After the first 30 minutes, a progressive decrease in low-gamma band power was observed in all ONS regimens lasting the first 90 mins, similar to that observed in the nitroglycerin cohort. This suppression was sustained from 90 minutes until the end of recordings with minor fluctuations in power. In the 89-119 time segment, thesuppression in low-gamma band power observed in the prophylactic treatment group, specifically at the right hemisphere HLS1 electrode, was statistically significant compared to saline (FDR-adjusted p<0.05).

[0134] Furthermore, similar to the NTG cohort, all three ONS regimens showed a more pronounced suppression in the power of the high gamma band (60- 100 Hz) compared to the saline-control cohort throughout the time course of EEG recordings post NTG administration. The suppression observed during the 89-119 min time segment in the prophylactic and short acute intermittent ONS treatments was statistically significant compared to saline (FDR-adjusted p<0.05) in both the left and right HLS1 electrodes (FIGS. 22A and 22B). Additionally, suppression during the 141-179 time segment for the long acute intermittent ONS was statistically significant compared to saline controls in both electrodes. Furthermore, the suppression in all three ONS regimens during the last time segment (209-239 min) was statistically significant compared to saline (FDR-adjusted p<0.05) only at the right hemisphere’s HLS1 electrode.

[0135] Changes in interhemispheric HLS1 coherence between the NTG and ONS-treated cohorts were only found to be significant in the low-gamma band and not in any other band (FIGS. 23A and 23B). Overall, the changes in coherence observed in the ONS-treated cohort were similar to those seen in the NTG cohort in frequency bands below 30 Hz and the high gamma band. In the alpha and beta bands, a more pronounced suppression in interhemispheric HLS1 coherence was observed in all three ONS regimens compared to saline. The suppression in coherence observed in the prophylactic ONS-treated cohort during the first 30 minutes in the alpha band was statistically significant compared to saline (FDR- adjusted p<0.05). Furthermore, in the high-gamma band, coherence between the right and left HLS1 regions was significantly enhanced compared to saline controls during the 89-119 min time segment.

[0136] On the other hand, two opposite profiles of coherence changes were observed in the low-gamma band. An enhancement in coherence in the prophylactic and short intermittent ONS-treated cohorts, while a suppression in coherence was observed in the NTG and long intermittent ONS cohorts. The differences in the coherence change between the NTG cohort and both the prophylactic and short intermittent ONS-treated animals were statistically significant (FDR-adjusted p<0.05) during the first 30 minutes (first-time segment), and the 209-239 time segment (last time segment). Additionally, the coherence change observed in the prophylactic ONS-treated cohort during the 141 -179 min and 179-209 min time segments was statistically significant (FDR-adjusted p<0.05) compared to the suppression in coherence observed in the NTG cohort.

[0137] Generally, the prophylactic ONS treatment induced more suppression in exploratory behaviors and enhanced further anxiety behaviors in all three-time point measurements compared to the nitroglycerin cohort (FIGS. 24A- 24H). In all behavioral readouts, the effects were more prominent at 2 hours compared to the other time points. Differences between the NTG 13 mg / kg cohort and the prophylactic treatment cohort at 2 hours were statistically significant (p<0.05) in specific behavioral readouts, including distance traveled, speed in the cage, freezing episodes, and time immobile. Similarly, the short intermittent ONS administered right before the behavioral measurements reduced exploratory behavior and enhanced anxiety behaviors more than the nitroglycerin cohort. However, differences between the NTG cohort and the short intermittent ONS were not statistically significant in all behaviors at all time point measurements. Additionally, differences in grooming time were not statistically significant between all groups.

[0138] Discussion and Conclusion: In this study, we investigated the effects of occipital nerve stimulation on reversing NTG-induced alterations in cortical oscillatory activity. All three occipital nerve stimulation regimens induced shorter suppression phases in the low-frequency band (0.5 - 30 Hz), lasting ~30 minutes after injection compared to the nitroglycerin cohort, which lasted ~90 minutes. There were no significant differences in the gamma band power (30-100 Hz) but significant differences in the low-gamma band (30-60 Hz) interhemispheric HLS1 coherence when comparing the NTG 13 mg / kg cohort and the prophylactic ONS treatment cohorts. While ONS treatment in those patients aims to reduce headache intensity and frequency, a painful stimulation of the nerve or areas innervated by it could trigger or potentiate headache and pain instead of relieving it. In this investigation, both prophylactic and intermittent ONS, using stimulation parameters that were shown successful in alleviating migraine headache pain in patients, were unable to reverse NTG-induced migraine-like spontaneous behaviors. On the contrary, we saw the opposite effect: more suppression in exploratory behavior at all time points (significant only at 2 hours in specific behavioral readouts) and no substantialchanges in gamma band power profiles compared to the no-treatment nitroglycerin cohort. Additionally, we saw a shorter suppression phase / early recovery in all frequency bands below 30 Hz compared to the NTG-no treatment cohort, which could be due to ONS neuromodulation of the central structure involved in pain processing.

[0139] Note that in Example IV, the ONS stimulation parameters were not modulated based on feedback provided from the EEG information. The results indicate there is a need for careful choice of ONS stimulation parameters to ensure a non-painful stimulation of the occipital nerves, which can be achieved by including incorporating electrophysiological information for controlling ONS parameters in real-time (i.e., as in the system of FIG. 1).

[0140] It should be understood from the foregoing that, while particular embodiments have been illustrated and described, various modifications can be made thereto without departing from the spirit and scope of the invention as will be apparent to those skilled in the art. Such changes and modifications are within the scope and teachings of this invention as defined in the claims appended hereto.

Claims

CLAIMSWhat is claimed is:

1. A system, comprising: a processor in communication with a memory, the memory including instructions executable by the processor to: access electrophysiological information for an individual, the electrophysiological information being acquired by two or more measuring electrodes recording from different cortical regions in a brain of the individual including a primary somatosensory cortex, the two or more measuring electrodes being in communication with the processor; and identify a presence of a quantitative biomarker for migraine of the individual based on the electrophysiological information, the quantitative biomarker indicating onset of migraine and a degree of migraine pain of the individual based on suppression of power for frequencies present within the electrophysiological information that are below a gamma frequency band from 0.5-30 Hz and in a high gamma band from 60-100 Hz.

2. The system of claim 1 , the memory further including instructions executable by the processor to: collect the electrophysiological information including electroencephalogram data over a predetermined time period to assess brain state changes3. The system of claim 1 , the memory further including instructions executable by the processor to: apply a Fourier Transform operation to the electrophysiological information resulting in a frequency-domain representation of the electrophysiological information.

4. The system of claim 1 , the memory further including instructions executable by the processor to: evaluate a change in power for frequencies present within the electrophysiological information relative to frequencies present within previous electrophysiological information for a previous time interval.

5. The system of claim 4, the memory further including instructions executable by the processor to: identify an onset of the quantitative biomarker for migraine in the individual based on a decrease in the power for frequencies present within the electrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz relative to corresponding frequencies present within the previous electrophysiological information.

6. The system of claim 4, the memory further including instructions executable by the processor to: identify an offset of the quantitative biomarker for migraine in the individual based on an increase in the power for frequencies present within the electrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz relative to corresponding frequencies present within the previous electrophysiological information, the previous electrophysiological information correlating with presence of the quantitative biomarker for migraine for the previous time interval.

7. The system of claim 1 , the quantitative biomarker further indicating onset and a degree of migraine pain of the individual based on suppression of power in frequencies present within the electrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz.

8. The system of claim 1 , the memory further including instructions executable by the processor to:apply one or more control signals to a stimulating electrode positionable along an occipital nerve of the individual based on the quantitative biomarker.

9. A method, comprising: capturing electrophysiological information for an individual, the electrophysiological information being acquired by two or more measuring electrodes recording from different cortical regions in a brain of the individual including a primary somatosensory cortex; and identifying a presence of a quantitative biomarkers for migraine of the individual based on the electrophysiological information, the quantitative biomarker indicating onset of migraine and a degree of migraine pain in the individual based on suppression of power for frequencies present within the electrophysiological information that are below a gamma frequency band from 0.5-30 Hz and in a high gamma band from 60-100 Hz.

10. The method of claim 9, further comprising: collecting the electrophysiological information including electroencephalogram data over a predetermined time period to assess brain state changes.11 . The method of claim 9, further comprising: applying a Fourier Transform operation to the electrophysiological information resulting in a frequency-domain representation of the electrophysiological information.

12. The method of claim 9, further comprising: evaluating a change in power for frequencies present within the electrophysiological information relative to frequencies present within previous electrophysiological information for a previous time interval.

13. The method of claim 12, further comprising: identify an onset of the quantitative biomarker for migraine in the individual based on a decrease in the power for frequencies present within theelectrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz relative to corresponding frequencies present within the previous electrophysiological information.

14. The method of claim 12, further comprising: identify an offset of the quantitative biomarker for migraine in the individual based on an increase in the power for frequencies present within the electrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz relative to corresponding frequencies present within the previous electrophysiological information, the previous electrophysiological information correlating with presence of the quantitative biomarker for migraine for the previous time interval.

15. The method of claim 9, the quantitative biomarker, further indicates the onset and the degree of migraine pain of the individual based on suppression of power in frequencies present within the electrophysiological information that are below the gamma band from 0.5-30 Hz and the high gamma band from 60-100 Hz.

16. The method of claim 9, further comprising: applying one or more control signals to a stimulating electrode based on the quantitative biomarker, the stimulating electrode being positionable along an occipital nerve of the individual.

17. A system for occipital nerve stimulation (ONS) to cause electrophysiological changes related to migraine, comprising: a neurostimulator device that is configured to stimulate an occipital nerve (ON) of an individual; and a processor in communication with a memory and the neurostimulator device, the memory including instructions executable by the processor to:access electrophysiological information for the individual, the electrophysiological information being acquired by two or more measuring electrodes recording from different cortical regions in the brain including the primary somatosensory cortex of the individual; the two or more measuring electrodes being in communication with the processor and identify a presence of a quantitative biomarker for migraine of the individual based on the electrophysiological information, the quantitative biomarker indicating onset of migraine and a degree of migraine pain of the individual based on suppression of power for frequencies present within the electrophysiological information that are below a gamma frequency band from 0.5-30 Hz and in a high gamma band from 60-100 Hz; and apply one or more control signals to the neurostimulator device based on the quantitative biomarker.

18. The system of claim 17, the neurostimulator device being placed by an injection or by a skin incision or other minimally invasive procedures.

19. The system of claim 17, the memory further including instructions executable by the processor to: evaluate a change in power for frequencies present within the electrophysiological information relative to frequencies present within previous electrophysiological information for a previous time interval.

20. The system of claim 19, the memory further including instructions executable by the processor to: adjust the one or more control signals applied to the neurostimulator device based on the change in power present within the electrophysiological information.

Images