Characterization of movement based on acceleration information

Abstract

Systems and methods for providing assessments of a disease using sensor data for a subject captured during day-to-day living are provided. The sensor data includes accelerometer data and is acquired by a sensor disposed on the subject's wrist while the subject performs both active and sedentary activities. The sensor data is analyzed for one or more bouts of activity; the sensor data is filtered, optionally rotated, and Fourier transformed. Spectral analysis parameters are then calculated from the transformed data to detect disease-relevant movement patterns. Movement distribution parameters for the accelerometer data associated with the disease-relevant movement patterns are calculated and compared to one or more disease scores to evaluate a progression of the disease or an efficacy of a treatment

Description

CHARACTERIZATION OF MOVEMENT BASED ON ACCELERATIONINFORMATIONCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This application claims the priority benefit of U.S. Application No. 63 / 661,504, filed June 18, 2024, which is incorporated herein by reference in its entirety for all purposes.BACKGROUND

[0002] Every day, millions of people deal with diseases that are diagnosed, evaluated, and monitored in laboratory settings or doctor’s offices. Treatments for diseases are often administered and evaluated in similar locations, which limits the duration and efficacy of these evaluations. Clinical environments may be poor analogs of the real-world conditions in which a person experiences symptoms or effects of their disease. This may lead doctors to conclude that a disease is progressing more quickly or slowly than in reality, or may lead doctors to assess that a treatment is more or less effective than in reality.

[0003] Characterizing or diagnosing some diseases (for example, Parkinson’s) may involve analyzing an aspect of a person’s physical behavior (such as their gait, or sedentary movements) in a controlled environment. However, these environments provide a limited window for doctors and clinicians to observe symptoms that may or may not be presenting in a person during an appointment. Further, these environments may cause the person to behave differently than they would during their daily life. For example, if a person is afraid of a negative diagnosis, they may artificially attempt to control their movements to exaggerate healthy characteristics or minimize unhealthy characteristics. Thus, improvements to the diagnosis and evaluation of diseases and related treatments are needed.SUMMARY

[0004] In some aspects, the techniques described herein relate to a method for assessing a progression of a disease in a subject, the method including: A) receiving acceleration data associated with the subject, the acceleration data including acceleration values for a plurality of axes and a timestamp associated with each respective acceleration value of the acceleration values; B) analyzing the acceleration data to identify a plurality of first bouts of sustainedmovement; C) calculating an activity threshold to characterize each bout of the plurality of first bouts; D) calculating a power spectrum for each bout of the first bouts; E) calculating, based on the power spectrum and for each bout of the first bouts, at least one spectral analysis parameter for that bout; F) comparing the at least one spectral analysis parameter to respective thresholds to identify a plurality of second bouts from the plurality of first bouts; G) calculating, for the plurality of second bouts and based on F), one or more bout distribution parameters for respective distributions of mobility parameters for the plurality of second bouts; H) comparing the one or more bout distribution parameters to one or more disease scores; and I) administering at least one of a treatment or a therapeutic based on H).

[0005] In some aspects, the techniques described herein relate to a method, wherein the one or more disease scores include at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0006] In some aspects, the techniques described herein relate to a method, wherein the one or more bout distribution parameters include at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0007] In some aspects, the techniques described herein relate to a method, wherein the disease includes Parkinson’s disease.

[0008] In some aspects, the techniques described herein relate to a method, wherein the activity threshold in C) is determined based on a mode of the acceleration data.

[0009] In some aspects, the techniques described herein relate to a method, wherein the mode of the acceleration data includes a mode of a substantially symmetrical portion of the acceleration data.

[0010] In some aspects, the techniques described herein relate to a method, wherein the first bouts include a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

[0011] In some aspects, the techniques described herein relate to a method, wherein the duration is between about 2 seconds and about 12 seconds.

[0012] In some aspects, the techniques described herein relate to a method, further including applying a correction to the acceleration data such that acceleration values for a first axis of the plurality of axes are substantially aligned with a vector of gravity prior to calculating the power spectrum.

[0013] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation calculated based on the power spectrum for each bout of the first bouts.

[0014] In some aspects, the techniques described herein relate to a method, wherein F) includes comparing, for each pitch-warping autocorrelation, at least one of an autocorrelation maximum of that pitch-warping autocorrelation, an autocorrelation minimum of that pitch-warping autocorrelation, or a frequency corresponding to the autocorrelation maximum of that pitchwarping autocorrelation to respective thresholds.

[0015] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 3 Hz to about 12 Hz.

[0016] In some aspects, the techniques described herein relate to a method, wherein F) includes comparing the GINI index to a GINI threshold to identify the plurality of second bouts.

[0017] In some aspects, the techniques described herein relate to a method for evaluating a progression of a disease in a subject based on time-stamped acceleration data associated with the subject, the method including: A) receiving, from a sensor, the time-stamped acceleration data associated with the subject during a first time period; B) analyzing the time-stamped acceleration data to identify a plurality of first bouts of sustained movement above a movement threshold by the subject; C) calculating, for each bout of the first bouts, at least one mobility parameter associated with a corresponding bout, the at least one mobility parameter based on at least one spectral analysis parameter associated with a power spectrum of each bout; D) identifying a plurality of second bouts associated with disease-relevant movement patterns based on the at least one spectral analysis parameter and calculating one or more bout distribution parameters based on the at least one mobility parameter for the plurality of second bouts; E) correlating the one or more bout distribution parameters with at least one disease score to determine the progression of the disease; and F) identifying, based on E), at least one of a treatment or therapeutic to be administered to the subject.

[0018] In some aspects, the techniques described herein relate to a method, wherein the disease includes Parkinson’s disease.

[0019] In some aspects, the techniques described herein relate to a method, wherein C) further includes calculating a frequency corresponding to a local maximum amplitude of a Fourier transformation of the time-stamped acceleration data corresponding to each bout of the first bouts.

[0020] In some aspects, the techniques described herein relate to a method, wherein C) further includes calculating, for each bout of the first bouts, a movement frequency associated with a maximum power of the power spectrum.

[0021] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation.

[0022] In some aspects, the techniques described herein relate to a method, wherein the movement threshold in B) is determined based on a mode of the time-stamped acceleration data.

[0023] In some aspects, the techniques described herein relate to a method, wherein the mode of the time-stamped acceleration data includes a mode of a substantially symmetrical portion of the time-stamped acceleration data.

[0024] In some aspects, the techniques described herein relate to a method, wherein the first bouts include a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

[0025] In some aspects, the techniques described herein relate to a method, wherein the duration is between about 2 seconds and about 12 seconds.

[0026] In some aspects, the techniques described herein relate to a method, wherein the sensor captures the time-stamped acceleration data at a capture frequency of at least 6 Hz.

[0027] In some aspects, the techniques described herein relate to a method, wherein the capture frequency is at least 25 Hz.

[0028] In some aspects, the techniques described herein relate to a method, wherein B) includes calculating one or more deviations of the time-stamped acceleration data and a corresponding local maximum associated with each of the one or more deviations.

[0029] In some aspects, the techniques described herein relate to a method, wherein the one or more deviations includes a mean absolute deviation.

[0030] In some aspects, the techniques described herein relate to a method, wherein B) further includes: Bl) determining at least one correction to the time-stamped acceleration data to orient the time-stamped acceleration data with an axis of a second coordinate system based on a gravity vector; B2) selecting one axis of the second coordinate system based on at least one portion of the time-stamped acceleration data; B3) converting the time-stamped acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data; and B4) identify the at least one bout of sustained movement based on the converted movement data.

[0031] In some aspects, the techniques described herein relate to a method, wherein the one or more bout distribution parameters includes at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0032] In some aspects, the techniques described herein relate to a method, wherein the at least one disease score includes at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0033] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes an autocorrelation associated with the power spectrum of each bout.

[0034] In some aspects, the techniques described herein relate to a method, wherein the autocorrelation includes a pitch-warping autocorrelation.

[0035] In some aspects, the techniques described herein relate to a method, wherein the disease-relevant movement patterns include walking patterns.

[0036] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 3 Hz to about 12 Hz.

[0037] In some aspects, the techniques described herein relate to a method, wherein the disease-relevant movement patterns include tremor patterns based on the GINI index exceeding a threshold.

[0038] In some aspects, the techniques described herein relate to a method for characterizing a movement in a subject, the method including: A) receiving acceleration data, the acceleration data including acceleration values for a plurality of axes and a timestamp associated with a respective acceleration value of the acceleration values; B) identifying a plurality of first bouts of movement indicated by the acceleration data, the first bouts including non-overlapping portions of the acceleration data centered around respective local maxima; C) identifying a plurality of second bouts from the plurality of first bouts based on comparison to a mean absolute deviation threshold; D) calculating a power spectrum for each bout of the second bouts; E) calculating, for each bout of the second bouts, at least one spectral analysis parameter based on the power spectrum for each bout of the second bouts; F) identifying, from the second bouts and based on the at least one spectral analysis parameter, a plurality of third bouts indicating disease-relevant movement patterns; G) calculating one or more bout distribution parameters for the third bouts, the one or more bout distribution parameters representing a distribution of one or more mobility parameters associated with each bout of the third bouts, wherein the one or more mobility parameters are associated with the movement; and H) classifying, based on the one or more bout distribution parameters, the movement of the subject as normal or abnormal.

[0039] In some aspects, the techniques described herein relate to a method, further including I) correlating the one or more bout distribution parameters with at least one disease score to characterize the movement.

[0040] In some aspects, the techniques described herein relate to a method, wherein the at least one disease score is associated with Parkinson’s disease.

[0041] In some aspects, the techniques described herein relate to a method, wherein the at least one disease score includes at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0042] In some aspects, the techniques described herein relate to a method, wherein the one or more bout distribution parameters includes at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0043] In some aspects, the techniques described herein relate to a method, wherein B) further includes: Bl) determining at least one correction to the acceleration data to orient the acceleration data with an axis of a second coordinate system based on a gravity vector; B2) selecting one axis of the second coordinate system based on at least one portion of the acceleration data; and B3) converting at least a portion of the acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data.

[0044] In some aspects, the techniques described herein relate to a method, wherein the acceleration data is received from a sensor.

[0045] In some aspects, the techniques described herein relate to a method, wherein the sensor captures the acceleration data at a capture frequency of at least 6 Hz.

[0046] In some aspects, the techniques described herein relate to a method, wherein the capture frequency is at least 25 Hz.

[0047] In some aspects, the techniques described herein relate to a method, wherein the first bouts include a contiguous portion of the acceleration data having a duration greater than or equal to a threshold.

[0048] In some aspects, the techniques described herein relate to a method, wherein the duration is between about 2 seconds and about 12 seconds.

[0049] In some aspects, the techniques described herein relate to a method, wherein the movement includes a gait of the subject.

[0050] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation based on the power spectrum for each bout of the second bouts.

[0051] In some aspects, the techniques described herein relate to a method, wherein the disease-relevant movement patterns include walking patterns identified based on the pitchwarping autocorrelation.

[0052] In some aspects, the techniques described herein relate to a method, wherein the second bouts include bouts exceeding the mean absolute deviation threshold.

[0053] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 3 Hz to about 12 Hz.

[0054] In some aspects, the techniques described herein relate to a method, wherein the disease-relevant movement patterns include tremor patterns identified based on the GINI index exceeding a threshold.

[0055] In some aspects, the techniques described herein relate to a method, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 4 Hz to about 7 Hz.

[0056] In some aspects, the techniques described herein relate to a method, wherein the second bouts include bouts below the mean absolute deviation threshold.

[0057] In some aspects, the techniques described herein relate to a system for assessing a progression of a disease in a subject, the system including: a processor; and a memory containing instructions thereon, the instructions configuring the processor to: A) receive acceleration data associated with the subject, the acceleration data including acceleration values for a plurality of axes and a timestamp associated with each respective acceleration value of the acceleration values; B) analyze the acceleration data to identify a plurality of first bouts of sustained movement; C) calculate an activity threshold to characterize each bout of the plurality of first bouts; D) calculate a power spectrum for each bout of the first bouts; E) calculate, based on the power spectrum and for each bout of the first bouts, at least one spectral analysis parameter for that bout; F) compare the at least one spectral analysis parameter to respective thresholds to identify a plurality of second bouts from the plurality of first bouts; G) calculate, for the plurality of second bouts and based on F), one or more bout distribution parameters for respective distributions of mobility parameters for the plurality of second bouts; H) compare the one or more bout distribution parameters to one or more disease scores; and I) administer at least one of a treatment or a therapeutic based on H).

[0058] In some aspects, the techniques described herein relate to a system, wherein the one or more disease scores include at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0059] In some aspects, the techniques described herein relate to a system, wherein the one or more bout distribution parameters include at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0060] In some aspects, the techniques described herein relate to a system, wherein the disease includes Parkinson’s disease.

[0061] In some aspects, the techniques described herein relate to a system, wherein the activity threshold in C) is determined based on a mode of the acceleration data.

[0062] In some aspects, the techniques described herein relate to a system, wherein the mode of the acceleration data includes a mode of a substantially symmetrical portion of the acceleration data.

[0063] In some aspects, the techniques described herein relate to a system, wherein the first bouts include a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

[0064] In some aspects, the techniques described herein relate to a system, wherein the duration is between about 2 seconds and about 12 seconds.

[0065] In some aspects, the techniques described herein relate to a system, the instructions further configuring the processor to apply a correction to the acceleration data such that acceleration values for a first axis of the plurality of axes are substantially aligned with a vector of gravity prior to calculating the power spectrum.

[0066] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation.

[0067] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 3 Hz to about 12 Hz.

[0068] In some aspects, the techniques described herein relate to a system for evaluating a progression of a disease in a subject based on time-stamped acceleration data associated with the subject, the system including: a processor; and a memory containing instructions thereon, the instructions configuring the processor to: A) receive, from a sensor, the time-stamped acceleration data associated with the subject during a first time period; B) analyze the time- stamped acceleration data to identify a plurality of first bouts of sustained movement above a movement threshold by the subject; C) calculate, for each bout of the first bouts, at least onemobility parameter associated with a corresponding bout, the at least one mobility parameter based on at least one spectral analysis parameter associated with a power spectrum of each bout; D) identify a plurality of second bouts associated with disease-relevant movement patterns based on the at least one spectral analysis parameter and calculating one or more bout distribution parameters based on the at least one mobility parameter for the plurality of second bouts; E) correlate the one or more bout distribution parameters with at least one disease score to determine the progression of the disease; and F) identify, based on E), at least one of a treatment or therapeutic to be administered to the subject.

[0069] In some aspects, the techniques described herein relate to a system, wherein the disease includes Parkinson’s disease.

[0070] In some aspects, the techniques described herein relate to a system, wherein C) further includes calculate a frequency corresponding to a local maximum amplitude of a Fourier transformation of the time-stamped acceleration data corresponding to each bout of the first bouts.

[0071] In some aspects, the techniques described herein relate to a system, wherein C) further includes calculate, for each bout of the first bouts, a movement frequency associated with a maximum power of the power spectrum.

[0072] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation.

[0073] In some aspects, the techniques described herein relate to a system, wherein the movement threshold in B) is determined based on a mode of the time-stamped acceleration data.

[0074] In some aspects, the techniques described herein relate to a system, wherein the mode of the time-stamped acceleration data includes a mode of a substantially symmetrical portion of the time-stamped acceleration data.

[0075] In some aspects, the techniques described herein relate to a system, wherein the first bouts include a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

[0076] In some aspects, the techniques described herein relate to a system, wherein the duration is between about 2 seconds and about 12 seconds.

[0077] In some aspects, the techniques described herein relate to a system, wherein the sensor captures the time-stamped acceleration data at a capture frequency of at least 6 Hz.

[0078] In some aspects, the techniques described herein relate to a system, wherein the capture frequency is at least 25 Hz.

[0079] In some aspects, the techniques described herein relate to a system, wherein B) includes calculating one or more deviations of the time-stamped acceleration data and a corresponding local maximum associated with each of the one or more deviations.

[0080] In some aspects, the techniques described herein relate to a system, wherein the one or more deviations includes a mean absolute deviation.

[0081] In some aspects, the techniques described herein relate to a system, wherein B) further includes: Bl) determine at least one correction to the time-stamped acceleration data to orient the time-stamped acceleration data with an axis of a second coordinate system based on a gravity vector; B2) select one axis of the second coordinate system based on at least one portion of the time-stamped acceleration data; B3) convert the time-stamped acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data; and B4) identify the at least one bout of sustained movement based on the converted movement data.

[0082] In some aspects, the techniques described herein relate to a system, wherein the one or more bout distribution parameters includes at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0083] In some aspects, the techniques described herein relate to a system, wherein the at least one disease score includes at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0084] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes an autocorrelation associated with the power spectrum of each bout.

[0085] In some aspects, the techniques described herein relate to a system, wherein the autocorrelation includes a pitch-warping autocorrelation.

[0086] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes calculating a GINI index from the power spectrum within a frequency range of about 3 Hz to about 12 Hz.

[0087] In some aspects, the techniques described herein relate to a system for characterizing a movement in a subject having a disease, the system including: a processor; and a memory containing instructions thereon, the instructions configuring the processor to: A) receive acceleration data, the acceleration data including acceleration values for a plurality of axes and a timestamp associated with a respective acceleration value of the acceleration values; B) identify a plurality of first bouts of movement indicated by the acceleration data, the first bouts including non-overlapping portions of the acceleration data centered around respective local maxima; C) identify a plurality of second bouts from the plurality of first bouts based on comparison to a mean absolute deviation threshold; D) calculate a power spectrum for each bout of the second bouts; E) calculate, for each bout of the second bouts, at least one spectral analysis parameter based on the power spectrum for each bout of the second bouts; F) identify, from the second bouts and based on the at least one spectral analysis parameter, a plurality of third bouts indicating disease-relevant movement patterns; G) calculate one or more bout distribution parameters for the third bouts, the one or more bout distribution parameters representing a distribution of one or more mobility parameters associated with each bout of the third bouts, wherein the one or more mobility parameters are associated with the movement; and H) classify, based on the one or more bout distribution parameters, the movement of the subject as normal or abnormal.

[0088] In some aspects, the techniques described herein relate to a system, the instructions further configuring the processor to: I) correlate the one or more bout distribution parameters with at least one disease score to characterize the movement.

[0089] In some aspects, the techniques described herein relate to a system, wherein the at least one disease score is associated with Parkinson's disease.

[0090] In some aspects, the techniques described herein relate to a system, wherein the at least one disease score includes at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

[0091] In some aspects, the techniques described herein relate to a system, wherein the one or more bout distribution parameters includes at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

[0092] In some aspects, the techniques described herein relate to a system, wherein B) further includes: Bl) determine at least one correction to the acceleration data to orient the acceleration data with an axis of a second coordinate system based on a gravity vector; B2) select one axis of the second coordinate system based on at least one portion of the acceleration data; and B3) convert the acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data.

[0093] In some aspects, the techniques described herein relate to a system, wherein the acceleration data is received from a sensor.

[0094] In some aspects, the techniques described herein relate to a system, wherein the sensor captures the acceleration data at a capture frequency of at least 6 Hz.

[0095] In some aspects, the techniques described herein relate to a system, wherein the capture frequency is at least 25 Hz.

[0096] In some aspects, the techniques described herein relate to a system, wherein the first bouts include a contiguous portion of the acceleration data having a duration greater than or equal to a threshold.

[0097] In some aspects, the techniques described herein relate to a system, wherein the duration is between about 2 seconds and about 12 seconds.

[0098] In some aspects, the techniques described herein relate to a system, wherein the movement includes a gait of the subject.

[0099] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes a pitch-warping autocorrelation.

[0100] In some aspects, the techniques described herein relate to a system, wherein the second bouts include bouts exceeding the mean absolute deviation threshold.

[0101] In some aspects, the techniques described herein relate to a system, wherein the at least one spectral analysis parameter includes a GINI index within a frequency range of about 3 Hz to about 12 Hz.

[0102] In some aspects, the techniques described herein relate to a system, wherein the second bouts include bouts below the mean absolute deviation threshold.

[0103] All combinations of the foregoing concepts and additional concepts discussed in greater detail below (provided such concepts are not mutually inconsistent) are part of the inventive subject matter disclosed herein. In particular, all combinations of claimed subject matter appearing at the end of this disclosure are part of the inventive subject matter disclosed herein. The terminology used herein that also may appear in any disclosure incorporated by reference should be accorded a meaning most consistent with the particular concepts disclosed herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0104] The skilled artisan will understand that the drawings primarily are for illustrative purposes and are not intended to limit the scope of the inventive subject matter described herein. The drawings are not necessarily to scale; in some instances, various aspects of the inventive subject matter disclosed herein may be shown exaggerated or enlarged in the drawings to facilitate an understanding of different features. In the drawings, like reference characters generally refer to like features (e.g., functionally similar and / or structurally similar elements).

[0105] FIG. 1 illustrates a sensor for collecting movement data from a subject.

[0106] FIG. 2 illustrates a block diagram of components of a system including a sensor and a device communicatively coupled to the sensor.

[0107] FIG. 3 illustrates acceleration data collected by a suitable accelerometer.

[0108] FIG. 4 illustrates a process for identifying a plurality of bouts of activity from accelerometer data.

[0109] FIG. 5 illustrates a plot of acceleration data as a function of value occurrence.

[0110] FIG. 6 illustrates a plot of unprocessed tri-axial accelerometer measurements with active zones and sedentary zones highlighted.

[0111] FIG. 7 illustrates successive signal processing plots in accordance with the present technology.

[0112] FIG. 8 illustrates a qualitative example of a pitch-warping autocorrelation.

[0113] FIG. 9 illustrates a graph of expected correlation between scaled and unsealed power spectra for a range of scaling factors between 1.5 and 2.5.

[0114] FIG. 10 illustrates a matrix of correlations between movement distribution parameters and PD characterization scores that may be utilized and / or indicated by acceleration data collected and analyzed in accordance with the present technology.

[0115] FIG. 11 is a flowchart of an example method for assessing progression of a disease.

[0116] FIG. 12 is a flowchart of an example method for evaluating progression of a disease.

[0117] FIG. 13 is a flowchart of an example method for characterizing movement in a subject.DETAILED DESCRIPTION

[0118] The present technology is directed toward systems and methods for evaluating Parkinson’s disease (PD) and other diseases (especially diseases impacting mobility), or for other purposes such as evaluating athletic movements (to recommend improvements, etc.), in a subject using sensor data for gait and other movement analysis. In traditional diagnostic and evaluation settings, a progression of and treatment for Parkinson’s in a subject is assessed in a clinical or doctor’s office setting. This has the drawback of significantly limiting the amount of time that clinicians and doctors have to observe a patient, which may fail to fully capture the symptoms that a subject is experiencing. However, recent developments in wearable devices and similar technology have enabled effectively continuous collection of data indicative of disease symptoms (e.g., PD tremors), baseline biometric data (e.g., heart rate, blood pressure), and activity level. Data collected by these wearable devices may provide additional insight into progression of a disease, efficacy of treatment, and / or steps to improve a subject’s quality of life. However, raw data collected by these devices is often messy, uncategorized, and of limited use without further processing. Additionally, because each subject is different, data collected by these devices may vary significantly from subject to subject. Drawing actionable conclusions from these data may therefore be non-trivial.

[0119] In an aspect, a subject with PD may wear an accelerometer on their wrist while engaging in day-to-day activities. The accelerometer may record accelerations in three dimensions as the subject moves, in particular as the subject walks. This data, representing (among other characteristics) a subject’s movement patterns, may then be analyzed for “bouts” of sustained movement activity, which may include active bouts or sedentary bouts occurringfor a threshold duration such as 5-8 seconds or more. A subject may have an accelerometer pattern associated with their gait or tremors, which may change as the subject’s PD progresses or as a treatment improves the subject’s PD. The present technology includes algorithms and methods for identifying these movement patterns (including both gait and tremor patterns) and changes over time and allowing correlations to be drawn between a subject’s movement characteristics and a status of their PD.

[0120] These correlations may be used to develop a better understanding of PD and other diseases, both in general as well as specifically for a subject with PD. Further, these improved understandings may enable personalized treatments, more efficient clinical trials, and lowered costs of care and medicine for subjects with a variety of diseases. In general, the techniques described herein may enable meaningful biomarkers of disease progression to be gleaned from high-dimensional and noisy sensor data.

[0121] FIG. 1 illustrates a sensor 110 for collecting movement data from a subject 120. Sensor 110 may be worn by subject 120 on their body (e.g., on a wrist, torso, ankle, head, etc.) or otherwise disposed on subject 120 (e.g., in a clothing pocket). Sensor 110 may include one or more individual sensors (e.g., an accelerometer, camera, cardiac sensor, pulse oximeter, blood pressure sensor, or the like) configured to measure movement or other biometric parameters. In particular, sensor 110 may be configured to record, detect, and / or measure time-stamped movement data, such as three-axis acceleration data, during everyday living, e.g., while the subject is ambulatory, at rest, going about their normal routine, engaging in day-to-day activities, etc., thereby capturing both active movement patterns and resting movement patterns. In some embodiments, the measured three-axis acceleration(s) (which may include all, or a portion of, the time-stamped movement data) may be relative to a reference frame of sensor 110 or may be relative to a reference frame of the earth, another device (e.g., a smartphone, tablet, computer, heart rate sensor, Bluetooth™ device, laptop computer, desktop computer, appliance, smart device [e.g., a smart watch, ankle sensor, or the like]).

[0122] A sensor 110 in accordance with the present technology may include any suitable wearable device, smart watch, smartphone, health monitor, etc., such as a GENEActiv™ from Activinsights, Fitbit™ from Google, and / or the like.

[0123] As subject 120 engages in daily activities (e.g., walks, sits, rests), sensor 110 will record the movement patterns of their body in accordance with a location on the subject’s body on which sensor 110 is disposed. The movement patterns may originate from voluntarymovements (e.g., walking pr brushing teeth) and / or involuntary movements (e.g., healthy movements like coughing or shivering and disease-relevant movements such as tremor). Sensor 110 may record data 130, which may include acceleration data measured by an accelerometer of sensor 110. Data 130 may be analyzed to identify both gross motor movements (such as walking) (which may be classified as active bouts) and fine motor movements (such as tremor) (which may be classified as sedentary bouts, although fine motor movements can occur simultaneously with gross ones as well) depending on the movement patterns present in the data. For example, data 130 may include acceleration values (e.g., acceleration data points) in three axes (x, y, and z) relative to sensor 110. Data 130 may further include timestamps for each acceleration data point to indicate when each data point was recorded and enable analysis of data 130.

[0124] Sensor 110 may record data over any suitable time period, including seconds, minutes, hours, days, weeks, months, or years based on available data storage, power, a willingness of subject 120 to wear the sensor 110, and / or other factors. Sensor 110 may be attached to subject 120 on a wrist, an arm, a torso, a leg, or any suitable portion of the subject’s body. Sensor 110 may be communicatively coupled with another device (e.g., device 220 shown in FIG. 2) via one or more wireless communication protocols such as Wi-Fi, Bluetooth™, NFC, Zigbee, radio transmission, cellular transmission, or any suitable protocol.Sensor Device

[0125] FIG. 2 illustrates a block diagram of components of a system 200 including sensor 210 and a device 220 that may be communicatively coupled to sensor 210. Sensor 210 may include one or more of an accelerometer 214, a cardiac sensor 216, temperature sensor, humidity sensor, conductivity sensor, and / or other suitable sensor(s). An accelerometer 214 of sensor 210 may record acceleration data in three orthogonal dimensions (e.g., triaxially), for example represented in Cartesian coordinates (x, y, and z). In an embodiment, a plurality of accelerometers may be utilized by system 200. Any measured and / or recorded data described herein may be stored in any suitable data structure, such as a list, array, tuple, matrix, dictionary, or the like. A data structure may contain a plurality of elements containing data, each element having a corresponding index by which the data contained in each element may be accessed or referenced.

[0126] Sensor 210 may be mounted by subject 120 manually, and analogous body parts such as wrist shape and / or a location on a subject which a sensor 210 may be worn (e.g., wrist, ankle,torso, etc.) may vary from subject to subject. Accordingly, an orientation of each axis of a worn accelerometer 214 may vary as well. An accelerometer may measure acceleration in its own reference frame; accordingly, the unprocessed or directly captured accelerometer data for each axis may not match a corresponding axis of a person’s body or of the ground.

[0127] Acceleration data 230 in x, y, and z dimensions may be measured by accelerometer 214 or received by a processor of device 220 from another source. The acceleration data 230 may include a plurality of data points, each data point having a measured acceleration value 232 and an acceleration timestamp data 234. Acceleration timestamp data 234 may represent time in a standardized format such as, for example, Greenwich Mean Time (GMT) or local clock time. A sampling frequency at which the acceleration data 230 is measured is determined (e.g., by calculating a difference between timestamps of temporally adjacent data points or by receiving the sampling frequency directly) and a magnitude r of acceleration at each time point is calculated by the processor as r = ^x2+ y2+ z2.

[0128] A time data structure (e.g., an array, list, tuple, etc.) of relative time values may be created by a processor 222 of the device 220 and stored on memory module 224 of device 220. The time data structure may contain values starting at 0 for the earliest timestamp and increment by any suitable increment (e.g., 1 (sampling frequency in Hz), an average difference between consecutive timestamps, a specified difference between consecutive timestamps, etc.). The time data structure may contain data points from a first timestamp (e.g., 0) until the last timestamp. A time data structure may use any suitable time denomination, such as seconds, milliseconds, microseconds, nanoseconds, or the like.

[0129] Acceleration timestamp data 234 (which may be GMT, local clock time, or other time zone or format) may be replaced by the relative time values of the time data structure. Alternatively, the relative time values of the time data structure may be added to acceleration timestamp data 234. Each relative time value may be matched with the measured acceleration value 232 and magnitude r of acceleration to which it corresponds.

[0130] For example, if a first data point of acceleration data 230 was measured at 10 AM on a first day and the last data point of acceleration data 230 was measured at 10 AM five days later, the relative time value of the first data point may be 0 and the relative time value of the last data point may be 5 days x 24 hours / day x 60 minutes / hour x 60 seconds / minute x 1000 milliseconds / second = 232,000,000 in milliseconds. Each relative time value of the time data structure may be incremented by a predetermined value divided by the sampling frequency inappropriate units. For example, each relative time value of the time data structure may be incremented by (1000 sampling frequency (milliseconds)), for example 20 milliseconds for a sampling frequency of 50 Hz. A relative time value may be incremented by about 1 ps, about 5 ps, about 10 ps, about 20 ps, about 50 ps, about 100 ps, about 200 ps, about 500 ps, about 1 ms, 2 ms, about 5 ms, about 10 ms, about 20 ms, about 50 ms, about 100 ms, about 200 ms, about 500 ms, about 1 s, or any suitable increment. Accordingly, in the above example, data recorded over 232,000,000 milliseconds may have a corresponding time data structure having 8,640,000 elements.

[0131] In some instances, measurement data 212 may be missing some values, e.g., if a communicative connection between the sensor 210 and device 220 is interrupted, if measurement data 212 is corrupted, if measurements by sensor 210 were interrupted, if sensor 210 lost power, or the like. Measurement data 212 may include second sensor data 240 including one or more second sensor values 242 and corresponding second sensor timestamps 244.

[0132] Sensor 210 may include a processor 218 such as a central processing unit (CPU), field- programmable gate array (FPGA), application-specific integrated circuit (ASIC), programmable logic controller (PLC), or any suitable microprocessor configured to manage one or more functions of sensor 210 including power management and distribution, wireless communication, data recording, startup, firmware operation, or the like.

[0133] Sensor 210 may further include one or more memory modules 208 that may be used to store data recorded by accelerometer 214. An exemplary memory module 208 may include flash memory, solid-state memory, or other suitable memory configured to temporarily store data recorded by sensor 210 until the data can be transferred to another device 220 or storage location such as a distributed storage, a cloud storage, a server, a desktop computer, a smartphone, or other storage location.

[0134] One or more processors and memory modules described herein may be used to perform any, some, or all of the steps, methods, or actions outlined herein. For example, any of the method steps or blocks, calculations, determination, identifications, etc., outlined herein may be performed by processor 218 and / or processor 222, or any other suitable processor.

[0135] Sensor 210 may include one or more communication modules 206 configured to transmit and receive data to and from another device 220 (such as a smartphone, smart watch, Bluetooth™-enabled intermediary device, tablet, wearable device, remote storage device, dataacquisition and control module, or the like). For example, a communications module 206 may include an antenna and signal generator. The signal generator may generate radio, Bluetooth™, Wi-Fi, or other signals and may use the antenna to transmit measurement data 212 to another device. Sensor 210 may transmit data periodically, or substantially continuously as it is collected, or may store data in a memory module until a threshold amount of data is stored, for example 1 megabyte (MB). Upon reaching the threshold amount of stored data, sensor 210 may transfer the data using one or more communications modules 206 to device 220. Device 220 may include one or more communications modules 226 configured to receive measurement data 212 transmitted by sensor 210.

[0136] Sensor 210 may transmit data to device 220 continuously unless a communicative connection to device 220 is interrupted or weakened below some predetermined threshold. When a communicative connection to device 220 is interrupted, sensor 210 may store data until the communicative connection to device 220 is restored, until no more memory is available on sensor 210, or may continuously overwrite the oldest stored data if no more memory is available on sensor 210.

[0137] Each sensor of sensor 210 may collect multiple data points each second. For example, accelerometer 214 may collect data points at a sampling frequency of about 1 Hz, about 2 Hz, about 5 Hz, about 10 Hz, about 12 Hz, about 15 Hz, about 20 Hz, about 25 Hz, about 30 Hz, about 50 Hz, about 75 Hz, about 100 Hz, about 125 Hz, about 150 Hz, about 200 Hz, about 250 Hz, about 500 Hz, between about 1 Hz and about 5 Hz, between about 2 Hz and about 10 Hz, between about 5 Hz and about 12 Hz, between about 10 Hz and about 15 Hz, between about 12 Hz and about 20 Hz, between about 15 Hz and about 25 Hz, between about 20 Hz and about 30 Hz, between about 25 Hz and about 50 Hz, between about 30 Hz and about 75 Hz, between about 50 Hz and about 100 Hz, between about 75 Hz and about 125 Hz, between about 100 Hz and about 150 Hz, between about 125 Hz and about 200 Hz, between about 150 Hz and about 250 Hz, between about 200 Hz and about 500 Hz, or any suitable sampling frequency.

[0138] System 200 may collect data substantially continuously subject to real -world constraint such as data storage space, power, a user removing and then reattaching the sensor 210, and / or the like. System 200 may collect data from sensor 210 at a predetermined sampling frequency for each sensor of sensor 210 (e.g., about 50 Hz for an accelerometer) for a plurality of days, weeks, months, years, etc., at a time. For example, system 200 may collect data fromsensor 210 for 7 consecutive days, which may correspond to hundreds of millions of data points and multiple gigabytes (GB) of data.

[0139] FIG. 3 illustrates approximately 3.5 minutes of unprocessed acceleration data 330 collected by a suitable accelerometer such as accelerometer 214. Acceleration data 330 may be relative to a reference frame of sensor 110 (e.g., not relative to an Earth reference frame or other gravity-based reference frame), or may be relative to an Earth reference frame or other gravity-based reference frame. Acceleration data 330 includes several bouts 332 (including bout 332a, bout 332b, bout 332c, bout 332d, and bout 332e). A bout may be a set of acceleration data points separated by no more than a threshold duration (e.g., no two sequential data points are separated by more than the threshold duration such as 100 ms, 500 ms, 1 s, 10 s, 30 s, 60 s, 500 s, 5 minutes, or any suitable threshold duration) and may include groupings of repetitive accelerometer values indicating one or more particular motor behaviors, including both active behaviors like walking and sedentary behaviors like tremor, and lasting approximately 3-20 seconds, though longer bouts are possible. Additionally or alternatively, a bout may include a set of temporally contiguous acceleration values (e.g., each acceleration value in the bout may be sampled immediately before or immediately after another acceleration value in the bout). Bouts can be computationally identified as further described herein with respect to FIG. 4. The acceleration data 330 may be collected at any suitable sampling frequency such as 50 Hz. In an embodiment, acceleration data 330 may be collected by sensor 110 or other suitable device.

[0140] In a non-limiting example, a bout may have a duration of about 1 second, about 2 seconds, about 3 seconds, about 4 seconds, about 5 seconds, about 8 seconds, about 10 seconds, about 12 seconds, about 15 seconds, about 20 seconds, about 25 seconds, about 30 seconds, about 45 seconds, about 60 seconds, between about 1 second and about 3 seconds, between about 2 seconds and about 5 seconds, between about 3 seconds and about 8 seconds, between about 5 seconds and about 10 seconds, between about 8 seconds and about 12 seconds, between about 10 seconds and about 15 seconds, between about 12 seconds and about 20 seconds, between about 15 seconds and about 25 seconds, between about 20 seconds and about 30 seconds, between about 25 seconds and about 45 seconds, between about 30 seconds and about 60 seconds, between about 2 seconds and about 12 seconds, between about 3 seconds and about 10 seconds, or any suitable duration.

[0141] Acceleration data 330 may be normalized to g (e.g., measured acceleration normalized relative to the acceleration due to Earth’s gravity, approximately 9.81 m / s2). In some nonlimiting examples, acceleration data 330 may vary between about 10 g and -10 g. A magnituder of acceleration may be equal to 1 g when sensor 110 is at rest, e.g., only measuring an acceleration due to gravity. Acceleration data 330 may include acceleration values for a plurality of axes, e.g., three axes (x, y, and z in Cartesian coordinates). During rest periods, while the overall magnitude may be close to 1 g due to gravity, small variations in this magnitude may indicate tremor or other involuntary movements that are clinically relevant. For example, small variations in this magnitude may correspond to a maximum mean absolute deviation that is below an activity threshold such as is described herein (e.g., in the description of FIG. 5).

[0142] FIG. 4 illustrates a process for identifying a plurality of 5 -second bouts of activity from accelerometer data. Plot 410 shows several minutes of unprocessed tri-axial accelerometer measurements 412 having amplitudes in a range from about -0.5 g to about 0.5 g. Unprocessed tri-axial accelerometer measurements 412 may be collected using a suitable sensor or sensor array such as sensor 110 and / or accelerometer 214. Unprocessed tri-axial accelerometer measurements 412 may be filtered using a low pass filter to remove noisy transient data.

[0143] Because a large number of data points may be collected over a period of, e.g., days, performing calculations over an entire dataset may be computationally intensive and / or infeasible. To improve computational efficiency, reduce an analytical duration, and reduce an overall number of datapoints to be analyzed, 5 -second bouts of activity are identified. These bouts may include either active bouts (e.g., gross motor activities such as walking) or sedentary bouts (e.g., fine motor activities such as tremor). These bouts are identified by first calculating the MAD of the Euclidean norm of the unprocessed tri-axial accelerometer measurements 412. MAD is defined by the following equation:where T is the number of data points (and a rolling measurement with a window of 5 seconds) and r(ti) is the magnitude of acceleration at time ti. For example, unprocessed tri-axial accelerometer measurements 412 are input into the above equation to produce rolling MAD 422 in plot 420, which may include one or more local maxima. These local maxima are then used to identify periods of activity which may include 5 -second bouts of walking. A moving average and / or lowpass filtered version of the unprocessed tri-axial accelerometermeasurements 412 may additionally or alternatively be computed to determine an absolute deviation of the acceleration data from a baseline.

[0144] To identify these 5 -second bouts of walking, a rolling maximum of the rolling MAD is calculated. Plot 430 illustrates rolling MAD 422 and a rolling maximum 434 of the rolling MAD 422. To calculate rolling maximum 434, a value of rolling maximum 434 at each time point will be the maximum value of rolling MAD 422 within a window (e.g., a 5-second window) with the time point centered at that window. For example, using a 5-second window, a rolling maximum 434 value at a time point will be the maximum value of rolling MAD 422 between 2.5 seconds before the time point and 2.5 seconds after the time point. This is repeated for each time point of rolling MAD 422.

[0145] A window used to calculate a rolling maximum may be any suitable duration, including about 1 second, about 2 seconds, about 3 seconds, about 4 seconds, about 5 seconds, about 8 seconds, about 10 seconds, about 12 seconds, about 15 seconds, about 20 seconds, about 25 seconds, about 30 seconds, about 45 seconds, about 60 seconds, between about 1 second and about 3 seconds, between about 2 seconds and about 5 seconds, between about 3 seconds and about 8 seconds, between about 5 seconds and about 10 seconds, between about 8 seconds and about 12 seconds, between about 10 seconds and about 15 seconds, between aboutl2 seconds and about 20 seconds, between about 15 seconds and about 25 seconds, between about 20 seconds and about 30 seconds, between about 25 seconds and about 45 seconds, between about 30 seconds and about 60 seconds, or any suitable window duration.

[0146] Active bouts 436 highlight identified periods within non-overlapping windows of predetermined duration (e.g., 5 seconds) of activity that may be analyzed for walking and gait patterns. These walking and gait patterns may then be correlated with one or more disease scores, as explained further herein. In an embodiment, a duration of non-overlapping windows may be selected such that both active and sedentary bouts are identified. For example, active bout 436a illustrates an active bout where a maximum mean absolute deviation value is relatively high (e.g., above an activity threshold), whereas bout 436b illustrates a sedentary bout where a maximum mean absolute deviation is relatively low (e.g., below the activity threshold).

[0147] An active bout may differ from a sedentary bout in that a sedentary bout may include one or more of a lower average or maximum acceleration values, lower frequency movements, lower amplitudes of an associated power spectrum, and / or the like. An active bout may beassociated with bulk, full-body movements such as walking, running, riding a bike, or climbing stairs, while a sedentary bout (e.g., an “inactive” bout) may be associated with smaller, more limited movements such as tapping fingers while sitting, scrolling on a phone, playing a piano, flossing, pathological movements such as tremor, or the like. In some cases, active bouts can be associated with gross motor skills, while sedentary bouts can be associated with fine motor movements. Gross motor movements may be movements that are performed using large muscle groups in the trunk, legs, and arms; fine motor movements may be movements that are performed using smaller muscle groups in the forearms, wrists, hands, and fingers. As PD progresses in a subject, they will typically be less mobile, and the frequency and intensity of active bouts tends to decline.

[0148] FIG. 5 illustrates a plot 510 of log2(MAD) vs. count number of values for rolling MAD 422. Plot 510 includes Log MAD distribution 512, which illustrates a number of occurrences of log2(MAD) for each value of rolling MAD 422 and may be multi-modal. In particular, plot 510 may indicate that a majority of rolling MAD values are relatively small (e.g., most values of log2(MAD) are less than -8). These relatively small values of rolling MAD may be used to differentiate between active and sedentary bouts of movement.

[0149] A mode 514, which may indicate a most commonly occurring value of log MAD distribution 512, may be calculated. Values of log MAD distribution 512 below mode 514 may be mirrored about mode 514 to indicate a substantially symmetric portion of log MAD distribution 512 indicating “sedentary” activity, which may correspond to non-walking movements and therefore may be excluded from calculations of gait and (in some cases) disease characterization. In some non-limiting embodiments, non-walking movements (e.g., tremor) may be identified using the techniques described herein. These sedentary measurements 526 are indicated in plot 520 as a darkened distribution on the left portion of plot 520. Active measurements 530 are indicated as the portion of MAD distribution 512 above an activity threshold 528. Activity threshold 528 may be used to identify a separation between active and sedentary bouts of movement. In an aspect, an activity threshold 528 may be calculated as a value at which a predetermined percentile of sedentary measurements 526 are below. The value at which a predetermined percentile of sedentary measurements 526 are below may be a percentile; for example, activity threshold 528 may be the value of sedentary measurements 526 at which 98% of values in sedentary measurements 526 are smaller.

[0150] For example, activity threshold 528 may be a 98thpercentile of sedentary measurements 526, a 99thpercentile of sedentary measurements 526, a 95thpercentile ofsedentary measurements 526, a 90thpercentile of sedentary measurements 526, an 85thpercentile of sedentary measurements 526, an 80thpercentile of sedentary measurements 526, a 75thpercentile of sedentary measurements 526, a 50thpercentile of sedentary measurements 526 (which may correspond to mode 514), or any suitable percentile.

[0151] Additionally or alternatively, a value of log2(MAD) may be selected as an activity threshold 528. For example, an activity threshold 528 may be selected to have a value of -8, - 8.1, -8.2, -8.3, -8.4, -8.5, -8.6, -8.7, -8.8, -8.9, -9.0, 9.1, -9.2, -9.3, -9.4, -9.5, or any suitable value.

[0152] FIG. 6 illustrates a plot 610 of unprocessed tri-axial accelerometer measurements 412 with active zones 620 and sedentary zones 630 highlighted based on calculations outlined with respect to FIGS. 4 and 5. In particular, active zones 620 and sedentary zones 630 may be determined by analyzing a multi-modal distribution such as MAD distribution 512 and differentiating between active and sedentary zones based on activity threshold 528. Each value of unprocessed tri-axial accelerometer measurements 412 may fall into either an active zone 620 or a sedentary zone 630. Values falling into a sedentary zone 630 may be discarded.

[0153] Once all active and sedentary bouts within unprocessed tri-axial accelerometer measurements 412 have been identified, one or more of the active bouts may be corrected or rotated such that one axis will, on average, capture the acceleration due to gravity. For each active bout, the corresponding accelerometer values for the active bout may be converted to spherical coordinates. Each time point of unprocessed tri-axial accelerometer measurements 412 will include an x acceleration, a y acceleration, and a z acceleration and an associated timestamp t. To convert to spherical coordinates, the following parameters are calculated for each timestamp t: a radius r(t) =inclination angle 9(t) =arctan(y(t) / x(t)), and azimuthal angle (p(t) = arctan(— ). For each active bout, anV (t)2+y(t)2average of 9 and cp are calculated. Then, a correction is calculated and applied to the accelerometer measurements for that active bout such that on average, 9 = 99° and cp = 0°. The result is a 3 x T matrix (where T is a total number of recorded timestamps) where the bulk of the acceleration due to gravity is aligned with the y-axis for that active bout.

[0154] FIG. 7 illustrates successive signal processing plots 710, 720, 730 in accordance with the present technology. Once accelerometer data such as unprocessed tri-axial accelerometer measurements 412 has been processed, one or more characteristic frequencies associated with movement patterns (such as gait or tremor patterns) may be identified in the filteredaccelerometer data 712. Filtered accelerometer data 712 may be analyzed for one or more bouts of sustained movement activity above a threshold. A Fourier transform may then be applied to bouts of sustained movement activity to produce a power spectrum. The power spectrum may indicate which frequencies appear the most in the accelerometer data as well as an overall power associated with active bouts (including walking bouts). Without being limited by any theory in particular, spectral characteristics of movement bouts (including power for active bouts such as walking bouts, and frequency concentration for sedentary bouts, such as tremor bouts) may change (e.g., power may decrease) over time as a disease such as PD progresses, and the present technology may provide methods to monitor such a progression as well is detect and evaluate the effects of a treatment or medication.

[0155] Plot 710 shows filtered accelerometer data 712 (which may be a magnitude r(t) of tri- axial accelerometer data x(t), y(t), and z(t)) within a 90 second window after input accelerometer data has been low pass filtered. Filtering may remove transient spikes or high- frequency noise from input accelerometer data to produce accelerometer data 712 and make it easier to extract a signal pertaining to a gait of a subject such as subject 120.

[0156] Plot 720 shows a portion of the accelerometer data 712 illustrating a sedentary bout 712a corresponding to a tremor in the subject — that is, sedentary bout 712a is a tremor bout. This accelerometer data representing sedentary bout 712a has several repeated patterns of varying frequencies that may be quantified using a Fourier transform.

[0157] Plot 730 is a graph of a Fourier transform of sedentary bout 712a. The Fourier transform of sedentary bout 712a shows that a spectral amplitude has local maxima 734a, 734b, and 734c at about 4 Hz, about 8 Hz, and about 12 Hz, respectively, both of which are emblematic of PD tremor. The frequencies associated with the local maxima may indicate an activity (or lack thereof) that a subject is engaged in based on the frequency peaks present in the Fourier transform. For example, Fourier transform 732 may indicate that the subject from which the acceleration was measured was experiencing a tremor while not moving.

[0158] A power spectrum may then be calculated from Fourier transform 732. A power f spectrum may be defined as R(f) = (Sn=oxne~l2n~^n)2where Ais a total number of samples (N = fs x Z, where fs is sampling rate and Z is a total duration of signal in seconds), / is the frequency at data point / / , and xnis the input acceleration value (which may be an acceleration value from an individual axis or a magnitude from two or more axes).

[0159] For each power spectrum calculated for each active bout, a pitch-warping autocorrelation is calculated. The pitch-warping autocorrelation may be used to determine if an active bout includes any well-pronounced harmonics, which can be characteristic of walking bouts and may serve to distinguish a signal associated with a walking motion from a noise signal having a first frequency from a subject’s gait having the same frequency. A pitch- warping autocorrelation may be calculated as P(s) * / ), where R( / ) is thepower spectrum as calculated above and R(.s * f) is the power spectrum scaled by the scaling factor 5.

[0160] FIG. 8 illustrates a qualitative example of a pitch-warping autocorrelation. Graph 800 shows a comparison of an unsealed power spectrum 832 of acceleration data representing a walking bout with scaled power spectrum 834 and scaled power spectrum 836. Scaled power spectrum 834 has been scaled using a scale factor 5 = 2, while scaled power spectrum 836 has been scaled using a scale factor .s = 1.5.

[0161] When harmonics are present in a signal, a scaling factor 5 equal to 2 should result in a second peak overlapping with a first peak, a fourth peak overlapping with a second peak, etc. A pure signal having a first frequency and no harmonics will have no overlap upon calculating the pitch-warping autocorrelation, and therefore will produce a small P(s) relative to a signal having harmonics for the same first frequency. For an ideal signal, non-integer scaling factors (e.g., .s = 1.4) will not produce resulting overlaps between a primary signal and associated harmonics. In reality, human biomarker signals may include harmonics that overlap at scaling factors other than 2 (e.g., in a range between about 1.5 and about 2.5). Thus, pitch-warping autocorrelations may be calculated for each active bout using a range of scaling factors between about 1.5 and about 2.5.

[0162] For example, pitch-warping autocorrelations may be calculated for each active bout using scaling factors of about 1.5, about 1.6, about 1.7, about 1.8, about 1.9, about 2, about 2.1, about 2.2, about 2.3, about 2.4, about 2.5, between about 1.4 to about 2.6, between about 1.5 to about 2.5, between about 1.7 to about 2.3, between about 1.9 to about 2.1, between about 1.4 to about 2.1, between about 1.9 to about 2.6, or any suitable scaling factors or ranges of scaling factors.

[0163] FIG. 8 illustrates a difference between scaling factors applied to an unsealed power spectrum 832. Scaled power spectrum 834 is scaled with a scaling factor s = 2 and a second frequency peak (i.e., a first harmonic frequency peak) of scaled power spectrum 834 overlapswith a first frequency peak of unsealed power spectrum 832 in region 840, indicating the presence of at least one harmonic frequency, which may in turn indicate that unsealed power spectrum 832 is representative of a walking bout.

[0164] In contrast, each peak of scaled power spectrum 836 (which is scaled using a scaling factor of .s = 1.5) does not overlap with any peak of unsealed power spectrum 832, resulting in a smaller P(s) for 5 = 1.5 than 5 = 2.

[0165] FIG. 9 illustrates a graph 900 of expected correlation between scaled and unsealed power spectra (e.g., a correlation between unsealed power spectrum 832 and scaled power spectrum 834) for a range of scaling factors between 1.5 and 2.5. Graph 900 shows that a maximum correlation between scaled and unsealed power spectra typically occurs at a scaling factor of 2, but correlations may occur for scaling factors greater than or less than 2.

[0166] For each pitch-warping autocorrelation P(s) and power spectrum R(f) calculated for each active bout, threshold metrics are calculated to determine if the active bout represents useful data (e.g., a walking bout). Threshold metrics may include a frequency at a maximum value of P(s), a ratio of the maximum value of P(s) to the minimum value of P(s), and a frequency at a maximum value of R(f) are calculated and compared to respective thresholds. If each threshold metric exceeds a corresponding threshold for the active bout, features relevant to mobility may be calculated from the acceleration data, power spectrum, and pitch-warping autocorrelation associated with the active bout.

[0167] Features relevant to mobility may include any one or more of the following:

[0168] VMC (vector magnitude count): A VMC may be defined as the mean absolute deviation of acceleration in the y-axis following a rotation of acceleration data to align y-axis accelerometer data with the gravity vector on average.

[0169] rVMC: The mean absolute deviation of the root mean square of the magnitude r(t) of acceleration following a rotation of acceleration data to align y-axis accelerometer data with the gravity vector on average.

[0170] xzVMC: The mean absolute deviation of the root mean square of the acceleration in the x-axis and z-axis (e.g., the horizontal plane in the rotated reference frame) following a rotation of acceleration data to align y-axis accelerometer data with the gravity vector on average.

[0171] R(f)maX: the maximum value of the discrete power spectrum R(f) for each active bout.

[0172] f corresponding to R(f)maX: the frequency at the maximum value of the discrete power spectrum R(f) for each active bout.

[0173] vv: the square root of the sum of the power spectra for an active bout.

[0174] Maxavg _R(f>: the maximum value of the averaged power spectra over all identified walking bouts.

[0175] N: the number of identified walking bouts.

[0176] For example, the techniques described herein may identify N = 1000 walking bouts from several days’ or weeks’ worth of acceleration data for a subject. For each of the 1000 walking bouts, VMC, rVMC, xzVMC, R(f)max, f corresponding to R(f)max, and vv are calculated, resulting in 1000 values for each metric. These 1000 values for each metric may form a distribution from which distribution parameters including mean, median, maximum, 75th, 95th, and 99thquantiles may be calculated. These distribution parameters are then correlated with Parkinson’s disease scores as illustrated in FIG. 10 to determine a status, progression, regression, and / or treatment efficacy associated with the PD for a subject.

[0177] For sedentary bouts containing tremor patterns (which may be called tremor bouts), relevant features may include any one or more of the following:

[0178] Spectral concentration measures such as GINI indices calculated within tremorrelevant frequency ranges (e.g., 3-12 Hz, or more specifically 4-7 Hz), which may quantify the concentration of spectral power within frequencies characteristic of pathological tremor.

[0179] VMC (like described above, but for tremor patterns).

[0180] rVMC (like described above, but for tremor patterns).

[0181] xzVMC (like described above, but for tremor patterns).

[0182] R(f)max(like described above, but for tremor patterns).

[0183] f corresponding to R(f)max (like described above, but for tremor patterns).

[0184] vv (like described above, but for tremor patterns).

[0185] Maxavg _R(f> (like described above, but for tremor patterns).

[0186] M: the number of identified tremor bouts.

[0187] For example, the techniques described herein may identify M tremor bouts from the same acceleration data for the subject. For each of the M tremor bouts, GINI indices calculatedwithin tremor-relevant frequency ranges may be determined, resulting in M values for each tremor-specific metric. These M values for each tremor metric may form a distribution from which distribution parameters including mean, median, maximum, 75th, 95th, and 99th quantiles may be calculated. These tremor-based distribution parameters are then correlated with Parkinson’s disease scores to determine tremor-specific aspects of the subject’s PD status, progression, regression, and / or treatment efficacy.Disease Scores

[0188] The techniques herein may be applied to characterize, evaluate, and treat a variety of disease models including Parkinson’s disease, stroke, multiple sclerosis (MS), muscular dystrophies, amyotrophic lateral sclerosis (ALS), spinal cord injuries, cerebral palsy, Huntington’s disease, myasthenia gravis, peripheral neuropathy, and obesity. Each of the aforementioned diseases may include associated motor behavior patterns such as gait characteristics (frequency, cadence, stride length, frequency of occurrences of walking, duration of occurrences of walking) and tremor characteristics (frequency, amplitude, constancy, and spectral concentration). The techniques described herein may be used to characterize these motor behavior patterns, including changes to these motor behavior patterns over time, and provide insight into a progression, regression, or status of these diseases as well as a related efficacy of therapeutics.

[0189] For example, for PD, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as levodopa, carbidopa, dopamine agonists, monoamine oxidase-B (MAO-B) inhibitors, anticholinergics, deep brain stimulation, and physical and / or occupational therapy.

[0190] For stroke, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as rehabilitation (including physical therapy, occupational therapy, and / or speech therapy), anticoagulants, antiplatelet agents, and / or an effectiveness of surgery.

[0191] For MS, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as disease modifying therapies (DMTs including oral and injectable DMTs), physical therapy, and corticosteroids.

[0192] For ALS, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as Riluzole, Edaravone, physical and occupational therapy, and breathing support such as ventilators.

[0193] For spinal cord injuries, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as physical and / or occupational therapy, medications to manage pain and spasticity, and / or surgical interventions (e.g., spine stabilization).

[0194] For cerebral palsy, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as physical therapy, occupational therapy, speech therapy, and medications to manage spasticity.

[0195] For Huntington’s disease, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as medications to address movement problems and psychiatric conditions (e.g., antipsychotics, antidepressants, antianxiety medications, etc.), physical therapy, and / or speech therapy.

[0196] For myasthenia gravis, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as cholinesterase inhibitors, immunosuppressants, plasmapheresis, intravenous immunoglobulin (IVIg), and / or thymectomies.

[0197] For peripheral neuropathy, the techniques described herein may be used to determine an effectiveness of therapeutics and / or treatments such as anti-seizure medications, antidepressants, physical therapy, and / or lifestyle changes.

[0198] Diseases may be evaluated using a score or scale. For example, PD may be evaluated based on the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS- UPDRS). The MDS-UPDRS includes a plurality of scores for a variety of aspects of PD including tremor; ease of movement; cognitive, mood, and mental characteristics; writing, walking, eating, and other daily activities; physiological characteristics; and other factors affecting a subject with PD.

[0199] The present techniques may evaluate a status of a disease in a subject by correlating a disease progression and / or therapeutic efficacy with some or all of the scores outlined in the MDS-UPDRS. A correlation between the present techniques and a standard scale provides an indication that the present techniques are providing information similar to assessments provided by the MDS-UPDRS. In particular, the following scores may be utilized:

[0200] MDS-UPDRS part I: nonmotor aspects of experiences of daily living. MDS-UPDRS part I may be assessed using a combination of secondary rater assessment of a subject and selfassessment by the subject. The subject is asked to rate 13 nonmotor aspects on a scale from 0to 4: 0 is considered “normal;” 1 is considered “slight;” 2 is considered “mild;” 3 is considered “moderate;” and 4 is considered “severe.”

[0201] The 13 nonmotor aspects for MDS-UPDRS part I are cognitive impairment, hallucinations and psychosis, depressed mood, anxious mood, apathy, features of dopamine dysregulation syndrome, sleep problems, daytime sleepiness, pain and other sensations, urinary problems, constipation problems, lightheadedness upon standing, and fatigue.

[0202] MDS-UPDRS part II: motor aspects of experiences of daily living. MDS-UPDRS partII may be assessed analogously to MDS-UPDRS part I. The subject is asked to rate 13 motor aspects related to daily living.

[0203] These 13 motor aspects indicate if a subject is experiencing problems or difficulties with: speech, saliva and drooling, chewing and swallowing, eating tasks, dressing, hygiene, handwriting, performing hobbies and other activities, turning in bed, tremor, getting out of bed, a car, or a deep chair, walking and balance, and freezing.

[0204] MDS-UPDRS part III: motor examination for motor signs of PD. MDS-UPDRS may be assessed by an examiner such as a clinician, doctor, or other disease expert. The subject is evaluated on 18 factors as observed by the examiner on a scale from 0 to 4 analogous to the scale as described with respect to MDS-UPDRS part I.

[0205] These 18 motor signs of PD evaluated in part III include: speech, facial expression, rigidity, finger tapping, hand movements, pronation-supination movements of hands, toe tapping, leg agility, ability to arise from a chair, gait (including stride amplitude, stride speed, height of foot lift, heel strike during walking, turning, and arm swing; this assessment does not include freezing, which is evaluated separately), freezing of gait (hesitation and stuttering movements, especially when turning and reaching the end of the task), postural stability in response to a forced displacement of the shoulders, global spontaneity of movement (e.g., body bradykinesia), postural tremor of the hands (evaluated based on tremor amplitude), kinetic tremor of the hands, rest tremor amplitude, and constancy of rest tremor. MDS-UPDRS partIII may further take into consideration dyskinesia impact on part III ratings as well as Hoehn and Yahr stage.

[0206] MDS-UPDRS part IV: motor complications. An examiner may determine scores for MDS-UPDRS part IV based on historical and objective information provided by a subject with PD, a caregiver for the subject, or someone familiar with the subject as well as the MDS-UPDRS assessment as a whole. MDS-UPDRS part IV includes six aspects describing motor complications.

[0207] The six aspects evaluated in MDS-UPDRS part IV include: time spent with dyskinesias (excluding OFF-state dystonia), functional impact of dyskinesias, time spent in the OFF state (where symptoms are more noticeable and movement becomes more difficult), functional impact of motor fluctuations, complexity of motor fluctuations, and painful OFF- state dystonia.

[0208] The present technology may utilize a score from the Hoehn and Yahr scale indicating a progression of PD. The Hoehn and Yahr scale indicates a plurality of stages of the disease from stage 1 to stage 5. Stage 1 indicates only one side of the body is exhibiting symptoms (i.e., unilateral involvement), typically with little to no functional disability or impairment. Stage 2 indicates bilateral involvement of symptoms with no impairment to balance. Stage 3 indicates mild to moderate bilaterial involvement including some postural instability, although the subject remains physically independent. Stage 4 indicates severe disability, though the subject retains the ability to walk and stand unassisted. Stage 5 indicates that the subject is wheelchair bound or bedridden unless aided.

[0209] The Hoehn and Yahr scale may further include stages 1.5 and 2.5. Stage 1.5 is characterized by unilateral involvement of symptoms with the addition of axial involvement. Stage 2.5 is characterized by mild bilateral disease with recovery on a pull test.

[0210] The present technology may utilize an MDS-UPDRS score on a second or subsequent clinical visit, for example a second clinical visit MDS-UPDRS part III score.

[0211] The present technology may utilize a sum of any suitable scores described herein, including a sum of MDS-UPDRS part III scores associated with postural tremor (which may be referred to herein as “posture total”), a sum of MDS-UPDRS part III scores associated with kinetic tremor (which may be referred to herein as “kinetic total”), a sum of MDS-UPDRS part III scores associated with resting tremor (which may be referred to herein as “rest total”), a constancy of resting tremor score (as scored in MDS-UPDRS part III), a tremor score limited to the upper limbs (which may be referred to herein as “limb total”), a sum of gait and freezing scores from MDS-UPDRS part III, a sum of scores related to rigidity from MDS-UPDRS part III, a sum of scores related to bradykinesia, the MDS-UPDRS part III gait score, a sum of scores related to gait from MDS-UPDRS part III (e.g., leg agility, arising from a chair, gait,and freezing of gait, and / or a sum of scores related to lower limb functioning (e.g., leg agility, arising from a chair, and gait).

[0212] FIG. 10 illustrates a matrix 1000 of correlations between movement distribution parameters 1010 described above and PD characterization scores 1020 that may be utilized and / or indicated by acceleration data collected and analyzed in accordance with the present technology.

[0213] Each cell contains a number between -100 and 100 indicating how well two metrics or scores are correlated with one another. For example, a resting tremor score indicated by rest total has a strong positive correlation of 95 out of 100 with an upper limb tremor score indicated by limb total.

[0214] The correlation between movement distribution parameters 1010 and PD characterization scores 1020 may be used to determine an efficacy of treatment or therapeutic, diagnose PD in a subject, evaluate a progression of PD over time, calculate a supplementary MDS-UPDRS score, or otherwise characterize PD in a subject using both gait and tremor analysis. Movement distribution parameters 1010 are broadly (though not always) negatively correlated with PD characterization scores 1020, indicating that a lower value of movement distribution parameters 1010 tends to correlate with a higher value (i.e., increased severity) of corresponding PD characterization scores 1020.

[0215] Matrix 1000 illustrates correlations between each movement distribution parameter of movement distribution parameters 1010 and each PD characterization score of PD characterization scores 1020. Matrix 1000 shows, for example, a relatively strong negative correlation of -0.61 between the 99thquantile of rVMC (rvmc_q99) and a sum of scores related to lower limb function (denoted lower limb function score in matrix 1000).

[0216] This may indicate that a decrease over time in rvmc_q99 may result in an increase over time in lower limb function score. An increased lower limb function score indicates that a subject is experiencing more severe symptoms related to lower limb function (e.g., more difficulty walking, balancing, controlling their legs, performing tasks related to standing such as getting out of a chair, and the like).

[0217] Acceleration data collected over time (e.g., acceleration data 330 collected over a period of six months) may be used as outlined above to indicate an efficacy of treatment. For example, a suitable processor such as processor 218, processor 222, or a similar processor may receive acceleration data from a suitable accelerometer such as accelerometer 214. Processor222 may apply a low pass filter to the acceleration data to remove high-frequency signal components. Processor 222 may then calculate a rolling MAD and rolling max of the rolling MAD to identify bouts of activity within one or more time windows (e.g., within 5-second time windows).

[0218] Once these bouts of activity have been identified, processor 222 may determine a threshold to separate active bouts from sedentary bouts, e.g., using techniques as described with respect to FIG. 5. Once active bouts have been separated from sedentary bouts, the acceleration data is rotated such that the acceleration data corresponding to the y-axis aligns with the gravity vector on average. A Fourier transform may then be applied to the rotated active bouts to generate a power spectrum. Using the power spectrum, processor 222 may identify peaks defining local maxima of the power spectrum for each active bout as well as the frequencies corresponding to the peaks.

[0219] Processor 222 may then calculate a pitch-warping autocorrelation P(s) for each active bout using a scaling factor of about 2, for example between about 1.5 and about 2.5. If P(s) indicates a presence of harmonic frequencies for an active bout, and movement metrics including A) a frequency at a maximum value of P(s), B) a ratio of the maximum value of P(s) to the minimum value of P(s), and / or C) a frequency at a maximum value of R(f) exceed respective movement thresholds, processor 222 may classify the active bout as a walking bout and calculate one or more mobility parameters for each walking bout. These mobility parameters may include VMC, rVMC, xzVMC, R(f)max, f corresponding to R(f)max, and vv.

[0220] For sedentary bouts, processor 222 may calculate a GINI index within a tremorrelevant frequency range (such as 3-12 Hz, or more specifically 4-7 Hz) and compare this to a GINI threshold to identify tremor bouts. Distribution parameters may then be calculated across identified tremor bouts in a manner similar to the active bout analysis.

[0221] Processor 222 may then form a distribution of one or more of VMC, rVMC, xzVMC, R(f)max, f corresponding to R(f)max, and vv across some or all identified walking bouts (though non-walking bouts may be used; for example, tremor bouts may be similarly analyzed), from which bout distribution parameters including mean, median, maximum, 75th, 95th, and 99th, quantiles may be calculated.

[0222] Processor 222 may then monitor changes to these bout distribution parameters over time (e.g., as a treatment or therapeutic is administered) and use the correlation scores of matrix 1000 to infer an efficacy or effectiveness of the treatment or therapeutic.

[0223] A medical expert such as a clinician or doctor may utilize these changes to bout distribution parameters over time to make more informed decisions about a progress, steadiness, and / or regression of a subject’s PD. For example, a medical expert may classify a movement of the subject as normal or abnormal based on the one or more bout distribution parameters. The analysis may incorporate both gait-based parameters (derived from active bout analysis) and tremor-based parameters (derived from sedentary bout analysis) to provide a comprehensive assessment of the subject’s condition.

[0224] A movement of the subject may be classified as normal if a bout distribution parameter has a predetermined value or falls within a predetermined range of values. Additionally or alternatively, a movement of the subject may be classified as normal based on movement characteristics such as gait frequency (e.g., a gait frequency between about 1.5 Hz and about 2.5 Hz) for walking analysis or tremor frequency and spectral concentration for tremor analysis, an average power of a walking bout above a predetermined threshold, a presence of harmonics in a pitch-warping autocorrelation of one or more walking bouts of the subject, or the like.

[0225] A movement of the subject may be classified as abnormal if a bout distribution parameter has a predetermined value or falls outside of a predetermined range of values. Additionally or alternatively, a movement of the subject may be classified as abnormal based on movement characteristics such as gait frequency (e.g., a gait frequency less than about 1.5 Hz) for walking analysis or tremor frequency and spectral concentration for tremor analysis, an average power of a walking bout below a predetermined threshold, a lack of harmonics in a pitch-warping autocorrelation of one or more walking bouts of the subject, or the like.Example Methods

[0226] FIG. 11 illustrates a flowchart of an example method 1100 for assessing progression of a disease in a subject, the method comprising blocks 1110-1185, some of which may be optional.

[0227] Block 1110 includes receiving acceleration data associated with the subject. For example, the acceleration data may comprise acceleration values for a plurality of axes and a timestamp associated with each respective acceleration value of the acceleration values. The acceleration data may be received from a sensor such as an accelerometer disposed on a subject’s wrist, torso, or other suitable location on the subject’s body. The acceleration data may be collected continuously or substantially continuously over a period of second, minutes, hours, days, weeks, months, or years during the subject’s normal daily activities.

[0228] Block 1120 includes identifying a plurality of first bouts of sustained movement from the acceleration data. The first bouts may be identified by analyzing the acceleration data for periods of activity that exceed baseline levels and have a duration greater than or equal to a threshold, such as between about 2 seconds and about 12 seconds. Each first bout may represent a contiguous portion of acceleration data, which may be centered around respective local maxima in the acceleration magnitude.

[0229] Block 1130 includes characterizing the first bouts by calculating an activity threshold to distinguish between different types of movement patterns. This characterization process may include sub-steps such as block 1131 and / or block 1132.

[0230] Block 1131 includes calculating an activity threshold based on a mode of the acceleration data. The activity threshold may be determined by analyzing a distribution of acceleration magnitudes, identifying a mode of the distribution, and calculating a threshold value (such as a 95th, 98th, or 99th percentile) that separates active periods from sedentary periods.

[0231] Block 1132 includes characterizing each bout of the plurality of first bouts based on the activity threshold. Bouts may be classified as either active bouts (exceeding the activity threshold) or sedentary bouts (below the activity threshold), with active bouts potentially corresponding to walking or other gross motor movements, and sedentary bouts potentially corresponding to fine motor movements or tremor activities.

[0232] Block 1140 may be optional and relate to active bout processing, and may include applying a correction to orient the acceleration data with a gravity vector. This step may involve determining one or more corrections to align acceleration values for a first axis of the plurality of axes substantially with a vector of gravity, thereby providing a consistent reference frame for subsequent analysis regardless of the sensor’s initial orientation on the subject’s body.

[0233] Block 1145 includes calculating a power spectrum for each bout of the first bouts. The power spectrum may be calculated using a Fourier transform of the acceleration data corresponding to each bout.

[0234] Block 1150 includes calculating a spectral analysis parameter for each bout based on the power spectrum. The method may use different paths for different types of movement analysis. Gait analysis of active bouts may include block 1151 and / or block 1152. Tremor analysis of sedentary bouts may include block 1155 and / or block 1156.

[0235] The method may include classifying windows (e.g., 5-second windows) of time as active or sedentary. This classification may be performed based on a threshold applied to MAD values, where values below the threshold may be classified as sedentary, and values above the threshold may be classified as active. This threshold may be estimated or calculated, such as is described herein.

[0236] Block 1151 includes calculating a pitch-warping autocorrelation for each bout. The pitch-warping autocorrelation may be calculated using scaling factors between about 1.5 and about 2.5 to detect the presence of harmonic frequencies characteristic of walking patterns.

[0237] Block 1152 includes determining thresholds for the autocorrelation parameters. These thresholds may be used to identify which bouts contain sufficient harmonic content to be classified as walking bouts.

[0238] For windows classified as sedentary bouts, the method may filter with a pass band, for example of 3-12 Hz. Then the method may calculate spectrograms, which may be used in block 1155.

[0239] Block 1155 includes calculating a GINI index of the Fourier transform within a frequency range of about 3 Hz to about 12 Hz (or more specifically about 4 Hz to about 7 Hz). The GINI index provides a measure of the concentration of spectral power within the tremorrelevant frequency range.

[0240] Block 1156 includes calculating a GINI threshold, for example, based on the mode method used to identify sedentary and active bouts. This threshold helps distinguish tremorcontaining bouts from other types of sedentary movement. For example, the GINI threshold may be used to identify bouts that are dominated by a small number of frequencies.

[0241] Block 1165 includes comparing the spectral analysis parameter to its respective threshold. For active bouts, this involves comparing pitch-warping autocorrelation metrics (such as autocorrelation maximum, autocorrelation minimum, or frequency corresponding to the autocorrelation maximum) to respective thresholds. For sedentary bouts, this involves comparing the GINI index to the GINI threshold.

[0242] Block 1170 includes identifying a plurality of second bouts from the plurality of first bouts based on the threshold comparisons. Second bouts may represent those first bouts that exceed their respective thresholds and are therefore classified as containing disease-relevant movement patterns (either walking patterns for active bouts or tremor patterns for sedentary bouts).

[0243] Block 1175 includes calculating one or more bout distribution parameters for the second bouts. These parameters may include statistical measures such as mean, median, maximum, quantiles (75th, 95th, 99th), root mean square, or mean absolute deviation calculated across the distributions of mobility parameters for the plurality of second bouts.

[0244] Block 1180 includes comparing the one or more bout distribution parameters to one or more disease scores. The disease scores may include UPDRS part I, II, III, or IV scores, Hoehn and Yahr scores, clinic visit scores, or specific subscores such as postural tremor scores, kinetic tremor scores, resting tremor scores, rigidity scores, bradykinesia scores, gait scores, freezing scores, or any other suitable score.

[0245] Block 1185 includes administering at least one of a treatment or a therapeutic based on the comparison with disease scores. The treatment selection may be informed by correlations between the bout distribution parameters and disease progression indicators, which may provide personalized therapeutic interventions.

[0246] FIG. 12 illustrates a flowchart of an example method 1200 for evaluating progression of a disease in a subject based on time-stamped acceleration data, the method comprising blocks 1210-1270, some of which may be optional.

[0247] Block 1210 includes receiving acceleration data, for example from a sensor. The acceleration data may be time-stamped and associated with the subject during a first time period. The sensor may capture the acceleration data at a frequency of at least 6 Hz, for example at least 25 Hz, to provide sufficient temporal resolution for movement analysis.

[0248] Block 1220 includes identifying first bouts of sustained movement above a movement threshold based on the acceleration data. This identification process may include sub-steps such as block 1221, block 1222, block 1223, block 1224, and / or block 1225.

[0249] Block 1221 includes calculating deviations and corresponding local maximum for each deviation. The deviations may include mean absolute deviations (MAD) calculated over windows of the acceleration data.

[0250] Block 1222 includes determining at least one correction to orient the time-stamped acceleration data with an axis of a second coordinate system based on a gravity vector. This may ensure consistent orientation regardless of initial sensor placement.

[0251] Block 1223 includes selecting one axis of the second coordinate system based on at least one portion of the acceleration data. The selected axis may correspond to periods of highest activity level during the measurement period.

[0252] Block 1224 includes converting the acceleration data from a first coordinate system (such as Cartesian coordinates) to the second coordinate system (such as spherical coordinates) based on the at least one correction to provide converted movement data.

[0253] Block 1225 includes identifying sustained movement bouts based on the converted movement data. These bouts may have durations between about 2 seconds and about 12 seconds and represent periods of continuous movement above the movement threshold.

[0254] Block 1230 includes calculating at least one mobility parameter for each bout of the first bouts, based on the power spectrum. For example, the at least one mobility parameter may be based on at least one spectral analysis parameter associated with a power spectrum of each bout. This process may involve different approaches for different types of movement. For active bouts, this process may include block 1231 and / or block 1232. For sedentary bouts, this process may include block 1235 and / or block 1236.

[0255] Block 1231 includes calculating a frequency corresponding to a local maximum amplitude of a Fourier transformation of the time-stamped acceleration data for each bout.

[0256] Block 1232 includes calculating a movement frequency associated with a maximum power of the power spectrum. This may provide additional characterization of the movement’s spectral properties.

[0257] Block 1235 includes calculating a GINI index of the Fourier transform within the tremor-relevant frequency range. This may quantify the concentration of spectral power within frequencies typical of pathological tremor.

[0258] Block 1236 includes calculating a GINI threshold, which may distinguish tremorcontaining bouts from other sedentary movements.

[0259] Block 1240 includes identifying second bouts from the first bouts. This identification process may use different criteria for different movement types. For active bouts, this identification may include block 1241. For sedentary bouts, this identification may include block 1245.

[0260] Block 1241 includes classifying walking bouts based on autocorrelation analysis. Bouts with sufficient harmonic content (as determined by pitch-warping autocorrelation) may be classified as walking bouts.

[0261] Block 1245 includes classifying tremor bouts based on GINI threshold comparison. Bouts with GINI indices exceeding the threshold may be classified as tremor bouts.

[0262] Block 1250 includes calculating bout distribution parameters for the second bouts, for example, based on the at least one mobility parameters. These bout distribution parameters may represent statistical distributions of mobility parameters across all identified disease-relevant bouts.

[0263] Block 1260 includes correlating the bout distribution parameters with at least one disease score to determine disease progression. This correlation analysis may reveal relationships between movement characteristics and clinical assessment scores.

[0264] Block 1270 includes identifying at least one treatment or therapeutic to be administered to the subject based on the correlation analysis. The treatment selection may be personalized based on the specific movement patterns and disease progression indicators identified through the analysis.

[0265] FIG. 13 illustrates a flowchart of an example method 1300 for characterizing movement in a subject, the method comprising blocks 1310-1390, some of which may be optional.

[0266] Block 1310 includes receiving acceleration data, which may comprise acceleration values for a plurality of axes and timestamps associated with respective acceleration values. The acceleration data may be received from a sensor disposed on the subject and may be collected during normal daily activities over extended periods.

[0267] Block 1320 includes identifying first bouts of movement based on local maxima of the acceleration data. The first bouts comprise non-overlapping portions of the acceleration data centered around respective local maxima. This identification process may include substeps such as block 1322, block 1323, and / or block 1324, some or all of which may be specific to active bout analysis.

[0268] Block 1322 includes determining corrections to orient the acceleration data with a gravity-based axis. This standardizes the coordinate system regardless of initial sensor orientation.

[0269] Block 1323 includes selecting one axis based on portions of the acceleration data. The selection may be based on identifying periods of highest activity that correspond to upright postures.

[0270] Block 1324 includes converting the acceleration data from a first coordinate system to a second coordinate system. This conversion may provide a standardized reference frame for subsequent analysis.

[0271] Block 1330 includes identifying second bouts from the first bouts based on a mean absolute deviation (MAD) threshold. The second bouts may be characterized as either exceeding the MAD threshold (active bouts) or falling below the MAD threshold (sedentary bouts), depending on the specific movement patterns being analyzed.

[0272] Block 1340 includes calculating a power spectrum for each second bout.

[0273] Block 1350 includes calculating at least one spectral analysis parameter for each second bout. These parameters may include pitch-warping autocorrelations for active bout analysis or GINI indices for sedentary bout analysis, depending on the type of movement being characterized.

[0274] Block 1360 includes identifying third bouts from the second bouts based on the spectral analysis parameters. Third bouts may represent those second bouts that contain disease-relevant movement patterns, such as walking patterns (identified through harmonic analysis) or tremor patterns (identified through spectral concentration analysis).

[0275] Block 1370 includes calculating bout distribution parameters for the third bouts. These parameters may represent distributions of one or more mobility parameters associated with each bout of the third bouts, where the one or more mobility parameters are associated with the movement (e.g., the movement patterns being analyzed).

[0276] Block 1380 includes classifying the movement of the subject as normal or abnormal based on the bout distribution parameters. The classification may be based on comparing the distribution parameters to normative values or predetermined thresholds that distinguish healthy movement patterns from pathological ones.

[0277] Block 1390 includes correlating the bout distribution parameters with disease scores to characterize the movement. This optional step may provide additional clinical context by relating the movement characteristics to established disease assessment scales, providing more detailed characterization of the subject’s condition and potential treatment needs.Conclusion

[0278] While various inventive embodiments have been described and illustrated herein, those of ordinary skill in the art will readily envision a variety of other means and / or structures for performing the function and / or obtaining the results and / or one or more of the advantages described herein, and each of such variations and / or modifications is deemed to be within the scope of the inventive embodiments described herein. More generally, those skilled in the art will readily appreciate that all parameters, dimensions, materials, and configurations described herein are meant to be exemplary and that the actual parameters, dimensions, materials, and / or configurations will depend upon the specific application or applications for which the inventive teachings is / are used. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. It is, therefore, to be understood that the foregoing embodiments are presented by way of example only and that, within the scope of the appended claims and equivalents thereto, inventive embodiments may be practiced otherwise than as specifically described and claimed. Inventive embodiments of the present disclosure are directed to each individual feature, system, article, material, kit, and / or method described herein. In addition, any combination of two or more such features, systems, articles, materials, kits, and / or methods, if such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent, is included within the inventive scope of the present disclosure.

[0279] Also, various inventive concepts may be embodied as one or more methods, of which an example has been provided. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

[0280] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0281] The terms “about,” “around,” “substantially,” “nearly,” and “approximately” shall be given their plain and ordinary meaning as would be understood by one of ordinary skill in the art given the context of their use. If not sufficiently clear from context, the terms “about,” “around,” “substantially,” “nearly,” and “approximately” shall mean + / - 10% of a stated value.

[0282] The indefinite articles “a” and “an,” as used herein in the specification and in the claims, unless clearly indicated to the contrary, should be understood to mean “at least one.”

[0283] The phrase “and / or,” as used herein in the specification and in the claims, should be understood to mean “either or both” of the elements so conjoined, i.e., elements that are conjunctively present in some cases and disjunctively present in other cases. Multiple elements listed with “and / or” should be construed in the same fashion, i.e., “one or more” of the elements so conjoined. Other elements may optionally be present other than the elements specifically identified by the “and / or” clause, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, a reference to “A and / or B”, when used in conjunction with open-ended language such as “comprising” can refer, in one embodiment, to A only (optionally including elements other than B); in another embodiment, to B only (optionally including elements other than A); in yet another embodiment, to both A and B (optionally including other elements); etc.

[0284] As used herein in the specification and in the claims, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” shall be interpreted as being inclusive, i.e., the inclusion of at least one, but also including more than one, of a number or list of elements, and, optionally, additional unlisted items. Only terms clearly indicated to the contrary, such as “only one of’ or “exactly one of,” or, when used in the claims, “consisting of,” will refer to the inclusion of exactly one element of a number or list of elements. In general, the term “or” as used herein shall only be interpreted as indicating exclusive alternatives (i.e., “one or the other but not both”) when preceded by terms of exclusivity, such as “either,” “one of,” “only one of,” or “exactly one of.” “Consisting essentially of,” when used in the claims, shall have its ordinary meaning as used in the field of patent law.

[0285] As used herein in the specification and in the claims, the phrase “at least one,” in reference to a list of one or more elements, should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed within the list of elements and not excluding any combinations of elements in the list of elements. This definition also allows that elements may optionally be present other than the elements specifically identified within the list of elements to which the phrase “at least one” refers, whether related or unrelated to those elements specifically identified. Thus, as a non-limiting example, “at least one of A and B” (or, equivalently, “at least one of A or B,” or, equivalently “at least one of A and / or B”)can refer, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including elements other than B); in another embodiment, to at least one, optionally including more than one, B, with no A present (and optionally including elements other than A); in yet another embodiment, to at least one, optionally including more than one, A, and at least one, optionally including more than one, B (and optionally including other elements); etc.

[0286] In the claims, as well as in the specification above, all transitional phrases such as “comprising,” “including,” “carrying,” “having,” “containing,” “involving,” “holding,” “composed of,” and the like are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases “consisting of’ and “consisting essentially of’ shall be closed or semi-closed transitional phrases, respectively, as set forth in the United States Patent Office Manual of Patent Examining Procedures, Section 2111.03.

Claims

CLAIMS1. A method for assessing a progression of a disease in a subject, the method comprising:A) receiving acceleration data associated with the subject, the acceleration data comprising acceleration values for a plurality of axes and a timestamp associated with each respective acceleration value of the acceleration values;B) analyzing the acceleration data to identify a plurality of first bouts of sustained movement;C) calculating an activity threshold to characterize each bout of the plurality of first bouts;D) calculating a power spectrum for each bout of the first bouts;E) calculating, based on the power spectrum and for each bout of the first bouts, at least one spectral analysis parameter for that bout;F) comparing the at least one spectral analysis parameter to respective thresholds to identify a plurality of second bouts from the plurality of first bouts;G) calculating, for the plurality of second bouts and based on F), one or more bout distribution parameters for respective distributions of mobility parameters for the plurality of second bouts;H) comparing the one or more bout distribution parameters to one or more disease scores; andI) administering at least one of a treatment or a therapeutic based on H).

2. The method of claim 1, wherein the one or more disease scores comprise at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

3. The method of claim 1, wherein the one or more bout distribution parameters comprise at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

4. The method of claim 1, wherein the disease comprises Parkinson’s disease.

5. The method of claim 1, wherein the activity threshold in C) is determined based on a mode of the acceleration data.

6. The method of claim 5, wherein the mode of the acceleration data comprises a mode of a substantially symmetrical portion of the acceleration data.

7. The method of claim 1, wherein the first bouts comprise a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

8. The method of claim 7, wherein the duration is between about 2 seconds and about 12 seconds.

9. The method of claim 1, further comprising applying a correction to the acceleration data such that acceleration values for a first axis of the plurality of axes are substantially aligned with a vector of gravity prior to calculating the power spectrum.

10. The method of claim 1, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation calculated based on the power spectrum for each bout of the first bouts.

11. The method of claim 10, wherein F) comprises comparing, for each pitch-warping autocorrelation, at least one of an autocorrelation maximum of that pitch-warping autocorrelation, an autocorrelation minimum of that pitch-warping autocorrelation, or a frequency corresponding to the autocorrelation maximum of that pitch-warping autocorrelation to respective thresholds.

12. The method of claim 1, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 3 Hz to about 12 Hz.

13. The method of claim 12, wherein F) comprises comparing the GINI index to a GINI threshold to identify the plurality of second bouts.

14. A method for evaluating a progression of a disease in a subject based on time-stamped acceleration data associated with the subject, the method comprising:A) receiving, from a sensor, the time-stamped acceleration data associated with the subject during a first time period;B) analyzing the time-stamped acceleration data to identify a plurality of first bouts of sustained movement above a movement threshold by the subject;C) calculating, for each bout of the first bouts, at least one mobility parameter associated with a corresponding bout, the at least one mobility parameter based on at least one spectral analysis parameter associated with a power spectrum of each bout;D) identifying a plurality of second bouts associated with disease-relevant movement patterns based on the at least one spectral analysis parameter and calculating one or more bout distribution parameters based on the at least one mobility parameter for the plurality of second bouts;E) correlating the one or more bout distribution parameters with at least one disease score to determine the progression of the disease; andF) identifying, based on E), at least one of a treatment or therapeutic to be administered to the subject.

15. The method of claim 14, wherein the disease comprises Parkinson’s disease.

16. The method of claim 14, wherein C) further comprises calculating a frequency corresponding to a local maximum amplitude of a Fourier transformation of the time-stamped acceleration data corresponding to each bout of the first bouts.

17. The method of claim 14, wherein C) further comprises calculating, for each bout of the first bouts, a movement frequency associated with a maximum power of the power spectrum.

18. The method of claim 17, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation.

19. The method of claim 14, wherein the movement threshold in B) is determined based on a mode of the time-stamped acceleration data.

20. The method of claim 19, wherein the mode of the time-stamped acceleration data comprises a mode of a substantially symmetrical portion of the time-stamped acceleration data.

21. The method of claim 14, wherein the first bouts comprise a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

22. The method of claim 21, wherein the duration is between about 2 seconds and about 12 seconds.

23. The method of claim 14, wherein the sensor captures the time-stamped acceleration data at a capture frequency of at least 6 Hz.

24. The method of claim 23, wherein the capture frequency is at least 25 Hz.

25. The method of claim 14, wherein B) comprises calculating one or more deviations of the time-stamped acceleration data and a corresponding local maximum associated with each of the one or more deviations.

26. The method of claim 25, wherein the one or more deviations comprises a mean absolute deviation.

27. The method of claim 14, wherein B) further comprises:Bl) determining at least one correction to the time-stamped acceleration data to orient the time-stamped acceleration data with an axis of a second coordinate system based on a gravity vector;B2) selecting one axis of the second coordinate system based on at least one portion of the time-stamped acceleration data;B3) converting the time-stamped acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data; andB4) identify the at least one bout of sustained movement based on the converted movement data.

28. The method of claim 14, wherein the one or more bout distribution parameters comprises at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

29. The method of claim 14, wherein the at least one disease score comprises at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

30. The method of claim 14, wherein the at least one spectral analysis parameter comprises an autocorrelation associated with the power spectrum of each bout.

31. The method of claim 30, wherein the autocorrelation comprises a pitch-warping autocorrelation.

32. The method of claim 30, wherein the disease-relevant movement patterns comprise walking patterns.

33. The method of claim 14, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 3 Hz to about 12 Hz.

34. The method of claim 33, wherein the disease-relevant movement patterns comprise tremor patterns based on the GINI index exceeding a threshold.

35. A method for characterizing a movement in a subject, the method comprising:A) receiving acceleration data, the acceleration data comprising acceleration values for a plurality of axes and a timestamp associated with a respective acceleration value of the acceleration values;B) identifying a plurality of first bouts of movement indicated by the acceleration data, the first bouts comprising non-overlapping portions of the acceleration data centered around respective local maxima;C) identifying a plurality of second bouts from the plurality of first bouts based on comparison to a mean absolute deviation threshold;D) calculating a power spectrum for each bout of the second bouts;E) calculating, for each bout of the second bouts, at least one spectral analysis parameter based on the power spectrum for each bout of the second bouts;F) identifying, from the second bouts and based on the at least one spectral analysis parameter, a plurality of third bouts indicating disease-relevant movement patterns;G) calculating one or more bout distribution parameters for the third bouts, the one or more bout distribution parameters representing a distribution of one or more mobility parameters associated with each bout of the third bouts, wherein the one or more mobility parameters are associated with the movement; andH) classifying, based on the one or more bout distribution parameters, the movement of the subject as normal or abnormal.

36. The method of claim 35, further comprising I) correlating the one or more bout distribution parameters with at least one disease score to characterize the movement.

37. The method of claim 36, wherein the at least one disease score is associated with Parkinson’s disease.

38. The method of claim 36, wherein the at least one disease score comprises at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

39. The method of claim 35, wherein the one or more bout distribution parameters comprises at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

40. The method of claim 35, wherein B) further comprises:Bl) determining at least one correction to the acceleration data to orient the acceleration data with an axis of a second coordinate system based on a gravity vector;B2) selecting one axis of the second coordinate system based on at least one portion of the acceleration data; andB3) converting at least a portion of the acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data.

41. The method of claim 35, wherein the acceleration data is received from a sensor.

42. The method of claim 41, wherein the sensor captures the acceleration data at a capture frequency of at least 6 Hz.

43. The method of claim 42, wherein the capture frequency is at least 25 Hz.

44. The method of claim 35, wherein the first bouts comprise a contiguous portion of the acceleration data having a duration greater than or equal to a threshold.

45. The method of claim 44, wherein the duration is between about 2 seconds and about 12 seconds.

46. The method of claim 35, wherein the movement comprises a gait of the subject.

47. The method of claim 35, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation based on the power spectrum for each bout of the second bouts.

48. The method of claim 47, wherein the disease-relevant movement patterns comprise walking patterns identified based on the pitch-warping autocorrelation.

49. The method of claim 35, wherein the second bouts comprise bouts exceeding the mean absolute deviation threshold.

50. The method of claim 35, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 3 Hz to about 12 Hz.

51. The method of claim 50, wherein the disease-relevant movement patterns comprise tremor patterns identified based on the GINI index exceeding a threshold.

52. The method of claim 35, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 4 Hz to about 7 Hz.

53. The method of claim 35, wherein the second bouts comprise bouts below the mean absolute deviation threshold.

54. A system for assessing a progression of a disease in a subject, the system comprising: a processor; and a memory containing instructions thereon, the instructions configuring the processor to:A) receive acceleration data associated with the subject, the acceleration data comprising acceleration values for a plurality of axes and a timestamp associated with each respective acceleration value of the acceleration values;B) analyze the acceleration data to identify a plurality of first bouts of sustained movement;C) calculate an activity threshold to characterize each bout of the plurality of first bouts;D) calculate a power spectrum for each bout of the first bouts;E) calculate, based on the power spectrum and for each bout of the first bouts, at least one spectral analysis parameter for that bout;F) compare the at least one spectral analysis parameter to respective thresholds to identify a plurality of second bouts from the plurality of first bouts;G) calculate, for the plurality of second bouts and based on F), one or more bout distribution parameters for respective distributions of mobility parameters for the plurality of second bouts;H) compare the one or more bout distribution parameters to one or more disease scores; andI) administer at least one of a treatment or a therapeutic based on H).

55. The system of claim 54, wherein the one or more disease scores comprise at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a posturaltremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

56. The system of claim 54, wherein the one or more bout distribution parameters comprise at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

57. The system of claim 54, wherein the disease comprises Parkinson’s disease.

58. The system of claim 54, wherein the activity threshold in C) is determined based on a mode of the acceleration data.

59. The system of claim 58, wherein the mode of the acceleration data comprises a mode of a substantially symmetrical portion of the acceleration data.

60. The system of claim 54, wherein the first bouts comprise a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

61. The system of claim 60, wherein the duration is between about 2 seconds and about 12 seconds.

62. The system of claim 54, the instructions further configuring the processor to apply a correction to the acceleration data such that acceleration values for a first axis of the plurality of axes are substantially aligned with a vector of gravity prior to calculating the power spectrum.

63. The system of claim 54, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation.

64. The system of claim 54, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 3 Hz to about 12 Hz.

65. A system for evaluating a progression of a disease in a subject based on time-stamped acceleration data associated with the subject, the system comprising: a processor; and a memory containing instructions thereon, the instructions configuring the processor to:A) receive, from a sensor, the time-stamped acceleration data associated with the subject during a first time period;B) analyze the time-stamped acceleration data to identify a plurality of first bouts of sustained movement above a movement threshold by the subject;C) calculate, for each bout of the first bouts, at least one mobility parameter associated with a corresponding bout, the at least one mobility parameter based on at least one spectral analysis parameter associated with a power spectrum of each bout;D) identify a plurality of second bouts associated with disease-relevant movement patterns based on the at least one spectral analysis parameter and calculating one or more bout distribution parameters based on the at least one mobility parameter for the plurality of second bouts;E) correlate the one or more bout distribution parameters with at least one disease score to determine the progression of the disease; andF) identify, based on E), at least one of a treatment or therapeutic to be administered to the subject.

66. The system of claim 65, wherein the disease comprises Parkinson’s disease.

67. The system of claim 65, wherein C) further comprises calculate a frequency corresponding to a local maximum amplitude of a Fourier transformation of the time-stamped acceleration data corresponding to each bout of the first bouts.

68. The system of claim 65, wherein C) further comprises calculate, for each bout of the first bouts, a movement frequency associated with a maximum power of the power spectrum.

69. The system of claim 65, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation.

70. The system of claim 65, wherein the movement threshold in B) is determined based on a mode of the time-stamped acceleration data.

71. The system of claim 70, wherein the mode of the time-stamped acceleration data comprises a mode of a substantially symmetrical portion of the time-stamped acceleration data.

72. The system of claim 65, wherein the first bouts comprise a contiguous portion of the acceleration data, the contiguous portion having a duration greater than or equal to a threshold.

73. The system of claim 72, wherein the duration is between about 2 seconds and about 12 seconds.

74. The system of claim 65, wherein the sensor captures the time-stamped acceleration data at a capture frequency of at least 6 Hz.

75. The system of claim 74, wherein the capture frequency is at least 25 Hz.

76. The system of claim 65, wherein B) comprises calculating one or more deviations of the time-stamped acceleration data and a corresponding local maximum associated with each of the one or more deviations.

77. The system of claim 76, wherein the one or more deviations comprises a mean absolute deviation.

78. The system of claim 65, wherein B) further comprises:Bl) determine at least one correction to the time-stamped acceleration data to orient the time-stamped acceleration data with an axis of a second coordinate system based on a gravity vector;B2) select one axis of the second coordinate system based on at least one portion of the time-stamped acceleration data;B3) convert the time-stamped acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data; andB4) identify the at least one bout of sustained movement based on the converted movement data.

79. The system of claim 65, wherein the one or more bout distribution parameters comprises at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

80. The system of claim 65, wherein the at least one disease score comprises at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

81. The system of claim 65, wherein the at least one spectral analysis parameter comprises an autocorrelation associated with the power spectrum of each bout.

82. The system of claim 81, wherein the autocorrelation comprises a pitch-warping autocorrelation.

83. The system of claim 65, wherein the at least one spectral analysis parameter comprises calculating a GINI index from the power spectrum within a frequency range of about 3 Hz to about 12 Hz.

84. A system for characterizing a movement in a subject having a disease, the system comprising: a processor; and a memory containing instructions thereon, the instructions configuring the processor to:A) receive acceleration data, the acceleration data comprising acceleration values for a plurality of axes and a timestamp associated with a respective acceleration value of the acceleration values;B) identify a plurality of first bouts of movement indicated by the acceleration data, the first bouts comprising non-overlapping portions of the acceleration data centered around respective local maxima;C) identify a plurality of second bouts from the plurality of first bouts based on comparison to a mean absolute deviation threshold;D) calculate a power spectrum for each bout of the second bouts;E) calculate, for each bout of the second bouts, at least one spectral analysis parameter based on the power spectrum for each bout of the second bouts;F) identify, from the second bouts and based on the at least one spectral analysis parameter, a plurality of third bouts indicating disease-relevant movement patterns;G) calculate one or more bout distribution parameters for the third bouts, the one or more bout distribution parameters representing a distribution of one or more mobility parameters associated with each bout of the third bouts, wherein the one or more mobility parameters are associated with the movement; andH) classify, based on the one or more bout distribution parameters, the movement of the subject as normal or abnormal.

85. The system of claim 84, the instructions further configuring the processor to: I) correlate the one or more bout distribution parameters with at least one disease score to characterize the movement.

86. The system of claim 85, wherein the at least one disease score is associated with Parkinson’s disease.

87. The system of claim 85, wherein the at least one disease score comprises at least one of a UPDRS part I score, a UPDRS part II score, a UPDRS part III score, a UPDRS part IV score, a Hoehn and Yahr score, a clinic visit score, a sum of a plurality of scores, a postural tremor score, a kinetic tremor score, a resting tremor score, a tremor constancy score, an upper limb tremor score, a rigidity score, a bradykinesia score, a gait score, a freezing score, or a lower limb score.

88. The system of claim 84, wherein the one or more bout distribution parameters comprises at least one of a mean absolute deviation, a mean, a median, a maximum, a quantile, or a root mean square.

89. The system of claim 84, wherein B) further comprises:Bl) determine at least one correction to the acceleration data to orient the acceleration data with an axis of a second coordinate system based on a gravity vector;B2) select one axis of the second coordinate system based on at least one portion of the acceleration data; andB3) convert the acceleration data from a first coordinate system to the second coordinate system based on the at least one correction to provide converted movement data.

90. The system of claim 84, wherein the acceleration data is received from a sensor.

91. The system of claim 90, wherein the sensor captures the acceleration data at a capture frequency of at least 6 Hz.

92. The system of claim 91, wherein the capture frequency is at least 25 Hz.

93. The system of claim 84, wherein the first bouts comprise a contiguous portion of the acceleration data having a duration greater than or equal to a threshold.

94. The system of claim 93, wherein the duration is between about 2 seconds and about 12 seconds.

95. The system of claim 84, wherein the movement comprises a gait of the subject.

96. The system of claim 84, wherein the at least one spectral analysis parameter comprises a pitch-warping autocorrelation.

97. The system of claim 84, wherein the second bouts comprise bouts exceeding the mean absolute deviation threshold.

98. The system of claim 84, wherein the at least one spectral analysis parameter comprises a GINI index within a frequency range of about 3 Hz to about 12 Hz.

99. The system of claim 84, wherein the second bouts comprise bouts below the mean absolute deviation threshold.

Images