Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
DEVICE AND METHOD FOR EVALUATING GAIT ASYMMETRY
Document Type and Number:
WIPO Patent Application WO/2019/180029
Kind Code:
A1
Abstract:
A device for evaluation gait asymmetry of the gait of a walking user, comprises a sensor package comprising an inertial measurement unit or an attitude and heading reference system, to be attached at the user, and a control unit connected with the sensor package to receive data from the sensor package, wherein the control unit is configured to determine the highest (71) and lowest (72) vertical displacement (36) of the center of mass of the walking user, to calculate, based on these highest (71) and lowest (72) vertical displacement values (36) over time (39), the downward acceleration for each right step (31) and for each left step (32) and to calculate an asymmetry value as a function of the difference of the determined step time (41) or downward accelerations (42) for right steps (31) and for left steps (32).

Inventors:
PEINDL, Richard (Carolinas Medical Center1000 Blythe Boulevar, Charlotte North Carolina, 28203, US)
HABET, Nahir (2341 Apple Glen Lane, Charlotte, North Carolina, 28269, US)
Application Number:
EP2019/056858
Publication Date:
September 26, 2019
Filing Date:
March 19, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
AO TECHNOLOGY AG (Grabenstrasse 15, 7000 Chur, 7000, CH)
International Classes:
A61B5/11
Domestic Patent References:
WO2017085914A12017-05-26
WO2014153201A12014-09-25
WO2016097746A12016-06-23
Foreign References:
US20170296116A12017-10-19
US20180325467A12018-11-15
CN107616798A2018-01-23
CN106908021A2017-06-30
Other References:
WIEBREN ZIJLSTRA: "Assessment of spatio-temporal parameters during unconstrained walking", EUROPEAN JOURNAL OF APPLIED PHYSIOLOGY, vol. 92, no. 1-2, 1 June 2004 (2004-06-01), pages 39 - 44, XP055165859, ISSN: 1439-6319, DOI: 10.1007/s00421-004-1041-5
F. BUGAN? ET AL: "Estimation of spatial-temporal gait parameters in level walking based on a single accelerometer: Validation on normal subjects by standard gait analysis", COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE., vol. 108, no. 1, 1 October 2012 (2012-10-01), NL, pages 129 - 137, XP055301546, ISSN: 0169-2607, DOI: 10.1016/j.cmpb.2012.02.003
Attorney, Agent or Firm:
LIEBETANZ, Michael (Postfach 1772, 8027 Zürich, 8027, CH)
Download PDF:
Claims:
CLAIMS

1. A device for evaluation gait asymmetry of the gait of a walking user (10), comprising:

a sensor package (20) comprising an inertial measurement unit or an attitude and heading reference system, to be attached at the user (10), and

a control unit (22) connected with the sensor package (20) to receive data from the sensor package (20),

wherein the control unit (22) is configured to determine the highest and lowest vertical displacement value of the center of mass of the walking user (10), to calculate, based on these highest (71) and lowest (72) vertical displacement values (36) over time (39), the step time (41) or downward acceleration (62) for each right step (31) and for each left step (32) and to calculate an asymmetry value as a function of the difference of the determined step time (41) or downward accelerations (42) for right steps (31) and for left steps (32).

2. The device according to claim 1, wherein the control unit (22) is adapted to calculate the asymmetry value as an acceleration asymmetry value, the acceleration asymmetry value being determined as a percentage

[ |(Right acceleration data - Left acceleration data)| / max {Right acceleration data, Left acceleration data}] x 100

wherein right acceleration data is an average value of the downward accelerations for all right steps of the walking user (10), wherein left acceleration data is an average value of the downward accelerations for all left steps of the walking user (10), and wherein the max-function relates to the highest downward acceleration value of the entire measurement series.

3. The device according to claim 1 or 2, wherein the sensor package (20) comprises a torso harness (21) adapted to fix the sensor package (20) on the central line of the user (10).

4. The device according to any one of claims 1 to 3, wherein the control unit (22) is provided as a separate device and wherein the connection between the sensor package and the control unit (22) to transmit data is a wireless connection (23).

5. The device according to any one of claims 1 to 4, wherein the control unit (22) is adapted to calculate the asymmetry value as a step time asymmetry value, wherein the step time asymmetry value is calculated as [ |(Right step time data - Left step time data)| / max {Right step time data, Left step time data}] x 100

wherein right step time data is an average value of the step time for all right steps of the walking user (10), wherein left step time data is an average value of the step time for all left steps of the walking user (10), and wherein the max-function relates to the highest step time value of the entire measurement series.

6. The device according to claim 2 and according to claim 5, wherein the control unit (22) is adapted to calculate the asymmetry value as a combined asymmetry value, which is determined based on a function of the acceleration asymmetry value and the step time asymmetry value.

7. The device according to claim 6, wherein the function of the acceleration asymmetry value and the step time asymmetry value is taken from the group comprising mean averaging and harmonic mean.

8. A method for evaluation gait asymmetry of the gait of a walking user (10), comprising the steps of:

providing a sensor package (20) comprising an inertial measurement unit or an attitude and heading reference system,

attaching the sensor package (20) at the user (10),

providing a control unit (22) connected (23) with the sensor package (20) to receive data from the sensor package (20) ,

determining, within the control unit (22), the highest (71) and lowest (72) vertical displacement value (36) of the center of mass of the walking user (10),

calculating, within the control unit (22), based on these highest (71) and lowest (72) vertical displacement values (36) over time (39), the step time (41) or downward acceleration (62) for each right step (31) and for each left step (32) made by the user (10), and

calculating, within the control unit (22), an asymmetry value as a function of the difference of the determined step time (41) or downward accelerations (42) for right steps (31) and for left steps (32).

9. The method according to claim 8, wherein the control unit (22) is configured to calculate the asymmetry value as an acceleration asymmetry value determined as percentage

[ [(Right acceleration data - Left acceleration data)| / max{Right acceleration data, Left acceleration data}] x 100

wherein right acceleration data is an average value of the downward accelerations for all right steps of the walking user (10), wherein left acceleration data is an average value of the downward accelerations for all left steps of the walking user (10), and wherein the max-function relates to the highest downward acceleration value of the entire measurement series.

10. The method according to claim 8 or 9, wherein the attachment step of the sensor package (20) comprises providing a torso harness (21) adapted to fix the sensor package (20) on the central line of the user (10).

11. The method according to any one of claims 8 to 10, wherein the control unit (22) is provided as a separate device from the sensor package (20) and wherein the connection between the sensor package (20) and the control unit (22) to transmit data is a wireless connection (23).

12. The method according to any one of claims 8 to 1 1, wherein the control unit (22) is configured to calculate the asymmetry value as a step time asymmetry value as [ [(Right step time data - Left step time data)| / max {Right step time data, Left step time data}] x 100

wherein right step time data is an average value of the step time for all right steps of the walking user (10), wherein left step time data is an average value of the step time for all left steps of the walking user (10), and wherein the max-function relates to the highest step time value of the entire measurement series.

13. The method according to claim 9 and 12, wherein the the control unit (22) is configured to calculate the asymmetry value as a combined asymmetry value, which is determined based on a function of the acceleration asymmetry value and the step time asymmetry value.

14. The method according to claim 13, wherein the function of the acceleration asymmetry value and the step time asymmetry value is taken from the group comprising mean averaging and harmonic mean.

Description:
TITLE

DEVICE AND METHOD FOR EVALUATING GAIT ASYMMETRY

TECHNICAL FIELD

The present invention relates to a device and method for evaluation gait asymmetry.

PRIOR ART

Gait assessment in performance based measures is important when a patient shows gait asymmetry. Such devices and methods for evaluating gait asymmetry, to be useful in both a clinical and research setting, should: a) be accurate and reproducible, b) objectively capture clinically meaningful functional information, and c) involve a quick and convenient measurement and report on results.

Commonly used performance based measures in the clinical setting include the timed-up- and-go test (TUG) and the 5 times sit-to-stand (5xSTS) test. However, these tests only measure the total time to accomplish the task, and cannot capture more subtle alterations in function.

CN 107 616 798 A provides a method for detecting gait asymmetry based on gravitational acceleration. Triaxial acceleration sensors acquire data of a walking user. Based on the changes in the displacement of the center of gravity during movement gait characteristic values are derived.

CN 106 908 021 A provides a wearable sensor and discloses a human body step length measuring method using of two inertial sensor units to collect and store the acceleration and angular velocity data of a human body in the walking process, and then uses an algorithm to calculate the user's step length information. The method can be used to estimate the user's gait asymmetry.

WO 2016/097746 Al discloses an arrangement for analyzing the biomechanical motion of a wearer comprising an acceleration sensing unit attached to the upper body of the wearer, for example to the ear, for sensing acceleration of the wearer's upper body and generating a series of acceleration values indicative of that motion; and a processor communicatively coupled to the acceleration sensing unit for processing the acceleration values, the processor being configured to gather real-time acceleration values from the acceleration sensing unit and to compare the real-time acceleration data with one or both of: (a) historic acceleration data derived from the acceleration sensing unit and (b) predetermined acceleration features indicative of biomechanical motion quality to form an output indicative of the current state of the biomechanical motion of the wearer.

SUMMARY OF THE INVENTION

Based on the prior art it is - inter alia - an object of the present invention to provide an improved method and device to not only detect gait asymmetry but to provide a reliable and valid result data to qualify gait asymmetry within a short test walk.

It is an aim of the present invention to provide a wearable with task-appropriate analytics to provide accurate and clinically meaningful performance-based gait data on orthopaedic trauma patients as part of normal clinic operations. This object can be reached with an IMU to generate kinematic variables which correlate with observed clinical pathology and also provide quantitative data to augment patient-reported outcome measures.

Within the framework of the present invention, algorithms had been developed to assess vertical acceleration, stance times and vertical displacement resulting from right and left strides along with peak-to-peak mean forward/backward pitch and side-to-side roll during the test. Vertical accelerations and vertical displacements are related to Left/Right push-off forces and torso up/down excursions, respectively. Stance times provide times spent on each limb during repetitive gait cycles. All of these measures can be visualized in the clinic but heretofore could not be quantitatively evaluated.

For two-to-three decades, microelectromechanical system (MEMS) inertial measurement units (IMUs) have been evaluated and/or commercially used to provide various measurements related to human body kinematics. These IMUs generally utilize tri-axial accelerometers and/or gyroscopes. Data from such systems attached to moving body segments is with respect to a sensor coordinate system with a continuously changing orientation. In contrast, affordable MEMS attitude and heading reference systems (AHRSs) were introduced around the year 2000 primarily for navigational use in various land, air and sea vehicles. AHRSs are IMUs that typically integrate an additional (non-inertial) tri- axial magnetometer and an on-board data processing system that computes continuous local sensor coordinate system attitude and heading information, in the form of orientation quaternions, in addition to 3-D linear acceleration and 3-D angular velocity data.

Unfortunately, AHRSs are often referred to simply as IMUs. The primary advantages of AHRSs over accelerometer- and gyroscope-based IMUs are: 1) all acceleration and angular velocity data can be collected with regard to a global coordinate system (e.g. North, East, Down (NED), 2) using a local coordinate system initialization procedure, data from a continuously moving sensor coordinate system can be converted to a non-moving, pre-established body coordinate system, and 3) when using multiple AHRSs, all sensor data streams can be transformed into the same global or fixed-local coordinate system. In this regard, commercially available AHRS systems for full body motion analysis can function similarly to reflective marker video motion tracking systems for medical research purposes. As configured for full-body motion capture, however, multiple AHRSs do not lend themselves to use in a busy clinical setting for routine patient mobility assessments.

In other words, the key difference between an inertial measurement unit (IMU) and an attitude and heading reference system (AHRS) is the addition of an on-board processing system in an AHRS which provides attitude and heading information versus an IMU which delivers sensor data to an additional device or control unit that computes attitude and heading. In addition to attitude determination an AHRS may also form part of an inertial navigation system. In efforts to obtain clinically relevant data, suitable for busy medical practices, it may be efficient to use only one IMU or AHRS to provide limited patient gait assessments. The present invention is related to determine a method for employing minimal instrumentation for kinematic and kinetic testing of patients in a busy orthopaedic clinic setting.

The gait measures that best indicate the degree of gait abnormality, from a clinical observation standpoint, are those that deal with spatio-temporal and/or kinetic asymmetries of motion. Therefore, the present invention starts to use a single chest-mounted AHRS to quantify vGRF asymmetry and step time asymmetry as measures of perceived kinetic and spatio-temporal gait abnormalities in a cadre of post-surgical lower extremity trauma patients versus matched healthy controls. A validated value of such measurements is the “lift acceleration” as a new measure for assessing gait kinetic asymmetry.

Spatio-temporal and kinematic asymmetry data from a single AHRS demonstrated the potential for using these methods and measures for limited gait assessment of lower extremity trauma patients in the clinic.

The measurements of the device according to the invention are showing that: 1) recovering patients with a lower limb injury spend less single stance time and have a diminished lift acceleration on their injured versus a lesser-injured or uninjured limb, 2) that AHRS vertical acceleration P2P times and P2T lift accelerations correlate well with spatio- temporal and kinetic measures as determined using a Vicon motion capture system wherein chest-mounted AHRS data filtering is important, and 3) step time and lift acceleration asymmetry measures show the relevance of these values in indicating degree of gait abnormality.

A device for evaluation gait asymmetry of the gait of a walking user can comprise a sensor package having an inertial measurement unit or an attitude and heading reference system, to be attached at the user, and a control unit connected with the sensor package to receive data from the sensor package, wherein the control unit is configured to determine the highest and lowest vertical displacement value of the center of mass of the walking user, to calculate, based on these highest and lowest vertical displacement values over time, the step time or downward acceleration for each right step and for each left step and to calculate an asymmetry value as a function of the difference of the determined step time or downward accelerations for right steps and for left steps.

As a technical indicator, the device can calculate an acceleration asymmetry value as a percentage from [ ((Right acceleration data - Left acceleration data)| / max{Right acceleration data, Left acceleration data}] x 100, wherein right acceleration data is an average value of the downward accelerations for all right steps of the walking user, wherein left acceleration data is an average value of the downward accelerations for all left steps of the walking user, and wherein the max-function relates to the highest downward acceleration value of the entire measurement series. This value is an objective value of gait asymmetry compared to the subjective assessment of a traumatologist who monitors a walking user over several gait cycles. Additionally, the value can be determined after only one or two passes of 10 meters without the necessity of the presence of a physician, thus reducing the costs of this automatic assessment of the gait asymmetry.

The average value of the downward accelerations for all steps of one kind can be taken from the group encompassing the traditional mean value or the harmonic mean value or the median value.

The sensor package of the device can comprise a torso harness adapted to fix the sensor package on the central line of the user. The fixation of the harness on the sternum as a near-bone fixation point has the advantage that no other elements interfere in the monitoring and gathering of the gait movement values. The central line is on the median plane of the body.

The control unit can be provided as a separate device, e.g. a computer being connected to a storage to directly store and use the determined measurement values. Then, the connection between the sensor package and the control unit to transmit data is a wireless connection; especially it can be a Bluetooth or wireless lan connection.

The device allows to determine a further asymmetry value, the step time asymmetry value which is calculated as [ ((Right step time data - Left step time data)| / max{Right step time data, Left step time data}] x 100, wherein right step time data is an average value of the step time for all right steps of the walking user, wherein left step time data is an average value of the step time for all left steps of the walking user, and wherein the max-function relates to the highest step time value of the entire measurement series. The step time asymmetry can be combined with the gait acceleration asymmetry value, which are preferably both provided as percentage from the value 0 for no presence of an asymmetry.

The average value of the step time for all steps of one kind can be taken from the group encompassing the traditional mean value or the harmonic mean value or the median value of this time value.

In an embodiment the device determines a combined asymmetry value, which is calculated as {[ ((Right acceleration data - Left acceleration data)| / max{Right acceleration data, Left acceleration data}] / 2} + [ ((Right step time data - Left step time data)| / max (Right step time data, Left step time data}] / 2,

wherein right acceleration data is an average value of the downward accelerations for all right steps of the walking user, wherein left acceleration data is an average value of the downward accelerations for all left steps of the walking user, and wherein the max- function relates to the highest downward acceleration value of the entire measurement series,

wherein right step time data is an average value of the step time for all right steps of the walking user, wherein left step time data is an average value of the step time for all left steps of the walking user, and wherein the max-function relates to the highest step time value of the entire measurement series.

In other words, it is possible to base the asymmetry value on the calculation of two different asymmetry values, i.e. the the acceleration asymmetry value as well as the step time asymmetry value giving as a result a combined asymmetry value, which is determined based on a function of the acceleration asymmetry value and of the step time asymmetry value. This function can especially be achieved by the mean average or the harmonic mean. It can be provided in percentage of asymmetry (x 100 as mentioned above) or as a value between 0 (no asymmetry) and 1 (full asymmetry, theoretical case).

A method device for evaluation gait asymmetry of the gait of a walking user comprises the steps of providing a sensor package comprising an inertial measurement unit or an attitude and heading reference system, attaching the sensor package at the user, providing a control unit connected with the sensor package to receive data from the sensor package, determining, within the control unit, the highest and lowest vertical displacement value of the center of mass of the walking user, calculating, within the control unit, based on these highest and lowest vertical displacement values over time, the step time or downward acceleration for each right step and for each left step made by the user, and calculating, within the control unit, an asymmetry value as a function of the difference of the determined step time or downward accelerations for right steps and for left steps.

Further embodiments of the invention are laid down in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the invention are described in the following with reference to the drawings, which are for the purpose of illustrating the present preferred embodiments of the invention and not for the purpose of limiting the same. In the drawings,

Fig. 1 shows an image of a harness with straps and an IMU noting location of the sensor package at the top of the sternum of a user;

Fig. 2 shows a schematic representation of the temporal (horizontal) and kinetic

(vertical) relationships vertical of normal human gait; and

Fig. 3 depicts within the three row of graphs show: simulated acceleration profiles for increasingly abnormal gait, result of direct double integration of sine series functions of the first row and evaluation of the resultant displacement function, and actual AHRS gait cycle acceleration patterns for several users.

DESCRIPTION OF PREFERRED EMBODIMENTS

Examples of preliminary measurements: Pain interference and physical function PROMIS measures on lower extremity trauma patients were collected along with a modified 10 m walk tests (mlOmWT) on patients instrumented with a single chest-mounted 10 degree-of- freedom IMU. An inertial measurement unit (IMU) is an electronic device that measures and reports a body's specific force, angular rate, using a combination of accelerometers and gyroscopes. The following Tables 1 provides PROMIS measure scores at the time of testing and Tables 2&3 provide kinematic gait test variables for three patients who were selected to illustrate various degrees of gait normalcy / abnormality. Subject 1 demonstrates a normal gait after surgery and recovery. Subject 2 demonstrates a noticeable limp and subject three demonstrates a continuing pathological gait pattern at 82 month post-surgery. However, subjects 2 and 3 routinely utilize an energy storing orthosis on their affected limb and were tested with and without the brace. Tables 2&3 provide data from gait tests without the brace. 50 is the U.S. general modulation mean value. +-10 is 1 standard deviation from said mean.

Table 1 : PROMIS Measures

Subject 3 demonstrated significantly more sidedness for vertical accelerations and stance times than subjects 1 or 2 or controls (Table 2) and significantly greater vertical displacement, pitch and roll than all others tested (Table 3). His PROMIS function score, however, is within 1 standard deviation of the mean and pain interference score is just outside of 1 standard deviation (Table 1).

o

O

00 o o

Table 2 - Quantitative Gait Parameters in Injured vs Control Subjects

P2P Vert Ace = Peak to Peak (+/-) Acceleration; indicative of vertical push off forces (L/R differences quantify

asymmetry in L/R push-off forces regardless of patient size or gait velocity)

DS2DS Time = Peak Acceleration to Peak Acceleration (+/+) Dual Stance Intervals; (L/R differences quantify

aperiodicity regarding mean time spent on each limb regardless of patient size or gait velocity).

Significantly different values from controls depicted in gray n H e¾

O

O

Os

00

00

Table 3:

In order to verify the method according to the invention five recovering, surgically-treated, lower-extremity trauma patients and five age and gender matched controls were tested in a laboratory equipped with a video motion capture (Vicon) system. Trauma patients were pre-selected with varying degrees of gait abnormality to assess the ability to quantify perceived gait abnormality based on step time and lift acceleration asymmetries. AHRS derived asymmetry data for step time and lift acceleration were then compared with Vicon- derived values for these same measures.

The result of raw AHRS vertical acceleration data streams strongly correlated with Vicon tracking data for markers attached to the AHRS (r > 0.9 in all cases). Pearson analysis results for aggregate patient and control AHRS versus Vicon step time data were: r = 0.699, with a 95% confidence interval of 0.372 to 0.872, considered very significant (p =

0.0006). Aggregate patient and control Pearson analysis results for AHRS versus Vicon lift acceleration data were: r = 0.993, with a 95% confidence interval of 0.983 to 0.997, considered extremely significant (p = 0.000001). Asymmetry rank values agreed with perceived gait abnormality as assessed by two orthopaedic traumatologists.

Fig. 1 shows an image of a harness with straps 21 and an IMU noting location of the sensor package 20 at the top of the sternum of a user 10 standing on ground 11. The sensor package 20 can be an AHRS or an IMU. It can also comprise a data control unit for an on- board handling of data. However, the sensor package 20 is preferably connected in a wireless manner 23 with an external control unit 22 like a desktop computer in the room where the tests are conducted. The user 10 is requested to walk 10 meters on a flat ground surface 11 , e.g. 5 meters in one direction and 5 meters back to the starting point, while the sensor package 20 takes a number of measurements and transmits them via the wireless connection (as e.g. Bluetooth or wireless lan) 23 to the control unit 22. The attachment of the sensor package 20 with straps 21 on the sternum of the user 10 has the advantage, that it is a central location on the body with no lateral offset when the user 10 is walking forward. Of course, the sensor package 20 comprises a power supply, e.g. a battery pack.

Fig. 2 is a schematic representation of the temporal (horizontal) and kinetic (vertical) relationships vertical of normal human gait. Vertical ground reaction force (VGRF) M- curves are shown at bottom for right step 31 and left step 32 along with the total ground reaction force 33 (GRF = dashed line) over time. The user 10 is making one right step 31 and one left step 32 in about 0.8 seconds. The vertical ground reaction force of a walking human is related with vertical acceleration 34, velocity 35 and displacement 36. These three values are shown in this order above the ground reaction force 33.

As mentioned above, the sensor package 20 comprises a control unit or the external control unit 22 comprising a programmable microprocessor having a computer program configured to calculate, based on the measurement of vertical displacement 36 and vertical velocity step time (P2P) 41 and lift acceleration (P2T) 42. The so-called push-off point 43 is denoted as the center of dual stance (DS) and mid-swing (MSw) 44 denotes the point of maximum single leg weight bearing. These are specific points in time and maximum value or zero crossings as seen in Fig. 2. As mentioned above, Fig.2 is the drawing of the curves measured for a user 10 in health. P2P time intervals 41 are shortened and P2T displacement magnitudes 42 are reduced for injured limbs in patients tested compared to contralateral limb values.

In Fig.2, troughs 52 in the vertical acceleration curve correspond to initiation of the mid- swing (MSw) phase 44 of each step 31 and 32 and also to maximal excursions in vertical displacement 36 of the center of mass (CoM). Likewise, peaks 53 on the acceleration curve correspond to mid-dual stance (DS) 43 and to minimal excursions 54 in CoM vertical displacement. Note that DS is actually a temporal range that lasts from heel-strike of one limb to toe-off of the contralateral limb. Note also that it is the downward-directed peak-to- trough (P2T) portion 62 of the acceleration curve that corresponds to the upward-directed trough-to-peak portion 66 of the CoM displacement curve. This P2T acceleration segment correlates with the terminal dual stance, pre-swing and initial swing phases of a step cycle. Since this P2T acceleration segment 62 is also associated with the upward forces applied to the CoM from push-off for one limb post dual stance to weight bearing by the contralateral limb at midstance in each step 31, 32, this segment is hereafter referred to as“lift acceleration” 62. Thus, the end points of lift acceleration 62 involve significant activities by both limbs of a user 10 for a given step 31 or 32. In abnormal gait, due to lower extremity trauma and surgical repair, both end points are affected and the kinematic and kinetic relationships in Fig. 2 are altered as will be addressed.

Five recovering, surgically-treated, lower-extremity trauma patients and five age and gender matched controls were tested in a Vicon-equipped laboratory. Patients were pre- selected to demonstrate full recovery or varying degrees of gait abnormality. Demographic data for all subjects is provided in Table 4 along with diagnostic information and time from surgery for each patient.

Table 4

Patients 1 and 4 were clinically assessed by orthopaedic traumatologists as having normal gait, while patients 2, 3 and 5 were assessed as having progressively mild to excessive gait abnormality. Patients 3 and 5 typically wear a dynamic brace for activities of daily living but were tested without their braces.

A modified lOmWalkTest (mlOmWT) (i.e. walk 5 meters, turn and walk five meters) was used due to walkway length constraints. A ten-camera VICON motion capture system (Vicon Motion Systems Ltd, Oxford, UK) was used to record motion data at 100 Hz. The measurement space was about 5.0m L x 2.0m W x2.5m H, and the 3-D residue of marker position tracking was lower than 1 mm after system calibration. All tests were also video- recorded at 60 fps. Accelerating and decelerating steps, at either end of the walkway, were removed from the step time and lift acceleration asymmetry analyses. These were steps on either end of the walkway where peak accelerations were less than 1 std of the remaining step acceleration peaks. All subjects were instrumented with reflective markers using a modified Helen Hayes marker set (further information relating thereto can be found in https://www.ncbi.nlm.nih.gov/pubmed/l9473844) and a single AHRS (Opal, APDM, with magnetometer engaged) 20 was strapped to each subject’s sternum using the manufacturer’s elastic chest harness with straps 21.

A marker triad was placed on the AHRS for Vicon corroboration of AHRS acceleration, angular velocity and orientation data. The mean of the AHRS marker triad data was also used to track the instantaneous position of the AHRS and to provide data as a Vicon clavicle CLAV marker. For Vicon motion analysis, the midpoint of the CLAV and C7 markers was used to approximate the CoM of the subject’s head, arms and thorax (HAT).

The control unit 22 comprises a computer program adapted to 1) determine P2T magnitudes (lift accelerations) and P2P intervals (step times) using AHRS data from the sensor package 20 transmitted via connection 23 and 2) to separate these values for right and left steps. P2T right lift acceleration 162 (see Fig. 3) and left lift acceleration 262 were calculated using sequential maxima and minima from AHRS vertical acceleration data. Vicon vertical acceleration data was determined using double-differentiated position data from a point midway between the Vicon CLAV and C7 markers as an approximation of the HAT CoM vertical acceleration. Step time and lift acceleration data are presented as the mean + sem for all right steps and left steps from four, 5-meter gait segments per subject.

A single AHRS as sensor package 20 was located on the upper sternum (i.e. manubrium) of each individual user 10 tested. This location was selected as a single-device, multi-test compromise that could be used for: 1) a 10-meter- walk-test (lOmWT), 2) Timed-Up-and- Go (TUG) tests and 3) the 5-times-Sit-To-Stand (5xSTS) tests in the clinic without repositioning the AHRS between tests. In the latter two tests the head, arms and upper torso (HAT) rotational movements are more pronounced than in gait and measurement of these angular displacement parameters is better served by an upper-torso-mounted AHRS.

In addition, the sternum, as opposed to belt-level device locations, positions the AHRS near a bony surface regardless of subject BMI and away from any metallic gait assistive devices that may be required for gait or metal-frame chairs (i.e. used for TUG and 5xSTS tests). Such accessories can affect AHRS magnetometer readings.

Initially, all Vicon data were converted to 128Hz to match the AHRS sampling frequency. Vicon CLAV markers and AHRS raw vertical acceleration waveform data were then compared to synchronize the waveforms obtained from the two separately triggered systems. Preferably, a signal processing function was used which shifts two continuous waveforms in time until an optimal fit (best synchronization) between waveforms is achieved.

Next, AHRS P2P vertical acceleration intervals and P2T lift accelerations were compared versus Vicon DS2DS step times and HAT lift accelerations on an individual step basis. The mean estimation error (i.e. AHRS vs Vicon) for N steps + sem for each lower extremity was obtained using the equation:

M

AHRS data(i)- Vicon data(i)

xlOO = est. error (%) Eq.1

Vicon data (i)

Mean P2P and P2T data were then used to determine a percent asymmetry for the two gait variables of each subject. Step time and lift acceleration percent asymmetry data (right vs left) was calculated using the mean step time and mean lift acceleration data as:

[ ((Right data - Left data)| / max{Right data, Left data}] x 100 = asymmetry (%) Eq.2 Asymmetry measures data for patients and controls were combined to perform Pearson correlations of AHRS and Vicon data for both step time and lift acceleration measures.

Vicon and AHRS vertical acceleration data for all gait segments and for all subjects tested were“highly correlated' ' with r > 0.9 in all cases. This indicates that the AHRS provided accurate acceleration data for the Vicon CLAV marker position for the mlOmWT for all patients and controls.

Step time and lift acceleration accuracy assessment data can be found in tabular format in Table 5 and 6. Aggregate mean right and left step time deviations for all AHRS P2P vs Vicon DS2DS step time data averaged 8.01% for all patients and 2.98% for all controls. Aggregate deviation for the patient group was significantly affected differences in AHRS and Vicon accuracy assessments. Pt 3 Vicon step time data, using forefoot and hindfoot markers was corroborated with video data. Examination of Pt 3 data showed that peak accelerations for the affected limb were affected by high frequency components that were not adequately filtered; this type of error has been previously reported when using chest- mounted AHRSs for gait analysis. All patients, however, demonstrated shorter step times for trauma affected limbs for Vicon data and for AHRS data with the exception of Pt 3.

Table 5

Table 6

Lift acceleration aggregate mean deviations for all P2T vs Vicon HAT CoM data averaged 10.92% for all patients and 6.22% for all controls tested. Affected limbs demonstrated lesser lift accelerations in all patients as compared with unaffected or lesser affected limbs using both AHRS and Vicon data. As is the case with all surface mounted sensors, percent error was affected by the fact that the AHRS and HAT CoM are not at the same location.

Table 7a and 7b provide asymmetry data for step times and lift accelerations for patients and controls using Vicon and AHRS data. Patient lift acceleration asymmetry ranking agrees with initial subjective traumatologists’ rankings but AHRS step time ranking was less discriminatory. Controls asymmetry data were small regardless of method used for determination. Pearson analysis results for aggregate (patient and control) AHRS P2P vs Vicon DS2DS step time data were: r = 0.795, with a 95% confidence interval of 0.330 to 0.949, considered very significant at p = 0.0060. AHRS and Vicon lift acceleration percent asymmetry data agreed with traumatologists’ rankings of relative asymmetry. Again, control values for asymmetry were small in all cases. Pearson analysis results for AHRS P2T vs Vicon lift acceleration data were: r = 0.977, with a 95% confidence interval of 0.903 to 0.995, considered extremely significant at p = 0.0001.

Table 7a

Table 7b

Note that the consistent overestimation of AHRS lift acceleration data versus Yicon data is effectively “normalized” by asymmetry calculations resulting in the exceptional concurrence of AHRS and Vicon data.

In locating a single AHRS (e.g. foot, shank, hip, lower back, sternum and upper back), for optimal accuracy one needs to consider the various tests to be performed and analyzed, prevalence of muscle movements or excessive soft tissue beneath the sensor (i.e. which can introduce sensor noise), and potential magnetic interference of metal contained in gait assistive devices or chairs used in tests.

The vertical acceleration 62, 162, 262 measured by an AHRS located on the sternum highly correlated (r > 0.9 for all tests using the MATLAB xcorr function) with the vertical acceleration of a Vicon CLAV marker at the same location.

It has to be noted that for each patient 10 in the tested group, however, consistent variations in acceleration peak-to-trough (P2T) patterns for affected versus unaffected limbs were measured. This allows evaluating lift acceleration 62 as a measure in assessing kinetic asymmetry in that it defines acceleration and vGRF range from push-off on one limb to single stance weight bearing for the contralateral limb. Both lift acceleration limits (i.e. max and min accelerations) are affected immediately post-injury and throughout recovery due to physiologic constraints and/or compensatory mechanisms and measuring them as described allows for a relevant objective value in assessing gait asymmetry.

Lift acceleration asymmetry also has a direct effect on various types of observable vertical displacement patterns during gait. Fig. 3 illustrates this point to demonstrate the capability of a single AHRS to provide reliable qualitative and quantitative vertical displacement data even when dealing with abnormal gait. Upper body kinetics and kinematics are major observational features of gait analysis separate from lower body spatio-temporal measures. The asymmetry results, particularly for lift acceleration 62, 162, 262 support the previous findings of 10% - 13% asymmetry as being an observational threshold for perceiving gait abnormality by a traumatologist, which represents a subjective point of view. Fig. 3 depicts within the first row of graphs simulated acceleration profiles for increasingly abnormal gait (i.e. left to right) constructed from a multiple sine series. The second row of graphs was created by direct double integration of these sine series functions and subsequent evaluation of the resultant displacement function over the same time periods as the accelerations. Row three presents actual AHRS gait cycle acceleration patterns for patients 1 , 3 and 5 along with double trapezoidal rule integration of the acceleration data.

There are multiple different variables that can be used to quantify gait characteristics in both normal and pathologic gait. The present invention focus on four parameters to explore the IMU’s potential to distinguish normal from pathologic gait: time in stance phase for each limb,“peak-to-peak” upper body vertical acceleration (i.e. which relates to observable vertical displacement), upper body oscillations in sagittal plane angulation (“pitching”), and upper body oscillations in coronal plane angulation (“rolling”). In gait pathology, alterations in all of these parameters are observed as compensatory mechanisms but they are generally not quantified.

During the gait cycle, there is maximal vertical acceleration between heel strike and toe-off when the center of mass is lowest and begins to displace upwards. Likewise, there is a peak negative vertical acceleration during mid-stance phase when the center of mass is highest and begins its downward descent toward the next heel strike. The time in stance phase for each leg can be measured as the difference between vertical acceleration peaks. This period is often shortened on the pathologic side as accommodative strategies are employed to minimize time on the painful/disabled limb. Similarly, the difference between minimum acceleration and the maximum acceleration on each limb can be calculated as a“peak to peak acceleration” (P2P A), which is likewise diminished for the affected limb during single-leg stance.

Similarly, normal gait is associated with cyclic changes in upper-body sagittal plane angulation (also called“pitch”) and upper-body coronal plane angulation (also called “roll”). The total difference between maximal positive and negative angulation in both planes can be calculated to measure the net angular displacement during the gait cycle. Pitch and/or roll typically increase in pathologic gait as the torso increasingly moves to more extreme positions to minimize joint reactive forces or accommodate for stiffness, weakness, or pain in the limb. For each gait segment, the three maximal pitch and three minimal pitch values can be used to calculate the peak-to-peak pitch value for that segment. The pitch results for four gait segments (i.e. two segments x two tests) were combined to produce the mean peak-to-peak pitch values. This same procedure was followed in calculating mean peak-to-peak roll data. The four gait cycle variables listed above (i.e. time in left and right stance phases, peak-to-peak acceleration, maximal pitch variations and maximal roll variations with repetitive gait cycles) were then compared for the previously-injured and matched control individuals.

LIST OF REFERENCE SIGNS user 43 centre of dual stance

ground 44 mid swing initiation

sensor package 52 trough

strap 53 peak

control unit 54 minimal excursion

wireless connection 62 P2T downward portion of right step VGRF acceleration / lift acceleration left step VGRF 66 upward portion of displacement total vertical ground force 71 highest displacement

vertical acceleration 72 lowest displacement

vertical velocity 162 asymmetric downward portion of vertical displacement acceleration, right step time 262 asymmetric downward portion of velocity step time P2P acceleration, left step

lift acceleration P2T