Patent Application
20080053253
Before explaining at least one embodiment of the present invention in detail, it is to be understood that the invention is not limited in its application to the details of construction and to the arrangements of the components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced and carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein are for the purpose of description and should not be regarded as limiting.
In a preferred embodiment, the present invention takes the form of a fully ambulatory, microcontroller-based, stride and gait evaluation monitor (SAGE-M) that is a small, self-contained device (weighing less than 100 grams and approximately the size of a pager) that is mountable on the shank just above the ankle. See
The device utilizes a combined accelerometer/gyroscope sensor array that is mounted on a patient's leg. This device provides improved accuracy (5 cm) over a wide range of stride length (0.2-1.5 m) by using a three stage process: (i) vertical acceleration of the shank detects periods of locomotion; (ii) an initial stride length estimate is calculated by integration of the gyroscope angular velocity signal, and (iii) a final accurate stride length value is determined using a novel calibration algorithm that accounts for forward motion of the body over the stance foot.
Frequency analysis of shank vertical acceleration data can be used to detect episodes of freezing. Meanwhile, custom analysis software is used to processes stride data to provide clinically-relevant information on the PD patient's response to dopaminergic therapy, such as latency from administration to improved stride length, abruptness of transition from ‘off’ to ‘on’ (using an Emax function), and time spent in the ‘off’ state.
A preferred embodiment of the present invention 10 includes an 8 bit, on board programmable flash microcontroller 12, a transducer array 14 that includes a ±6 g accelerometer (a triaxial package in which only one channel is logged and which has a frequency response 0-40 Hz), a ±1200°/sec angular gyroscopic velocity sensor (frequency response 0-40 Hz), appropriate signal conditioning and filtering circuitry 16, a 12 bit A/D converter 18 which has a 100 Hz signal sampling rate, a 35 Mbyte flash memory 20 having 24 hour recording capacity at 400 bytes/second, a USB 2.0 interface 22 (alternatively, the data could be wirelessly transmitted to a remote personal computer), a AAA 9V NiMH rechargeable battery that is chargeable through the USB port, and an external event button 24 to allow a user to flag the occurrence of 1 s an event (medication administration, freezing episode, etc) that is pertinent to the analysis of the wearer's gait characteristics. See
The firmware 26 that was developed for the present invention utilized structured programming implemented in C and assembly language. It is interrupt driven firmware and includes the following functions: (a) timing clock, (b) A/D conversion ready, (c) USB port data reception, (d) command setting with host computer, (e) synchronization of start of recording time with host computer (upon command), (f) during recording, sequential memory storage of the following data blocks at a constant 100 Hz rate: vertical acceleration (12 bits), angular rate (12 bits), event status (8 bits).
The present invention is also equipped with a elasticized strap 30 having a hook and loop fastener that allows the unit to be mounted around a patient's shank (e.g., just above the ankle).
Data can be transferred (e.g., using a USB cable) to a personal computer (PC) and processed using a Windows-based interface and custom analysis software (SAGE-S) written in Labview G (National Instruments, Austin, Tex.). The interface aspect of this software allows the user to program the start date and time for data acquisition, enter patient information into a data file, upload a data file to a PC using the USB connection, and check SAGE-M battery status.
The analysis aspect of the present invention enables one to use a leg's vertical acceleration and angular velocity measurements to compute any one of a host of clinical parameters relating to PD gait dysfunction and a patient's response to dopaminergic therapy.
A calibration algorithm is used to correct for movement of the body over the stance foot to determine the length of the stride. Other clinical parameters of interest include: stride length variability, ‘off’ time (when the stride length is less than 50% (a defined percentage) above a baseline value), latency of locomotor response to levodopa, abruptness of transition from ‘off’ to ‘on’, etc.
The performance of the present invention is illustrated in
An estimate of the angular extent of leg swing was obtained by integration of the negative portion of the angular velocity trace during periods of locomotion (as determined from the RMS vertical acceleration). An initial stride length estimate (SLi) was calculated as follows:
SLi=2×l×sin(α/2) (1)
where l is the length of the leg from the trochanter (hip joint) to the ground, which can be measured directly or estimated as 53% of participant height, and a is the angular extent of the swing phase.
Determining stride length from leg swing alone is reasonably accurate for small stride lengths (<1 m) as there is minimal forward motion of the pelvis during the swing phase. However, this technique underestimates larger strides due to the considerable forward motion of the body over the stance foot in addition to the component generated by leg swing. To overcome this problem, a novel calibration algorithm was developed that could provide accurate stride length measurements from a single, shank-mounted gyroscope. See Moore et al., “Long-term Monitoring of Gait In Parkinson's Disease,” Gait & Posture, 26, pp 200-207 (2007).
A ‘one-size-fits-all’ group calibration algorithm was developed for clinical uses of the present invention and where individual calibrations were not practical (e.g., for patients with advanced PD). This “group calibration” algorithm utilizes a direct measure of the stride length obtained from 10 healthy participants walking along a 30 meter hallway. Healthy participants/controls were utilized as it was necessary to acquire angular velocity data over a wide range of stride lengths (˜0.2-1.5 m) from each participant.
An aluminum tube was taped to the heel of the left shoe and a whiteboard marker inserted such that the tip left a single dot on the floor during each foot placement (pen technique). Simultaneous estimates of stride length were obtained from the present invention attached to the left leg. Actual stride length was determined from measurement of the distance between successive dots on the floor. Participants were instructed to walk at a natural pace but to vary gait according to verbal commands to produce a wide range of stride lengths, including small shuffling steps typical of Parkinson's disease. The pen technique was chosen as it enabled the calibration over a wide range of stride lengths, was relatively accurate (˜5 mm error), and enabled calibration outside of the laboratory.
Plotting height-normalized true-versus-estimated stride lengths from the ten controls revealed a non-linear but consistent relationship, such that it was possible to generalize a correction algorithm applicable to all participants. To correct for forward motion of the body over the stance foot, a least-squares fit (Labview Advanced Analysis Package, National Instruments, Austin Tex.) was applied to the height-normalized initial stride length estimate (SLni) of the form:
where SLnc is the height-normalized corrected stride length, and the group calibration coefficients ai were [−43.34, 21.86, 14.91, −1.42, 2.25]. The mean error was 2.8% of participant height (maximum error 9%), or 5 cm for the average participant height of 167 cm.
Application of Equation (2) to the height-normalized initial stride estimates and multiplication by participant height yielded an accurate stride measure over the full range of stride length. The error per stride was also estimated by comparing the total distance traveled down the hallway (cumulative stride length of the true and corrected values) and dividing by the number of strides taken for each participant. Mean error was similar to that calculated from the height-normalized data at 4.8 cm, with a maximum error of 8 cm.
An alternative to a group calibration is to derive the coefficients of Equation (2) for each individual subject. Using the data from the ten control subjects, individual subject calibration algorithms were computed and were found to reduced the mean error by 33%, from 4.8 cm to 3.2 cm. Thus, if increased accuracy is required subjects can be individually calibrated rather than using the group calibration coefficients. However, this may not be possible for patients with advanced PD, who cannot vary stride over a sufficient range to provide adequate calibration.
The present invention (SAGE-M) was then used to obtain stride data from two PD participants in the ‘off’ state (no dopaminergic medication in the previous 12 hours). A participant with a relatively mild form of PD walked a distance of 4.5 m (5 strides) and simultaneous pen and SAGE-M measures of stride length (left leg) were obtained. The average stride length was 90.1 cm (pen) and 89.2 cm (SAGE-M). A second participant with severe locomotor impairment traversed a distance of 89 cm utilizing small shuffling steps (7 strides) that yielded an average stride length of 12.7 cm (video analysis) and 10.4 cm (SAGE-M), and the mean difference was 2.5 cm. Thus, at two extremes of locomotor impairment in the PD ‘off’ state, the accuracy of the present invention was within that established in the ten healthy controls.
Differences between healthy and Parkinsonian gait over extended periods were also monitored with the present invention. Over four hours a healthy participant covered a total of 3.9 km with 3071 strides. Stride length was stable at 1.5 m and consistent with the typical value for adult males. In contrast, four hours of data from a PD patient during normal daily activities outside of the clinic demonstrates the cardinal features of Parkinsonian gait; namely a small (˜0.5 m), highly variable stride length, covering a distance of 492 m with 923 strides.
Freezing of gait (FOG) and falls in PD patients are generally thought to be closely related; both occur sporadically, are often resistant to dopaminergic treatment, and greatly diminish quality of life. Recent studies have demonstrated a high-frequency movement of the leg (2-6 Hz) during FOG, which may be preceded by higher stride-to-stride variability. To date there is no objective measure of FOG and subsequent falls outside the laboratory.
Using a preferred embodiment of the present invention, a pilot study (N=11) demonstrated that FOG could be identified in PD subjects from the appearance of high frequency components (2-6 Hz band) in the vertical acceleration of the leg that were not apparent during quiet stance or walking. See
Thus, the potential exists for the present invention to be used for extended real-time monitoring of gait in PD that can both identify FOG and predict an impending FOG episode, based on high-frequency vertical leg acceleration and changes in stride length, respectively. See Moore et al., “Ambulatory Monitoring of Freezing of Gait in Parkinson's Disease,” Movement Disorders, 22, Suppl. 16, pp. S78-79(2007).
The foregoing is considered as illustrative only of the principles of the invention. Further, since numerous modifications and changes will readily occur to those skilled in the art, and because of the wide extent of the teachings disclosed herein, the foregoing disclosure should not be considered to limit the invention to the exact construction and operation shown and described herein. Accordingly, all suitable modifications and equivalents of the present disclosure may be resorted to and still considered to fall within the scope of the invention as hereinafter set forth in claims to the present invention.
This application claims the benefit of U.S. Provisional Patent Application No. 60/842,598, filed Sep. 6, 2006 by the present inventors.
| Number | Date | Country | |
|---|---|---|---|
| 60842598 | Sep 2006 | US |
