Gait Analysis and Fall Risk Assessment

People:

Overview:

Falls are associated with numerous physical and psychological morbidities, decreased quality of life, increased mortality, and high healthcare costs. Patients with end stage renal diseases (ESRD) on hemodialysis (HD) have high morbidity and mortality due to multiple causes, one of which is dramatically higher fall rates than the general population. The incidence of falls in HD patients has been found to be 1.18 falls/patient-year, in comparison to the people living in community who have incidence of falls ranging from 0.32-0.70 falls/person-year [1].

The mobility mechanisms that contribute to falls in this population must be understood if adequate interventions for fall prevention are to be achieved. Current methods for kinematic analysis involve videographic or optoelectronic systems, which is expensive, non-portable, and invasive, requiring the application of markers to the subject. This is an undesirably long process for patients, who are already burdened with multi-hour dialysis sessions up to three times per week. In contrast, the TEMPO platform for wearable, six degrees-of-freedom motion capture is being used to continuously and non-invasively collect gait and posture data on ESRD patients who are on HD.

In a recent preliminary study, we collected data from ESRD patients on HD before and after dialysis treatments. The TEMPO platform, as well as a custom torque-sensing leg brace, were used in performing four tests:

  • Get-Up-and-Go: movement time and walking time for a sit-to-stand and walk transition
  • Posture-Locomotion-Manual (PLM): measures postural function, gait, and a goal directed reaching arm movement and the efficacy with which these movements compose a whole-person dynamic performance
  • Strength test: plantarflexion strength is measured by the torque-sensing brace
  • Local Dynamic Stability: stability measurement during walking

Preliminary results show decreased degradation before and after dialysis with respect to get-up-and-go, PLM, and strength, consistent with expectations from traditional results, thus showing promise for fall risk assessment using portable, wearable technology.

Related Projects:

References:

1. Desmet, C., Beguin, C., Swine, C., and Jadoul, M. Falls in hemodialysis patients: Prospective study of incidence, risk factors, and complications. American Journal of Kidney Diseases 45, 1 (2005), 148-153.

Lockhart TE, Barth AT, Zhang X, Songra R, Lach J, Abdel-Rahman E.  2010.  Portable, Non-Invasive Fall Risk Assessment in End Stage Renal Disease Patients on Hemodialysis. Wireless Health.

Gait Analysis for Diagnosing Normal Pressure Hydrocephalus

NPH is a neuropathy caused by abnormal accumulation of cerebrospinal fluid (CSF) that normally surrounds the brain . The symptoms are usually described as a classic triad of gait disturbance, dementia or mental decline and urinary incontinence . To diagnose NPH, a high volume lumbar puncture (HVLP) procedure to remove excess fluid is usually given the suspected NPH subjects followed by the evaluation of clinical response to CSF removal. If diagnosed successfully as NPH, patients will be treated with an invasive, long term intervention – the ventriculo-peritoneal (VP) cerebral shunt to drain excess CSF to the abdomen where it is absorbed. However, risks associated with a cerebral shunt include intracranial hematoma, cerebral edema, crushed brain tissue and herniation, revealing the importance of an accurate diagnosis. Medical literature has suggested that improvement in gait pre- to post-HVLP is often a good marker of diagnosis in the decision to proceed with shunt surgery.  However, such gait improvement is usually based on clinical observation rather than objective, quantitative gait analysis, and even assessments performed with modern gait laboratory equipment only provide snapshots of the patient’s gait. This is especially problematic given that patients have variable HVLP response times, so longer continuous gait monitoring is necessary to provide high confidence NPH diagnosis.
  

This project seeks to evaluate the TEMPO system as an NPH diagnostic tool. As a inertial BSN system, TEMPO provides continuous and non-invasive gait data collection in any location over an extended period of time. Through an in-clinic human subjects pilot study (IRB approval has already been obtained), this project will lay the groundwork and provide preliminary data for a follow up study (and corresponding NIH proposal) in which TEMPO systems are deployed for gait data collection over the two days pre- and post-HVLP. This project will serve to evaluate continuous gait assessment using TEMPO and personalized signal processing techniques as a tool for improving physicians’ abilities to diagnose – and ultimately treat – NPH without the tremendous cost and inconvenience of inpatient monitoring or multiple outpatient visits, and with greater safety and comfort to an aging, stress-vulnerable, and growing population.

Research Plan

The proposed project is a one year pilot study for a subsequent multi-year study (and corresponding NIH R01 proposal) to fully evaluate the efficacy of continuous, non-invasive gait assessment for NPH diagnosis through multi-day data collections pre- and post-HPLV. Towards the preparation for that study, we propose the following research activities over the following year:

1.    Perform in-clinic TEMPO data collections on NPH patients (five nodes – both ankles, both wrists, and sacrum) during the pre- and post-HVLP gait assessments that are currently performed. TEMPO data collections will also be performed in VICON motion capture laboratories, which are the gold-standard for gait analysis. These collections will help inform, validate, and evaluate the use of TEMPO in NPH gait assessment.

2.    Develop a signal processing plan and perform a series of longitudinal data collection sessions on each patient to measure intra-individual gait variations (e.g., differences pre- and post-HVLP). The type of learning machine and data filtering process chosen will depend on both the nature of the information sought by the clinicians and the computational capabilities of TEMPO. We will consider a simple neural network, a support vector machine, the CMAX, as well as a myriad of feature extraction procedures (principal component analysis, information theoretic entropy maximizing techniques, wavelet decomposition, etc.). Both data compression and classification will be addressed. Performance in terms of classification error and loss of signal fidelity will be compared to similar algorithms trained across individuals and to more powerful techniques developed for non-resource-constrained machines.

 

 Future Work:

Prepare TEMPO for the multi-day data collections to be performed in the second year of the study. In particular, when TEMPO is operating at the highest data rate with all sensors active, the battery supply will be exhausted in approximately four hours. While this is sufficient for the short, in-clinic data collection sessions that will be performed in this pilot phase, the following steps will be taken to achieve the battery life extension necessary for the follow study:

A.   The longer-term data collections do not require real-time analysis. Instead, data can be processed on-node and the essential information downloaded from the system when the subjects return to the clinic. Therefore, no wireless transmission is necessary, which is the largest consumer of power in the TEMPO system. Writing to non-volatile flash memory consumes significantly less energy per bit than radio frequency transmissions. The TEMPO nodes will therefore be adapted accordingly – replacing radio transceivers with flash memory modules.

B.   We will explore how data can be compressed on-node without compromising the subsequent gait assessment fidelity. Compression can greatly reduce the number of bits that must be written to flash, saving both energy and memory capacity.

C.   Given that on-node signal processing can be used to easily determine if the wearer is walking, the system can enter a low power mode when possible, including stopping flash writes altogether and turning off the gyroscopes (the second largest power consumers in TEMPO). The microcontroller need only occasionally and briefly re-enter active mode and turn on the accelerometers to determine if the subject is walking and, if so, activate the rest of the system. This enables extended operation at well below 1 mW.

Reference:

Shrinivasan A, Brandt-Pearce M, Barth A, Lach J.  2011.  Analysis of gait in patients with normal pressure hydrocephalus. Proceedings of the First ACM Workshop on Mobile Systems, Applications, and Services for Healthcare. :3:1–3:6.

Gait Speed Estimation using Inertial Body Sensor Networks

 Gait speed is a particularly important parameter in geriatrics, as it is the number one predictor of mortality in adults over 65 years old, with differences of just a couple tenths of a meter per second predicting statistically significant outcome differences. The most common method for gait speed estimation in medical research and clinical practice is to simply use a stopwatch and a tape measure. This typically provides good accuracy but is insufficient for applications that require more continuous and longitudinal data, especially given that speed and many other gait parameters can vary significantly day-to-day and even hour-to-hour in geriatric and gait impaired populations. It is therefore highly desirable to be able to estimate gait speed using inertial BSNs and to do so with a resolution of better than 0.1 m/s.

However, for inertial BSNs, while simple temporal gait parameters such as step time and double stance time can be extracted from accelerometer and gyroscope data, parameters that depend on both temporal and spatial information are much more challenging to accurately assess due to integration drift (e.g., acceleration to velocity and position, rotational rate to angular displacement) and node placement uncertainty. And gait speed is one such spatio-temporal parameter, as it includes both stride length and stride time. With the abovementioned challenges, inertial BSNs have been prevented significant progress towards accurate gait speed estimation.

 

People:

Methodology:

To tackle this problem, mounting calibration using simple pre-defined movements and rotation matrices to ensure accurate spatial analysis regardless of how the BSN nodes are placed on-body. In addition, application specific methods are developed and applied that leverage knowledge of biomechanics and human gait – including temporal knowledge of gait phases – in order to minimize integration drift and to better model stride length. [1]

  • Mounting Calibration:

 

 

  • Gait Cycle Segmentation and Drift Elimination:

     Gait cycle extraction is critical to extract parameters such as gait phase, step time, and stride length, all of which are important for gait speed estimation. Based on the assumption that during the foot on ground event, the angular velocity should be near zero, a local maximum peak detection algorithm is selected for gait cycle extraction. (This portion of the gait cycle was chosen because it also supports integration drift cancelation) To suppress the ripples in the gyroscope signal, a zero-phase, 3rd order, Butterworth low-pass filter with a cutoff frequency of 3Hz is used. The cutoff frequency is determined empirically by inspecting the spectrum of the gyroscope signal, in which the main frequency components lie below 3Hz. The spike representing the Heel-Strike event is removed after the filtering, illuminating the foot-on-ground identified by the peak detection algorithm. Then the time point of foot-on-ground event is recorded and the original gyroscope signal is kept for later integration.

 

 

  • Refined Gait Model
To better examine the human gait model, the gait cycle is divided into 8 phases as shown in Figure 5. Research has shown that the angular velocity of the shank reaches its maximum when the leg is fully extended, and the angle of the shank reaches its maximum after this when the leg is flexed. These two events do not overlap in time as illustrated in Figure 5. and verified by the data in Figure 6. Thus using leg length and the maximum shank angle for computing step length during backward swing (the simplified pendulum model in Figure 2. ) is imprecise. This discrepancy suggests a more refined compound pendulum model to compute step length as shown in Figure 7.

 

        

As shown in Figure 7. , the step length calculation of our model differs from the model in the reference. One stride’s length is defined as the sum of the step length of the right leg and the step length of the left leg in one gait cycle. The total distance travelled is the sum of the stride lengths of all cycles. Finally, the average gait speed is the distance travelled divided by the total time elapsed.   

 

Results:

 

The RMSE is computed comparing treadmill speed, with a resolution of 0.2 MPH (0.09 m/s) from 1MPH to 3MPH, to the calculated gait speed. The accuracy of the proposed model was significantly higher than that of the reference model, which commonly overestimates gait speed.  The largest RMSE was only 0.095m/s after mounting calibration as shown in . However, at very low and high speeds, the thigh angle can be critical for controlling the step length. At very low speeds, the thigh tends to swing forward ahead of plumb line so as to maintain a very short step length on the treadmill, resulting in a step length that is shorter than predicted, and vice versa at high speeds. Thus, correction factors are needed to further reduce errors at very slow or fast walking speeds.

 

Future Work:

Work is underway to evaluate the estimation accuracy among various gaits, including both healthy and pathological gait at a greater range of speeds (including running), through experiments with more subjects. For healthy gait, a training set of data can be used to calibrate the algorithm for each individual subject. For certain types of pathological gait, including those with shuffling, a wide base, and out-of-plane motion, more refined gait models will be developed based on biomechanical knowledge.

 

 

Fall Risk Assessment in Parkinson's Patients

 Coming soon


References