Using Wearable Sensors and Machine Learning to Automatically Detect Freezing of Gait during a FOG-Provoking Test

Freezing of gait (FOG) is a debilitating motor phenomenon that is common among individuals with advanced Parkinson's disease. Objective and sensitive measures are needed to better quantify FOG. The present work addresses this need by leveraging wearable devices and machine-learning methods to develop and evaluate automated detection of FOG and quantification of its severity. Seventy-one subjects with FOG completed a FOG-provoking test while wearing three wearable sensors (lower back and each ankle). Subjects were videotaped before (OFF state) and after (ON state) they took their antiparkinsonian medications. Annotations of the videos provided the "ground-truth" for FOG detection. A leave-one-patient-out validation process with a training set of 57 subjects resulted in 84.1% sensitivity, 83.4% specificity, and 85.0% accuracy for FOG detection. Similar results were seen in an independent test set (data from 14 other subjects). Two derived outcomes, percent time frozen and number of FOG episodes, were associated with self-report of FOG. Bother derived-metrics were higher in the OFF state than in the ON state and in the most challenging level of the FOG-provoking test, compared to the least challenging level. These results suggest that this automated machine-learning approach can objectively assess FOG and that its outcomes are responsive to therapeutic interventions.

What happens before the first step? A New Approach to Quantifying Gait Initiation Using a Wearable Sensor

BACKGROUND

Walking is a volitional behavior that requires planning and initiation before a step is observed. Following a signal to begin walking, studies of gait initiation in specialized labs have identified three phases that occur during the transition from a standing position via anticipatory postural adjustment (APA) to the first step. Routine instrumented gait testing outside of the laboratory setting focuses on gait execution and does not include gait initiation measures.

RESEARCH QUESTION

Can a single IMU sensor be used for performing gait initiation evaluations outside the lab?

METHODS

We recorded walking in young (N = 41) and older (N = 26) adults using an instrumented gait mat while they were wearing a 3D accelerometer on their lower back. Subjects were instructed to begin walking following an auditory signal. An algorithm was developed to extract the following measures from the acceleration signal: gait initiation time, measured from the start of the auditory cue to begin walking and ends at the heel-strike of the swing leg, time-to-APA (reaction time), APA duration and swing time (execution of the first step).

RESULTS

Intraclass correlation coefficient analysis showed good to excellent agreement between gait initiation metrics obtained with the gait mat and the wearable sensor (mean 0.88, range [0.75-0.96]). Except for swing time, all measures were longer in the older subjects, compared to the young adults (p < 0.01).

SIGNIFICANCE

Extracting gait initiation measures from routine instrumented gait testing may facilitate studies that can better determine the extent to which impaired gait planning and execution contribute to mobility impairments.

AUTOMATIC QUANTIFICATION OF TANDEM WALKING USING A WEARABLE DEVICE: VALIDITY OF THE INSTRUMENTED TANDEM WALK

Tandem walk (TW) is typically assessed by the time to complete the task and the number of missteps, however, these measures suffer from limitations and may not fully capture the range of performance in this task. We developed metrics of TW by using a body-fixed, wearable sensor in young and older adults. Healthy young men (n=40) and older adult men (n=362) were studied. While wearing a 3D accelerometer on their lower back, subjects performed three different tasks: TW, usual-walking, and quiet standing. The extracted measures for TW were: High-to-Low frequency band ratio from the power spectral density from the ML axis [nu], signal vector magnitude[g], step duration[s], sample entropy from ML, AP axis[nu] and CV[%]. All of the TW metrics were significantly different in the young and older men (p<0.001). Older men completed the TW with higher CV, suggesting greater stride-to-stride variability and they walked more slowly, as seen by their step duration. Additionally, the frequency ratio measure suggests that the older adults displayed less complex corrective movements in the ML axis. TW measures were modestly correlated with usual-walking (e.g., average stride time with TW step time, r=0.3; p<0.001) and with quiet standing postural control (e.g., acceleration path length in the ML and AP axis with TW sample entropy in the ML axis, r=0.13; p=0.014). Metrics derived from a wearable device complement conventional TW measures and vary with age. Further work is needed to determine if TW, gait and posture metrics are differentially associated with distinct adverse health outcomes.

A new approach to quantifying the EEG during walking: Initial evidence of gait related potentials and their changes with aging and dual tasking

BACKGROUND

The electroencephalogram (EEG) can be a useful tool to investigate the neurophysiology of gait during walking. Our aims were to develop an approach that identify and quantify event related potentials (ERPs) during a gait cycle and to examine the effects of aging and dual tasking on these gait related potentials (GRPs).

METHODS

10 young and 10 older adults walked on a treadmill while wearing a wireless 20-channels EEG and accelerometers on the ankles. Each heel strike extracted from the accelerometers was used as an event to which the electrical brain activity pattern was locked. The subjects performed usual and dual task walking that included an auditory oddball task. GRPs amplitude and latency were computed, and a new measure referred to as Amplitude Pattern Consistency (APC) was developed to quantify the consistency of these GRP amplitudes within a gait cycle. The results were compared between and within groups using linear mixed model analysis.

RESULTS

The electrical pattern during a gait cycle consisted of two main positive GRPs. Differences in these GRPs between young and older adults were observed in Pz and Cz. In Pz, older adults had higher GRPs amplitude (p = 0.006, p = 0.010), and in Cz lower APC (p = 0.025). Alterations were also observed between the walking tasks. Both groups showed shorter latency during oddball walking compared to usual walking in Cz (p = 0.040). In addition, the APC in Cz was correlated with gait speed (r = 0.599, p = 0.011) in all subjects and with stride time variability in the older adults (r = -0.703, p = 0.023).

CONCLUSIONS

This study is the first to define specific gait related potentials within a gait cycle using novel methods for quantifying waveforms. Our findings show the potential of this approach to be applied broadly to study the EEG during gait in a variety of contexts. The observed changes in GRPs with aging and walking task and the relationship between GRPs and gait may suggest the neurophysiologic foundation for studying walking and for developing new approaches for improving gait.

Using wearables to assess bradykinesia and rigidity in patients with Parkinson's disease: a focused, narrative review of the literature

The potential of using wearable technologies for the objective assessment of motor symptoms in Parkinson's disease (PD) has gained prominence recently. Nonetheless, compared to tremor and gait impairment, less emphasis has been placed on the quantification of bradykinesia and rigidity. This review aimed to consolidate the existing research on objective measurement of bradykinesia and rigidity in PD through the use of wearables, focusing on the continuous monitoring of these two symptoms in free-living environments. A search of PubMed was conducted through a combination of keyword and MeSH searches. We also searched the IEEE, Google Scholar, Embase, and Scopus databases to ensure thorough results and to minimize the chances of missing relevant studies. Papers published after the year 2000 with sample sizes greater than five were included. Studies were assessed for quality and information was extracted regarding the devices used and their location on the body, the setting and duration of the study, the "gold standard" used as a reference for validation, the metrics used, and the results of each paper. Thirty-one and eight studies met the search criteria and evaluated bradykinesia and rigidity, respectively. Several studies reported strong associations between wearable-based measures and the gold-standard references for bradykinesia, and, to a lesser extent, rigidity. Only a few, pilot studies investigated the measurement of bradykinesia and rigidity in the home and free-living settings. While the current results are promising for the future of wearables, additional work is needed on their validation and adaptation in ecological, free-living settings. Doing so has the potential to improve the assessment and treatment of motor fluctuations and symptoms of PD more generally through real-time objective monitoring of bradykinesia and rigidity.

Is every-day walking in older adults more analogous to dual-task walking or to usual walking? Elucidating the gaps between gait performance in the lab and during 24/7 monitoring

Background

The traditional evaluation of gait in the laboratory during structured testing has provided important insights, but is limited by its “snapshot” character and observation in an unnatural environment. Wearables enable monitoring of gait in real-world environments over a week. Initial findings show that in-lab and real-world measures differ. As a step towards better understanding these gaps, we directly compared in-lab usual-walking (UW) and dual-task walking (DTW) to daily-living measures of gait.

Methods

In-lab gait features (e.g., gait speed, step regularity, and stride regularity) derived from UW and DTW were compared to the same gait features during daily-living in 150 elderly fallers (age: 76.5 ± 6.3 years, 37.6% men). In both settings, features were extracted from a lower-back accelerometer. In the real-world setting, subjects were asked to wear the device for 1 week and pre-processing detected 30-s daily-living walking bouts. A histogram of all walking bouts was determined for each walking feature for each subject and then each subject’s typical (percentile 50, median), worst (percentile 10) and the best (percentile 90) values over the week were determined for each feature. Statistics of reliability were assessed using Intra-Class correlations and Bland-Altman plots.

Results

As expected, in-lab gait speed, step regularity, and stride regularity were worse during DTW, compared to UW. In-lab gait speed, step regularity, and stride regularity during UW were significantly higher (i.e., better) than the typical daily-living values (p < 0.001) and different (p < 0.001) from the worst and best values. DTW values tended to be similar to typical daily-living values (p = 0.205, p = 0.053, p = 0.013 respectively). ICC assessment and Bland-Altman plots indicated that in-lab values do not reliably reflect the daily-walking values.

Conclusions

Gait values measured during relatively long (30-s) daily-living walking bouts are more similar to the corresponding values obtained in the lab during dual-task walking, as compared to usual walking. Still, gait performance during most daily-living walking bouts is worse than that measured during usual and dual-tasking in the lab. The values measured in the lab do not reliably reflect daily-living measures. That is, an older adult’s typical daily-living gait cannot be estimated by simply measuring walking in a structured, laboratory setting.

Changes in event-related potentials during dual task walking in aging and Parkinson's disease

OBJECTIVE

To investigate EEG changes during an auditory odd-ball task while walking (dual-task) in young adults, older adults, and patients with Parkinson's disease.

METHODS

11 young adults, 10 older adults, and 10 patients with Parkinson's disease (PD) performed an auditory oddball task during standing and walking on a treadmill, while wearing a wireless EEG cap. The amplitude and latency of P300 were compared between groups and within conditions using linear mix model analysis. Gait was evaluated using wearable sensors and cognition was assessed using the Color Trail Test.

RESULTS

P300 latency became longer during walking in all groups (p = 0.005). During walking, older adults (p = 0.005) and patients with PD (p = 0.001) showed prolonged P300 latency compared to young adults. Significant task by group interaction was found in P300 amplitude (p = 0.008). Patients with PD demonstrated reduced P300 amplitude during walking compared to standing (p = 0.023). Among all subjects, better motor and cognitive performance correlated with shorter P300 latency (r = 0.457, p = 0.014 and r = 0.431, p = 0.040, respectively).

CONCLUSIONS

These findings provide direct evidence of the physiological recruitment of attentional networks during walking and their impact by ageing and disease.

SIGNIFICANCE

This study is the first to report on changes in P300 latency and amplitude during dual-task oddball walking in older adults and patients with PD.

Treadmill walking reduces pre-frontal activation in patients with Parkinson's disease

BACKGROUND

Among patients with Parkinson's disease (PD), gait is typically disturbed and less automatic. These gait changes are associated with impaired rhythmicity and increased prefrontal activation, presumably in an attempt to compensate for reduced automaticity.

RESEARCH QUESTION

We investigated whether during treadmill walking, when the pace is determined and fixed, prefrontal activation in patients with PD is lower, as compared to over-ground walking.

METHODS

Twenty patients with PD (age: 69.8 ± 6.5 yrs.; MoCA: 26.9 ± 2.4; disease duration: 7.9 ± 4.2 yrs) walked at a self-selected walking speed over-ground and on a treadmill. A wireless functional near infrared spectroscopy (fNIRS) system measured prefrontal lobe activation, i.e., oxygenated hemoglobin (Hb02) in the pre-frontal area. Gait was evaluated using 3D-accelerometers attached to the lower back and ankles (Opal™, APDM). Dynamic gait stability was assessed using the maximum Lyapunov exponent to investigate automaticity of the walking pattern.

RESULTS

Hb02 was lower during treadmill walking than during over-ground walking (p = 0.001). Gait stability was greater on the treadmill, compared to over-ground walking, in both the anteroposterior and medio-lateral axes (p < 0.001).

SIGNIFICANCE

These findings support the notion that when gait is externally paced, prefrontal lobe activation is reduced in patients with PD, perhaps reflecting a reduced need for compensatory cognitive mechanisms.

Age-associated changes in obstacle negotiation strategies: Does size and timing matter?

INTRODUCTION

Tripping over an obstacle is one of the most common causes of falls among older adults. However, the effects of aging, obstacle height and anticipation time on negotiation strategies have not been systematically evaluated.

METHODS

Twenty older adults (ages: 77.7±3.4years; 50% women) and twenty young adults (age: 29.3±3.8years; 50% women) walked through an obstacle course while negotiating anticipated and unanticipated obstacles at heights of 25mm and 75mm. Kinect cameras captured the: (1) distance of the subject's trailing foot before the obstacles, (2) distance of the leading foot after the obstacles, (3) clearance of the leading foot above the obstacles, and (4) clearance of the trailing foot above the obstacles. Linear-mix models assessed changes between groups and conditions.

RESULTS

Older adults placed their leading foot closer to the obstacle after landing, compared to young adults (p<0.001). This pattern was enhanced in high obstacles (group*height interaction, p=0.033). Older adults had lower clearance over the obstacles, compared to young adults (p=0.007). This was more pronounced during unanticipated obstacles (group*ART interaction, p=0.003). The distance of the leading foot and clearance of the trailing foot after the obstacles were correlated with motor, cognitive, and functional abilities.

CONCLUSIONS

These findings suggest that there are age-related changes in obstacle crossing strategies that are dependent on the specific characteristics of the obstacle. The results have important implications for clinical practice, suggesting that functional exercise should include obstacle negotiation training with variable practice of height and available response times. Further studies are needed to better understand the effects of motor and cognitive abilities.

Identification of Characteristic Motor Patterns Preceding Freezing of Gait in Parkinson's Disease Using Wearable Sensors

Freezing of gait (FOG) is a disabling symptom that is common among patients with advanced Parkinson’s disease (PD). External cues such as rhythmic auditory stimulation can help PD patients experiencing freezing to resume walking. Wearable systems for automatic freezing detection have been recently developed. However, these systems detect a FOG episode after it has happened. Instead, in this study, a new approach for the prediction of FOG (before it actually happens) is presented. Prediction of FOG might enable preventive cueing, reducing the likelihood that FOG will occur. Moreover, understanding the causes and circumstances of FOG is still an open research problem. Hence, a quantitative characterization of movement patterns just before FOG (the pre-FOG phase) is of great importance. In this study, wearable inertial sensors were used to identify and quantify the characteristics of gait during the pre-FOG phase and compare them with the characteristics of gait that do not precede FOG. The hypothesis of this study is based on the threshold-based model of FOG, which suggests that before FOG occurs, there is a degradation of the gait pattern. Eleven PD subjects were analyzed. Six features extracted from movement signals recorded by inertial sensors showed significant differences between gait and pre-FOG. A classification algorithm was developed in order to test if it is feasible to predict FOG (i.e., detect it before it happens). The aim of the classification procedure was to identify the pre-FOG phase. Results confirm that there is a degradation of gait occurring before freezing. Results also provide preliminary evidence on the feasibility of creating an automatic algorithm to predict FOG. Although some limitations are present, this study shows promising findings for characterizing and identifying pre-FOG patterns, another step toward a better understanding, prediction, and prevention of this disabling symptom.

Triple paternal contribution to a normal/complete molar chimeric singleton placenta

A comprehensive study of unusual cases of placental pathology may provide insight into mechanisms of normal human fertilization and early embryonic development by examining the exception to the rule. A gravida three para two 39-year-old woman was monitored by ultrasound from 16 weeks of gestation for cystic placenta. A female newborn was born at 36 weeks gestation. Pathologic examination of the partially cystic placenta revealed a singleton placenta comprised of 2/3 normal placenta and 1/3 complete hydatidiform mole, largely degenerated. Immunostaining for p57 was negative in stromal cells of the molar villi. Chromogenic in-situ hybridization revealed diploidy in both normal and molar parts. A total of 16 microsatellites were studied by short tandem repeat analysis, 11 of which were informative. The analysis revealed bipaternal molar tissue of dispermic origin. The paternal monospermic contribution to the normal part was different from that in the molar part, thus resulting in tripaternal contribution to the conceptus. A chimera is a single organism composed of two or more different populations of genetically distinct cells that originated from different zygotes (tetragametic) whereas mosaic is a mixture of two cell lines in one organism originating from one zygote. The possible mechanisms leading to the formation of chimeric/mosaic placenta in our case (one of the components being complete hydatidiform mole), including twinning, fusion at an early embryonic stage and diploidization of triploids, are discussed.

Hausdorff J, Tröster G. A Wearable Assistant for Gait Training for Parkinson’s Disease with Freezing of Gait in Out-of-the-Lab Environments

People with Parkinson’s disease (PD) suffer from declining mobility capabilities, which cause a prevalent risk of falling. Commonly, short periods of motor blocks occur during walking, known as freezing of gait (FoG). To slow the progressive decline of motor abilities, people with PD usually undertake stationary motor-training exercises in the clinics or supervised by physiotherapists. We present a wearable system for the support of people with PD and FoG. The system is designed for independent use. It enables motor training and gait assistance at home and other unsupervised environments. The system consists of three components. First, FoG episodes are detected in real time using wearable inertial sensors and a smartphone as the processing unit. Second, a feedback mechanism triggers a rhythmic auditory signal to the user to alleviate freeze episodes in an assistive mode. Third, the smartphone-based application features support for training exercises. Moreover, the system allows unobtrusive and long-term monitoring of the user’s clinical condition by transmitting sensing data and statistics to a telemedicine service.

We investigate the at-home acceptance of the wearable system in a study with nine PD subjects. Participants deployed and used the system on their own, without any clinical support, at their homes during three protocol sessions in 1 week. Users’ feedback suggests an overall positive attitude toward adopting and using the system in their daily life, indicating that the system supports them in improving their gait. Further, in a data-driven analysis with sensing data from five participants, we study whether there is an observable effect on the gait during use of the system. In three out of five subjects, we observed a decrease in FoG duration distributions over the protocol days during gait-training exercises. Moreover, sensing data-driven analysis shows a decrease in FoG duration and FoG number in four out of five participants when they use the system as a gait-assistive tool during normal daily life activities at home.

Changes in oxygenated hemoglobin link freezing of gait to frontal activation in patients with Parkinson disease: an fNIRS study of transient motor-cognitive failures

Recent studies have suggested that deficits in executive function contribute to freezing of gait (FOG), an episodic disturbance common among patients with Parkinson's disease (PD). To date, most findings provide only indirect evidence of this relationship. Here, we evaluated a more direct link between FOG and frontal lobe dysfunction. Functional, near infrared spectroscopy measured frontal activation, i.e., oxygenated hemoglobin (HbO2) levels in Brodmann area 10 before and during FOG. Eleven patients with PD and eleven healthy older adults were studied. Changes in frontal lobe activation before and during FOG that occurred during turns were determined. Altogether, 49 FOG episodes were observed-28 occurred during turns that were anticipated (i.e., the patient knew in advance that the turn was coming), 21 during unanticipated turns that were performed "abruptly", according to the examiner's request. During anticipated turns, HbO2 increased by 0.22 ± 0.08 µM (p = 0.004) before FOG and by an additional 0.19 ± 0.13 µM (p = 0.072) during FOG. In contrast, during unanticipated turns, HbO2 did not increase before or during FOG. HbO2 decreased by 0.32 ± 0.08 µM (p = 0.004) during turns without FOG; in healthy controls HbO2 did not change during turns. These findings support the existence of an association between FOG episodes and changes in frontal lobe HbO2. Increased activation in Brodmann area 10 before FOG, specifically during anticipated turns, highlights the connections between motor planning, information processing, and FOG. These results support the idea that alterations in executive control play a role in this debilitating motor disturbance.

Automated detection of missteps during community ambulation in patients with Parkinson's disease: a new approach for quantifying fall risk in the community setting

BACKGROUND

Falls are a leading cause of morbidity and mortality among older adults and patients with neurological disease like Parkinson's disease (PD). Self-report of missteps, also referred to as near falls, has been related to fall risk in patients with PD. We developed an objective tool for detecting missteps under real-world, daily life conditions to enhance the evaluation of fall risk and applied this new method to 3 day continuous recordings.

METHODS

40 patients with PD (mean age ± SD: 62.2 ± 10.0 yrs, disease duration: 5.3 ± 3.5 yrs) wore a small device that contained accelerometers and gyroscopes on the lower back while participating in a protocol designed to provoke missteps in the laboratory. Afterwards, the subjects wore the sensor for 3 days as they carried out their routine activities of daily living. An algorithm designed to automatically identify missteps was developed based on the laboratory data and was validated on the 3 days recordings.

RESULTS

In the laboratory, we recorded 29 missteps and more than 60 hours of data. When applied to this dataset, the algorithm achieved a 93.1% hit ratio and 98.6% specificity. When we applied this algorithm to the 3 days recordings, patients who reported two falls or more in the 6 months prior to the study (i.e., fallers) were significantly more likely to have a detected misstep during the 3 day recordings (p = 0.010) compared to the non-fallers.

CONCLUSIONS

These findings suggest that this novel approach can be applied to detect missteps during daily life among patients with PD and will likely help in the longitudinal assessment of disease progression and fall risk.

The RNA polymerase flow model of gene transcription

Gene expression is a fundamental cellular process by which proteins are synthesized based on the information coded in the genes. The two major steps of this process are the transcription of the DNA segment corresponding to a gene to mRNA molecules and the translation of the mRNA molecules to proteins by the ribosome. Thus, understanding, modeling and engineering the different stages of this process have both important biotechnological applications and contributions to basic life science. In previous studies we have introduced the Homogenous Ribosome Flow Model (HRFM) and demonstrated its advantages in analyses of the translation process. In this study we introduce the RNA Polymerase Flow Model (RPFM), a non trivial extension of the HRFM, which also includes a backward flow and can be used for modeling transcription and maybe other similar processes. We compare the HRFM and the RPFM in the three regimes of the transcription process: rate limiting initiation, rate limiting elongation and rate limiting termination via a simulative and analytical analysis. In addition, based on experimental data, we show that RPFM is a better choice for modeling transcription process.

Living related liver transplant following bone marrow transplantation from same donor: long-term survival without immunosuppression

We report long-term (seven yr) immunological tolerance in a 16-yr-old boy, to a liver allograft donated by his father following a bone marrow transplant at age 2.5 yr from the same donor. The bone marrow transplant was complicated by severe GVHD leading to liver failure and the ensuing need for a liver transplant, performed under planned avoidance of immunosuppression. At one wk post-transplant, although a liver biopsy was histologically compatible with acute rejection, favorable clinical and biochemical evolution precluded initiating immunosuppressive therapy, thus highlighting the need for caution when interpreting early histological changes so that administration of unnecessary immunosuppression can be avoided. Induction of tolerance in transplant recipients remains an elusive goal. In those patients who had received conventional bone marrow transplants and had endured the consequences of GVHD, development of macrochimerism may allow immunosuppression-free solid organ transplantation from the same donor.

Patterned arrays of ordered peptide nanostructures

Methods for the deposition of ordered nanostructures on various substrates are a key factor in nanotechnological devices. There is a special interest in the development of methods for the organization of organic nanostructures that are not compatible with some of the conventional fabrication methods. The unique chemical and physical properties of the peptide nanotubes make them excellent component in various devices and the useful application were already demonstrated in the case of biosensors. Here we demonstrate the ability to deposited aromatic dipeptide nanotubes using electron beam treatment of surfaces to control their wettability. The use of a low energy electron irradiation results in the formation of pre-defined surfaces with controlled level of wettability. This treatment allows the precise patterning of the organic tubular assemblies at high resolution. The differential wettability of the surface resulted in organization of the peptide assemblies according to the properties of the different areas of the surface. In the current work, we describe the use of wettability patterned surfaces for the control patterning of horizontal peptide nanotubes and nanospheres. Furthermore, lift-off lithography is used to make patterned arrays of peptide nano-forests, vertically aligned peptide nanotubes. In summary, the novel patterning techniques together with the unique properties of the peptide nanostructures represent an important step in the integration of these assemblies into functional nanosystems and devices.

HLA-G levels in serum and plasma

Soluble human leukocyte antigen (HLA)-G levels are in most cases higher in the plasma than in the serum obtained from the same individual. This is probably due to trapping of the protein during clot formation. In studies where soluble HLA-G is quantified, it is therefore recommended that plasma or serum levels should be compared with the same blood product, namely, serum to serum and plasma to plasma. Because of possible gender differences in HLA-G levels it is also recommended that this should be considered in the construction of a control group especially in studies where there is a preponderance of one of the sexes.