Quantification of human movement is a challenge in many areas, ranging from physical therapy to robotics. We quantify of human movement for the purpose of providing automated exercise coaching in the home. We developed a model-based assessment and inference process that combines biomechanical constraints with movement assessment based on the Microsoft Kinect camera. To illustrate the approach, we quantify the performance of a simple squatting exercise using two model-based metrics that are related to strength and endurance, and provide an estimate of the strength and energy-expenditure of each exercise session. We look at data for 5 subjects, and show that for some subjects the metrics indicate a trend consistent with improved exercise performance.
Early and reliable detection of cognitive decline is one of the most important challenges of current healthcare. In this project we developed an approach whereby a frequently played computer game can be used to assess a variety of cognitive processes and estimate the results of the pen-and-paper Trail-Making Test (TMT) – known to measure executive function, as well as visual pattern recognition, speed of processing, working memory, and set-switching ability. We developed a computational model of the TMT based on a decomposition of the test into several independent processes, each characterized by a set of parameters that can be estimated from play of a computer game designed to resemble the TMT. An empirical evaluation of the model suggests that it is possible to use the game data to estimate the parameters of the underlying cognitive processes and using the values of the parameters to estimate the TMT performance. Cognitive measures and trends in these measures can be used to identify individuals for further assessment, to provide a mechanism for improving the early detection of neurological problems, and to provide feedback and monitoring for cognitive interventions in the home.
We explored the relationship between sleep disturbances and mild cognitive impairment (MCI) in community-dwelling seniors. Recent evidence suggests that sleep habits are differentially compromised in different subtypes of MCI, but the relationship between sleep disruption and MCI remains poorly understood. We gathered daily objective measures of sleep disturbance from 45 seniors, including 16 with MCI (mean age 86.9 ± 4.3 years), over a six month period. We also collected self-report measures of sleep disturbance. Although there were no differences between groups in any of our self-report measures, we found that amnestic MCI (aMCI) volunteers had less disturbed sleep than both non-amnestic MCI (naMCI) and cognitively intact volunteers, as measured objectively by movement in bed at night (F2,1078=4.30, p=0.05), wake after sleep onset (F2,1078=41.6, p<0.001), and times up at night (F2,1078=26.7, p<0.001). The groups did not differ in total sleep time. In addition, the aMCI group had less day-to-day variability in these measures than the intact and naMCI volunteers. In general, the naMCI volunteers showed a level of disturbed sleep that was intermediate to that of aMCI and intact volunteers. These differences in sleep disruption between aMCI and naMCI may be related to differences in the pathology underlying these MCI subtypes.
MCI (Mild Cognitive Impairment); Assessment of cognitive disorders/dementia; Sleep Habits; Cohort studies
Creative use of new mobile and wearable health information and sensing technologies (mHealth) has the potential to reduce the cost of health care and improve well-being in numerous ways. These applications are being developed in a variety of domains, but rigorous research is needed to examine the potential, as well as the challenges, of utilizing mobile technologies to improve health outcomes. Currently, evidence is sparse for the efficacy of mHealth. Although these technologies may be appealing and seemingly innocuous, research is needed to assess when, where, and for whom mHealth devices, apps, and systems are efficacious.
In order to outline an approach to evidence generation in the field of mHealth that would ensure research is conducted on a rigorous empirical and theoretic foundation, on August 16, 2011, researchers gathered for the mHealth Evidence Workshop at NIH. The current paper presents the results of the workshop. Although the discussions at the meeting were cross-cutting, the areas covered can be categorized broadly into three areas: (1) evaluating assessments; (2) evaluating interventions; and, (3) reshaping evidence generation using mHealth. This paper brings these concepts together to describe current evaluation standards, future possibilities and set a grand goal for the emerging field of mHealth research.
With the rising age of the population, there is increased need to help elderly maintain their independence. Smart homes, employing passive sensor networks and pervasive computing techniques, enable the unobtrusive assessment of activities and behaviors of the elderly which can be useful for health state assessment and intervention. Due to the multiple health benefits associated with socializing, accurately tracking whether an individual has visitors to their home is one of the more important aspects of elders’ behaviors that could be assessed with smart home technology. With this goal, we have developed a preliminary SVM model to identify periods where untagged visitors are present in the home. Using the dwell time, number of sensor firings, and number of transitions between major living spaces (living room, dining room, kitchen and bathroom) as features in the model, and self report from two subjects as ground truth, we were able to accurately detect the presence of visitors in the home with a sensitivity and specificity of 0.90 and 0.89 for subject 1, and of 0.67 and 0.78 for subject 2, respectively. These preliminary data demonstrate the feasibility of detecting visitors with in-home sensor data, but highlight the need for more advanced modeling techniques so the model performs well for all subjects and all types of visitors.
Quality of sleep is an important attribute of an individual’s health state and its assessment is therefore a useful diagnostic feature. Changes in the patterns of mobility in bed during sleep can be a disease marker or can reflect various abnormal physiological and neurological conditions. This paper describes a method for detection of movement in bed that is evaluated on data collected from patients admitted for regular polysomnography. The system is based on load cells installed at the supports of a bed. Since the load cell signal varies the most during movement, the approach uses a weighted combination of the short-term mean-square differences of each load cell signal to capture the variations in the signal caused by movement. We use a single univariate Gaussian model to represent each class: movement versus non-movement. We assess the performance of the method against manual annotation performed by a sleep clinic technician from seventeen patients. The proposed detection method achieved an overall sensitivity of 97.9% and specificity of 98.7%.
Physical performance measures predict health and function in older populations. Walking speed in particular has consistently predicted morbidity and mortality. However, single brief walking measures may not reflect a person’s typical ability. Using a system that unobtrusively and continuously measures walking activity in a person’s home we examined walking speed metrics and their relation to function. In 76 persons living independently (mean age, 86) we measured every instance of walking past a line of passive infra-red motion sensors placed strategically in their home during a four-week period surrounding their annual clinical evaluation. Walking speeds and the variance in these measures were calculated and compared to conventional measures of gait, motor function and cognition. Median number of walks per day was 18 ± 15. Overall mean walking speed was 61 ± 17 cm/sec. Characteristic fast walking speed was 96 cm/sec. Men walked as frequently and fast as women. Those using a walking aid walked significantly slower and with greater variability. Morning speeds were significantly faster than afternoon/evening speeds. In-home walking speeds were significantly associated with several neuropsychological tests as well as tests of motor performance. Unobtrusive home walking assessments are ecologically valid measures of walking function. They provide previously unattainable metrics (periodicity, variability, range of minimum and maximum speeds) of everyday motor function.
Gait; Home-based clinical assessment; Technology
Motor speed is an important indicator and predictor of both cognitive and physical function. One common assessment of motor speed is the finger tapping test (FTT), which is typically administered as part of a neurological or neuropsychological assessment. However, the FTT suffers from several limitations including infrequent in-person administration, the need for a trained assessor and dedicated equipment, and potential short term sensory-motor fatigue. In this article, we propose an alternative method of measuring motor speed, with face validity to the FTT, that addresses these limitations by measuring the interkeystroke interval (IKI) of familiar and repeated login data collected in the home during a subject’s regular computer use. We show significant correlations between the mean tapping speed from the FTT and the median IKIs of the non-dominant (r=0.77) and dominant (r=0.70) hands, respectively, in an elderly cohort of subjects living independently. Finally, we discuss how the proposed method for measuring motor speed fits well into the framework of unobtrusive and continuous in-home assessment.
Motor speed; Typing; Finger tapping test
An important component of future proactive healthcare is the detection of changes in the individual’s physical or cognitive performance, especially for aging and for those with neurodegenerative diseases. For a variety of reasons, the current techniques for neuropsychological assessment are not suitable for continuous monitoring and assessment. This paper proposes a technique for continuous monitoring of behaviors that could potentially be used to complement the traditional assessment techniques. In particular we monitor the movements of a computer pointing device (mouse) to assess cognitive and sensory-motor functionality of human users unobtrusively. The focus of this paper is on an approach that can be used to identify moves so that they can later be used as the basis for constructing sensory-motor measures. Due to the nature of the data the distinction between moves and pauses between moves is not immediately apparent. The segmentation of the data into moves is done by constructing an estimated distribution of the mouse cursor velocity for the entire computer session and locating a particular minimum which indicates a likely point of division between active moves and inter-move intervals. We analyzed computer usage data for 113 elderly participants over a period of two years, and the technique applied to that data was able to divide data from a session of computer usage into a series of mouse moves in 98% of observed computer sessions with a physically sensible value for the cutoff dividing moves from stops.
Gait velocity has repeatedly been shown to be an important indicator and predictor of both cognitive and physical function, especially in elderly. However, clinical gait assessments are conducted infrequently and cannot distinguish between abrupt changes in function and changes that occur more slowly over time. Collecting gait measurements continuously in-home has recently been proposed and validated to overcome these clinical limitations. In this paper, we describe the longitudinal analysis of in-home gait velocity collected unobtrusively from passive infrared motion sensors. We first describe a model for the probability density function of the in-home gait velocities. We then describe estimation of the evolution of the density function over time and report empirically determined algorithm parameters that have performed well over a wide variety of different gait velocity data. Finally, we demonstrate how this approach allows detection of significant events (abrupt changes in function) and slower changes over time in gait velocity data collected from a sample of two elderly subjects in the Intelligent Systems for Assessing Aging Changes (ISAAC) study.
Ubiquitous and unobtrusive in-home monitoring has the potential to detect physical and mental decline earlier and with higher precision than current clinical methods. However, given that this field is in its infancy, the specific metrics through which these changes are detected are not well defined. The work presented here offers room-transitions, the act of physically moving from one area of a home to another, as a quantifiable measure for total daily activity that can be inferred from a network of passive infrared sensors. We describe a method to calculate this value from raw sensor data and validate this method on an acute health event: an 18-day quarantine at a retirement community that was initiated in the midst of a norovirus outbreak. The results from this case study show that room-transition values increased significantly as subjects remained in their homes during the quarantine, demonstrating a mean increase of 12 transitions per day. Furthermore, a time-adjusted measure of room-transitions is examined that did not significantly change across the group. Finally, the healthy subjects and those that fell ill were analyzed separately, and significant differences were found between them for both the raw and time-adjusted metrics. As detection algorithms improve, these types of measures may be useful in the early detection of a change in health status.
Sleep disturbances are prevalent, financially taxing, and have a negative effect on health and quality of life. One of the most common sleep disturbances is obstructive sleep apnea-hypopnea syndrome (OSAHS) which frequently goes undiagnosed. The gold standard for diagnosing OSAHS is polysomnography (PSG)--a procedure that is inconvenient, time-consuming, and interferes with normal sleep patterns. We are investigating an alternative to PSG in which unobtrusive load cells fitted under the bed are used to monitor movement, heart rate, and respiration. In this paper we describe how load cell data can be used to distinguish between clinically relevant disordered breathing (apneas and hypopneas) and normal respiration. The method correctly classified disordered breathing segments with a sensitivity of 0.77 and a specificity of 0.91.
To determine whether low concentrations of a dopamine agonist worsen parkinsonism, which would suggest that activation of presynaptic dopamine autoreceptors causes a super-off state.
Randomized, double-blind, placebo-controlled, crossover clinical trial.
Academic movement disorders center.
Patients with Parkinson disease and motor fluctuations.
Fourteen patients with Parkinson disease and motor fluctuations were randomized to receive 1 of 6 possible sequences of placebo, low-dose (sub-threshold) apomorphine hydrochloride, and high-dose (threshold to suprathreshold) apomorphine hydrochloride infusions. Subthreshold doses of apomorphine hydrochloride (12.5 μg/kg/h every 2 hours and 25 μg/kg/h every 2 hours), threshold to suprathreshold doses of apomorphine hydrochloride (50 μg/kg/h every 2 hours and 100 μg/kg/h every 2 hours), and placebo were infused for 4 hours daily for 3 consecutive days.
Main Outcome Measures
Finger and foot tapping rates.
There was no decline in finger or foot tapping rates during the low-dose apomorphine hydrochloride infusions relative to placebo. The high-dose infusions increased foot tapping (P<.001) and trended toward increasing finger tapping compared with placebo infusions.
Subthreshold concentrations of apomorphine did not worsen parkinsonism, suggesting that pre-synaptic dopamine autoreceptors are not important to the motor response in moderate to advanced Parkinson disease.
To describe a longitudinal community cohort study, Intelligent Systems for Assessing Aging Changes, that has deployed an unobtrusive home-based assessment platform in many seniors homes in the existing community.
Several types of sensors have been installed in the homes of 265 elderly persons for an average of 33 months. Metrics assessed by the sensors include total daily activity, time out of home, and walking speed. Participants were given a computer as well as training, and computer usage was monitored. Participants are assessed annually with health and function questionnaires, physical examinations, and neuropsychological testing.
Mean age was 83.3 years, mean years of education was 15.5, and 73% of cohort were women. During a 4-week snapshot, participants left their home twice a day on average for a total of 208 min per day. Mean in-home walking speed was 61.0 cm/s. Participants spent 43% of days on the computer averaging 76 min per day.
These results demonstrate for the first time the feasibility of engaging seniors in a large-scale deployment of in-home activity assessment technology and the successful collection of these activity metrics. We plan to use this platform to determine if continuous unobtrusive monitoring may detect incident cognitive decline.
Cognitive assessment; Dementia; Home-based clinical assessment; Technology
Executive dysfunction has previously been found to be a risk factor for falls. The aim of this study is to investigate the association between executive dysfunction and risk of falling and to determine if this association is independent of balance.
Participants were 188 community-dwelling individuals aged 65 and older. All participants underwent baseline and annual evaluations with review of health history, standardized neurologic examination, neuropsychological testing, and qualitative and quantitative assessment of motor function. Falls were recorded prospectively using weekly online health forms.
During 13 months of follow-up, there were 65 of 188 participants (34.6%) who reported at least one fall. Univariate analysis showed that fallers were more likely to have lower baseline scores in executive function than non-fallers (p = 0.03). Among participants without balance impairment we found that higher executive function z-scores were associated with lower fall counts (p = 0.03) after adjustment for age, sex, health status and prior history of falls using negative binomial regression models. This relationship was not present among participants with poor balance.
Lower scores on executive function tests are a risk factor for falls in participants with minimal balance impairment. However, this effect is attenuated in individuals with poor balance where physical or more direct motor systems factors may play a greater role in fall risk.
In-home monitoring of gait velocity with passive PIR sensors in a smart home has been shown to be an effective method of continuously and unobtrusively measuring this important predictor of cognitive function and mobility. However, passive measurements of velocity are nonspecific with regard to who generated each measurement or walking event. As a result, this method is not suitable for multi-person homes without additional information to aid in the disambiguation of gait velocities. In this paper we propose a method based on Gaussian mixture models (GMMs) combined with infrequent clinical assessments of gait velocity to model in-home walking speeds of two or more residents. Modeling the gait parameters directly allows us to avoid the more difficult problem of assigning each measured velocity individually to the correct resident. We show that if the clinically measured gait velocities of residents are separated by at least 15 cm/s a GMM can be accurately fit to the in-home gait velocity data. We demonstrate the accuracy of this method by showing that the correlation between the means of the GMMs and the clinically measured gait velocities is 0.877 (p value < 0.0001) with bootstrapped 95% confidence intervals of (0.79, 0.94) for 54 measurements of 20 subjects living in multi-person homes. Example applications of using this method to track in-home mean velocities over time are also given.
Gait; passive infrared (PIR) motion detectors; smart homes; unobtrusive monitoring; walking speed
Gait velocity has been shown to quantitatively estimate risk of future hospitalization, has been shown to be a predictor of disability, and has been shown to slow prior to cognitive decline. In this paper, we describe a system for continuous and unobtrusive in-home assessment of gait velocity, a critical metric of function. This system is based on estimating walking speed from noisy time and location data collected by a “sensor line” of restricted view passive infrared (PIR) motion detectors. We demonstrate the validity of our system by comparing with measurements from the commercially available GAITRite® Walkway System gait mat. We present the data from 882 walks from 27 subjects walking at three different subject-paced speeds (encouraged to walk slowly, normal speed, or fast) in two directions through a sensor line. The experimental results show that the uncalibrated system accuracy (average error) of estimated velocity was 7.1cm/s (SD = 11.3cm/s), which improved to 1.1cm/s (SD = 9.1cm/s) after a simple calibration procedure. Based on the average measured walking speed of 102 cm/s our system had an average error of less than 7% without calibration and 1.1% with calibration.
Eldercare; unobtrusive monitoring; ubiquitous computing; gait; walking speed; passive infrared (PIR) motion detectors
Disrupted sleep patterns are a significant problem in the elderly, leading to increased cognitive dysfunction and risk of nursing home placement. A cost-effective and unobtrusive way to remotely monitor changing sleep patterns over time would enable improved management of this important health problem. We have developed an algorithm to derive sleep parameters such as bed time, rise time, sleep latency, and nap time from passive infrared sensors distributed around the home. We evaluated this algorithm using 404 days of data collected in the homes of 8 elderly community-dwelling elders. Data from this algorithm were highly correlated to ground truth measures (bed mats) and were surprisingly robust to variability in sensor layout and sleep habits.
In this paper we describe a preliminary modeling and analysis of a unique data set comprising unobtrusive and continuous measurements of gait velocity in the elder participants' residences. The data have been collected as a part of a longitudinal study aimed at early detection of cognitive decline. We motivate these analyses by first presenting evidence that suggests significant relationship between gait parameters and cognitive functions. We then describe a simple, model-based approach to the analysis of gait velocity using a weighted correlation function estimates. One of the main challenges is due to the fact that the daily estimates of the gait parameters vary with the number of walks. We illustrate the importance of using weighted as opposed to unweighted estimates on a sample of different houses. The correlation functions appear to capture behavioral differences that can be related to the cognitive functioning of the participants.
Poor medication adherence is one of the major causes of illness and of treatment failure in the United States. The objective of this study was to conduct an initial evaluation of a context-aware reminder system, which generated reminders at an opportune time to take the medication. Ten participants aged 65 or older, living alone and managing their own medications, participated in the study. Participants took a low-dose vitamin C tablet twice daily at times that they specified. Participants were considered adherent if they took the vitamin within 90 minutes (before or after) of the prescribed time. Adherence and activity in the home was measured using a system of sensors, including an instrumented pillbox. There were three phases of the study: baseline, in which there was no prompting; time-based, in which there was prompting at the prescribed times for pill-taking; and context-aware, in which participants were only prompted if they forgot to take their pills and were likely able to take their pills. The context-based prompting resulted in significantly better adherence (92.3%) as compared to time-based (73.5%) or no prompting (68.1%) conditions (p < 0.0002, χ2 = 17.0). In addition, subjects had better adherence in the morning than in the evening. We have shown in this study that a system that generates reminders at an opportune time to take the medication significantly improves adherence. This study indicates that context-aware prompting may provide improved adherence over standard time-based reminders.
home health telemonitoring; telehealth; information management
Walking speed and activity are important measures of functional ability in the elderly. Our earlier studies have suggested that continuous monitoring may allow us to detect changes in walking speed that are also predictive of cognitive changes. We evaluated the use of passive infrared (PIR) sensors for measuring walking speed in the home on an ongoing basis. In comparisons with gait mat estimates (ground truth) and the results of a timed walk test (the clinical gold standard) in 18 subjects, we found that the clinical measure overestimated typical walking speed, and the PIR sensor estimations of walking speed were highly correlated to actual gait speed. Examination of in-home walking patterns from more than 100,000 walking speed samples for these subjects suggested that we can accurately assess walking speed in the home. We discuss the potential of this approach for continuous assessment.
Unobtrusive in-home computer monitoring could one day be used to deliver cost-effective diagnostic information about the cognitive abilities of the elderly. This could allow for early detection of cognitive impairment and would additionally be coupled with the cost advantages that are associated with a semi-automated system. Before using the computer usage data to draw conclusions about the participants, we first needed to investigate the nature of the data that was collected. This paper represents a forensics style analysis of the computer usage data that is being collected as part of a larger study of cognitive decline, and focuses on the isolation and removal of non user-generated activities that were recorded by our computer monitoring software (CMS).
Assessment of cognitive functionality is an important aspect of care for elders. Unfortunately, few tools exist to measure divided attention, the ability to allocate attention to different aspects of tasks. An accurate determination of divided attention would allow inference of generalized cognitive decline, as well as providing a quantifiable indicator of an important component of driving skill. We propose a new method for determining relative divided attention ability through unobtrusive monitoring of computer use. Specifically, we measure performance on a dual-task cognitive computer exercise as part of a health coaching intervention. This metric indicates whether the user has the ability to pay attention to both tasks at once, or is primarily attending to one task at a time (sacrificing optimal performance). The monitoring of divided attention in a home environment is a key component of both the early detection of cognitive problems and for assessing the efficacy of coaching interventions.
We present an ambulatory cognitive state classification system to assess the subject's mental load based on EEG measurements. The ambulatory cognitive state estimator is utilized in the context of a real-time augmented cognition (AugCog) system that aims to enhance the cognitive performance of a human user through computer-mediated assistance based on assessments of cognitive states using physiological signals including, but not limited to, EEG. This paper focuses particularly on the offline channel selection and feature projection phases of the design and aims to present mutual-information-based techniques that use a simple sample estimator for this quantity. Analyses conducted on data collected from 3 subjects performing 2 tasks (n-back/Larson) at 2 difficulty levels (low/high) demonstrate that the proposed mutual-information-based dimensionality reduction scheme can achieve up to 94% cognitive load estimation accuracy.