Ensemble deep model for continuous estimation of Unified Parkinson's Disease Rating Scale III.

BACKGROUND: Unified Parkinson Disease Rating Scale-part III (UPDRS III) is part of the standard clinical examination performed to track the severity of Parkinson's disease (PD) motor complications. Wearable technologies could be used to reduce the need for on-site clinical examinations of people with Parkinson's disease (PwP) and provide a reliable and continuous estimation of the severity of PD at home. The reported estimation can be used to successfully adjust the dose and interval of PD medications. METHODS: We developed a novel algorithm for unobtrusive and continuous UPDRS-III estimation at home using two wearable inertial sensors mounted on the wrist and ankle. We used the ensemble of three deep-learning models to detect UPDRS-III-related patterns from a combination of hand-crafted features, raw temporal signals, and their time-frequency representation. Specifically, we used a dual-channel, Long Short-Term Memory (LSTM) for hand-crafted features, 1D Convolutional Neural Network (CNN)-LSTM for raw signals, and 2D CNN-LSTM for time-frequency data. We utilized transfer learning from activity recognition data and proposed a two-stage training for the CNN-LSTM networks to cope with the limited amount of data. RESULTS: The algorithm was evaluated on gyroscope data from 24 PwP as they performed different daily living activities. The estimated UPDRS-III scores had a correlation of [Formula: see text] and a mean absolute error of 5.95 with the clinical examination scores without requiring the patients to perform any specific tasks. CONCLUSION: Our analysis demonstrates the potential of our algorithm for estimating PD severity scores unobtrusively at home. Such an algorithm could provide the required motor-complication measurements without unnecessary clinical visits and help the treating physician provide effective management of the disease.

Wearable Sensors for Estimation of Parkinsonian Tremor Severity during Free Body Movements.

Tremor is one of the main symptoms of Parkinson's Disease (PD) that reduces the quality of life. Tremor is measured as part of the Unified Parkinson Disease Rating Scale (UPDRS) part III. However, the assessment is based on onsite physical examinations and does not fully represent the patients' tremor experience in their day-to-day life. Our objective in this paper was to develop algorithms that, combined with wearable sensors, can estimate total Parkinsonian tremor as the patients performed a variety of free body movements. We developed two methods: an ensemble model based on gradient tree boosting and a deep learning model based on long short-term memory (LSTM) networks. The developed methods were assessed on gyroscope sensor data from 24 PD subjects. Our analysis demonstrated that the method based on gradient tree boosting provided a high correlation (r = 0.96 using held-out testing and r = 0.93 using subject-based, leave-one-out cross-validation) between the estimated and clinically assessed tremor subscores in comparison to the LSTM-based method with a moderate correlation (r = 0.84 using held-out testing and r = 0.77 using subject-based, leave-one-out cross-validation). These results indicate that our approach holds great promise in providing a full spectrum of the patients' tremor from continuous monitoring of the subjects' movement in their natural environment.

Technology in Parkinson's disease: Challenges and opportunities.

The miniaturization, sophistication, proliferation, and accessibility of technologies are enabling the capture of more and previously inaccessible phenomena in Parkinson's disease (PD). However, more information has not translated into a greater understanding of disease complexity to satisfy diagnostic and therapeutic needs. Challenges include noncompatible technology platforms, the need for wide-scale and long-term deployment of sensor technology (among vulnerable elderly patients in particular), and the gap between the "big data" acquired with sensitive measurement technologies and their limited clinical application. Major opportunities could be realized if new technologies are developed as part of open-source and/or open-hardware platforms that enable multichannel data capture sensitive to the broad range of motor and nonmotor problems that characterize PD and are adaptable into self-adjusting, individualized treatment delivery systems. The International Parkinson and Movement Disorders Society Task Force on Technology is entrusted to convene engineers, clinicians, researchers, and patients to promote the development of integrated measurement and closed-loop therapeutic systems with high patient adherence that also serve to (1) encourage the adoption of clinico-pathophysiologic phenotyping and early detection of critical disease milestones, (2) enhance the tailoring of symptomatic therapy, (3) improve subgroup targeting of patients for future testing of disease-modifying treatments, and (4) identify objective biomarkers to improve the longitudinal tracking of impairments in clinical care and research. This article summarizes the work carried out by the task force toward identifying challenges and opportunities in the development of technologies with potential for improving the clinical management and the quality of life of individuals with PD. © 2016 International Parkinson and Movement Disorder Society.

Neurophysiological correlates of motor and working memory performance following subthalamic nucleus stimulation.

Subthalamic nucleus (STN) deep brain stimulation (DBS) has become an accepted treatment for the motor manifestations of Parkinson disease (PD). The beneficial motor effects of STN DBS are likely due to modulation of BG output to frontal cortical regions associated with motor control, but the underlying neurophysiology of STN DBS effects, especially at the level of the cortex, is not well understood. In this study, we examined the effects of STN DBS on motor disability and visual working memory, a cognitive process supported by pFC. We tested 10 PD participants off medications, ON and OFF stimulation, along with 20 normal controls on a visual working memory task while simultaneously recording cortical EEG. In the OFF state, PD patients had poor motor function, were slower and less accurate in performing the working memory task, and had greater amplitudes and shorter latencies of the N200 ERP response. DBS improved clinical motor function, reduced N200 amplitudes, and increased N200 latencies but had little effect on working memory performance. We conclude that STN DBS normalizes neurophysiological activity in fronto striatal circuits and this may independently affect motor and cognitive function.

Dyskinesia estimation during activities of daily living using wearable motion sensors and deep recurrent networks.

Levodopa-induced dyskinesias are abnormal involuntary movements experienced by the majority of persons with Parkinson's disease (PwP) at some point over the course of the disease. Choreiform as the most common phenomenology of levodopa-induced dyskinesias can be managed by adjusting the dose/frequency of PD medication(s) based on a PwP's motor fluctuations over a typical day. We developed a sensor-based assessment system to provide such information. We used movement data collected from the upper and lower extremities of 15 PwPs along with a deep recurrent model to estimate dyskinesia severity as they perform different activities of daily living (ADL). Subjects performed a variety of ADLs during a 4-h period while their dyskinesia severity was rated by the movement disorder experts. The estimated dyskinesia severity scores from our model correlated highly with the expert-rated scores (r = 0.87 (p < 0.001)), which was higher than the performance of linear regression that is commonly used for dyskinesia estimation (r = 0.81 (p < 0.001)). Our model provided consistent performance at different ADLs with minimum r = 0.70 (during walking) to maximum r = 0.84 (drinking) in comparison to linear regression with r = 0.00 (walking) to r = 0.76 (cutting food). These findings suggest that when our model is applied to at-home sensor data, it can provide an accurate picture of changes of dyskinesia severity facilitating effective medication adjustments.

Amyloid imaging of Lewy body-associated disorders.

Clinicopathologic studies of Parkinson disease dementia (PDD) and dementia with Lewy bodies (DLB) commonly reveal abnormal ß-amyloid deposition in addition to diffuse Lewy bodies (α-synuclein aggregates), but the relationship among these neuropathologic features and the development of dementia in these disorders remains uncertain. The purpose of this study was to determine whether amyloid-ß deposition detected by PET imaging with Pittsburgh Compound B (PIB) distinguishes clinical subtypes of Lewy body-associated disorders. Nine healthy controls, 8 PD with no cognitive impairment, 9 PD with mild cognitive impairment, 6 DLB, and 15 PDD patients underwent [(11)C]-PIB positron emission tomography imaging, clinical examination, and cognitive testing. The binding potential (BP) of PIB for predefined regions and the mean cortical BP (MCBP) were calculated for each participant. Annual longitudinal follow-up and postmortem examinations were performed on a subset of participants. Regional PIB BPs and the proportion of individuals with abnormally elevated MCBP were not significantly different across participant groups. Elevated PIB binding was associated with worse global cognitive impairment in participants with Lewy body disorders but was not associated with any other clinical or neuropsychological features, including earlier onset or faster rate of progression of cognitive impairment. These results suggest that the presence of fibrillar amyloid-ß does not distinguish between clinical subtypes of Lewy body-associated disorders, although larger numbers are needed to more definitively rule out this association. Amyloid-ß may modify the severity of global cognitive impairment in individuals with Lewy body-associated dementia.

Dyskinesia Severity Estimation in Patients with Parkinson's Disease Using Wearable Sensors and A Deep LSTM Network.

Dyskinesias are abnormal involuntary movements that patients with mid-stage and advanced Parkinson's disease (PD) may suffer from. These troublesome motor impairments are reduced by adjusting the dose or frequency of medication levodopa. However, to make a successful adjustment, the treating physician needs information about the severity rating of dyskinesia as patients experience in their natural living environment. In this work, we used movement data collected from the upper and lower extremities of PD patients along with a deep model based on Long Short-Term Memory to estimate the severity of dyskinesia. We trained and validated our model on a dataset of 14 PD subjects with dyskinesia. The subjects performed a variety of daily living activities while their dyskinesia severity was rated by a neurologist. The estimated dyskinesia severity ratings from our developed model highly correlated with the neurologist-rated dyskinesia scores (r=0.86 (p<0.001) and 1.77 MAE (6%)) indicating the potential of the developed the approach in providing the information required for effective medication adjustments for dyskinesia management.

Acute readmission following deep brain stimulation surgery for Parkinson's disease: A nationwide analysis.

INTRODUCTION: Deep brain stimulation (DBS) surgery is an efficacious, underutilized treatment for Parkinson's disease (PD). Studies of DBS post-operative outcomes are often restricted to data from a single center and consider DBS in isolation. National estimates of DBS readmission and post-operative outcomes are needed, as are comparisons to commonly performed surgeries. METHODS: This study used datasets from the 2013 and 2014 Nationwide Readmissions Database (NRD). Our sample was restricted to PD patients discharged alive after hospitalization for DBS surgery. Descriptive analyses examined patient, clinical, hospital and index hospitalization characteristics. The all-cause, non-elective 30-day readmission rate after DBS was calculated, and logistic regression models were built to examine factors associated with readmission. Readmission rates for the most common surgical procedures were calculated and compared to DBS. RESULTS: There were 6058 DBS surgeries for PD in our sample, most often involving a male aged 65 and older, who lived in a high socioeconomic status zip code. DBS patients had an average of four comorbidities. With respect to outcomes, the majority of patients were discharged home (95.3%). Non-elective readmission was rare (4.9%), and was associated with socioeconomic status, comorbidity burden, and teaching hospital status. Much higher acute, non-elective readmission rates were observed for common procedures such as upper gastrointestinal endoscopy (16.2%), colonoscopy (14.0%), and cardiac defibrillator and pacemaker procedures (11.1%). CONCLUSION: Short-term hospitalization outcomes after DBS are generally favorable. Socioeconomic disparities in DBS use persist. Additional efforts may be needed to improve provider referrals for and patient access to DBS.

Assessment of response to medication in individuals with Parkinson's disease.

BACKGROUND AND OBJECTIVE: Motor fluctuations between akinetic (medication OFF) and mobile phases (medication ON) states are one of the most prevalent complications of patients with Parkinson's disease (PD). There is a need for a technology-based system to provide reliable information about the duration in different medication phases that can be used by the treating physician to successfully adjust therapy. METHODS: Two KinetiSense motion sensors were mounted on the most affected wrist and ankle of 19 PD subjects (age: 42-77, 14 males) and collected movement signals as the participants performed seven daily living activities in their medication OFF and ON phases. A feature selection and a classification algorithm based on support vector machine with fuzzy labeling was developed to detect medication ON/OFF states using gyroscope signals. The algorithm was trained using approximately 15% of the data from four activities and tested on the remaining data. RESULTS: The algorithm was able to detect medication ON and OFF states with 90.5% accuracy, 94.2% sensitivity, and 85.4% specificity. It performed equally well for all the activities with an average accuracy of 91.3% for the activities that were used in the training phase and 88.4% for the new activities. CONCLUSIONS: The developed sensor-based algorithm could provide objective and accurate assessment of medication states that can lead to successful adjustment of the therapy resulting in considerably improved care delivery and quality of life of PD patients.

Continuous Assessment of Levodopa Response in Parkinson's Disease Using Wearable Motion Sensors.

OBJECTIVE: Fluctuations in response to levodopa in Parkinson's disease (PD) are difficult to treat as tools to monitor temporal patterns of symptoms are hampered by several challenges. The objective was to use wearable sensors to quantify the dose response of tremor, bradykinesia, and dyskinesia in individuals with PD. METHODS: Thirteen individuals with PD and fluctuating motor benefit were instrumented with wrist and ankle motion sensors and recorded by video. Kinematic data were recorded as subjects completed a series of activities in a simulated home environment through transition from off to on medication. Subjects were evaluated using the unified Parkinson disease rating scale motor exam (UPDRS-III) at the start and end of data collection. Algorithms were applied to the kinematic data to score tremor, bradykinesia, and dyskinesia. A blinded clinician rated severity observed on video. Accuracy of algorithms was evaluated by comparing scores with clinician ratings using a receiver operating characteristic (ROC) analysis. RESULTS: Algorithm scores for tremor, bradykinesia, and dyskinesia agreed with clinician ratings of video recordings (ROC area > 0.8). Summary metrics extracted from time intervals before and after taking medication provided quantitative measures of therapeutic response (p < 0.01). Radar charts provided intuitive visualization, with graphical features correlated with UPDRS-III scores (R = 0.81). CONCLUSION: A system with wrist and ankle motion sensors can provide accurate measures of tremor, bradykinesia, and dyskinesia as patients complete routine activities. SIGNIFICANCE: This technology could provide insight on motor fluctuations in the context of daily life to guide clinical management and aid in development of new therapies.

Implementing Levodopa-Carbidopa Intestinal Gel for Parkinson Disease: Insights from US Practitioners.

BACKGROUND: Levodopa-carbidopa intestinal gel (LCIG, designated in the United States as carbidopa-levodopa enteral suspension, CLES) was approved in the United States in 2015 for the treatment of refractory motor fluctuations in individuals with Parkinson disease (PD). Many neurologists in the United States have not had personal experience with implementation and management of the unique delivery system for this treatment. METHODS AND FINDINGS: This educational review was developed to provide practitioners with an understanding of LCIG use from the clinician's point of view. Practical recommendations for the use of LCIG from the early planning stages through long-term patient management were compiled from the published literature, regulatory guidance, and clinical experience. Among the topics reviewed were: assembling a multidisciplinary treatment team, identifying treatment candidates, patient/care partner education, procedural considerations, post-procedural care, LCIG initiation and titration, troubleshooting issues, and ongoing monitoring. For most of these steps, a considerable amount of individualization is possible, which allows clinicians to tailor protocols based on the needs of their teams, the healthcare system, and the patient and care partner. Although clinical practices are heterogeneous, themes of early planning, ongoing education, and a team-based approach to management are universal. CONCLUSIONS: By using established protocols and insights gleaned from experienced practitioners, clinicians who are unfamiliar with LCIG can more feasibly incorporate this treatment option into their armamentarium for treating PD motor fluctuations.

Telehealth Management of Parkinson's Disease Using Wearable Sensors: An Exploratory Study.

BACKGROUND: Parkinson's disease (PD) motor symptoms can fluctuate and may not be accurately reflected during a clinical evaluation. In addition, access to movement disorder specialists is limited for many with PD. The objective was to assess the impact of motion sensor-based telehealth diagnostics on PD clinical care and management. METHODS: Eighteen adults with PD were randomized to control or experimental groups. All participants were instructed to use a motion sensor-based monitoring system at home one day per week, for seven months. The system included a finger-worn motion sensor and tablet-based software interface that guided patients through tasks to quantify tremor, bradykinesia, and dyskinesia. Data were processed into motor symptom severity reports, which were reviewed by a movement disorders neurologist for experimental group participants. After three months and six months, control group participants visited the clinic for a routine appointment, while experimental group participants had a videoconference or phone call instead. RESULTS: Home based assessments were completed with median compliance of 95.7%. For a subset of participants, the neurologist successfully used information in the reports such as quantified response to treatment or progression over time to make therapy adjustments. Changes in clinical characteristics from study start to end were not significantly different between groups. DISCUSSION: Individuals with PD were able and willing to use remote monitoring technology. Patient management aided by telehealth diagnostics provided comparable outcomes to standard care. Telehealth technologies combined with wearable sensors have the potential to improve care for disparate PD populations or those unable to travel.

Utilization of rehabilitation therapy services in Parkinson disease in the United States.

OBJECTIVE: To examine rehabilitation therapy utilization for Parkinson disease (PD). METHODS: We identified 174,643 Medicare beneficiaries with a diagnosis of PD in 2007 and followed them through 2009. The main outcome measures were annual receipt of physical therapy (PT), occupational therapy (OT), or speech therapy (ST). RESULTS: Outpatient rehabilitation fee-for-service use was low. In 2007, only 14.2% of individuals with PD had claims for PT or OT, and 14.6% for ST. Asian Americans were the highest users of PT/OT (18.4%) and ST (18.4%), followed by Caucasians (PT/OT 14.4%, ST 14.8%). African Americans had the lowest utilization (PT/OT 7.8%, ST 8.2%). Using logistic regression models that accounted for repeated measures, we found that African American patients (adjusted odds ratio [AOR] 0.63 for PT/OT, AOR 0.63 for ST) and Hispanic patients (AOR 0.97 for PT/OT, AOR 0.91 for ST) were less likely to have received therapies compared to Caucasian patients. Patients with PD with at least one neurologist visit per year were 43% more likely to have a claim for PT evaluation as compared to patients without neurologist care (AOR 1.43, 1.30-1.48), and this relationship was similar for OT evaluation, PT/OT treatment, and ST. Geographically, Western states had the greatest use of rehabilitation therapies, but provider supply did not correlate with utilization. CONCLUSIONS: This claims-based analysis suggests that rehabilitation therapy utilization among older patients with PD in the United States is lower than reported for countries with comparable health care infrastructure. Neurologist care is associated with rehabilitation therapy use; provider supply is not.

Automatic assessment of medication states of patients with Parkinson's disease using wearable sensors.

Motor fluctuations are a major focus of clinical managements in patients with mid-stage and advance Parkinson's disease (PD). In this paper, we develop a new patient-specific algorithm that can classify those fluctuations during a variety of activities. We extract a set of temporal and spectral features from the ambulatory signals and then introduce a semi-supervised classification algorithm based on K-means and self-organizing tree map clustering methods. Two different types of cluster labeling are introduced: hard and fuzzy labeling. The developed algorithm is evaluated on a dataset from triaxial gyroscope sensors for 12 PD patients. The average result of using K-means and fuzzy labeling on the trunk and the more affected leg sensors' readings was 75.96%, 70.57%, and 86.93% for accuracy, sensitivity, and specificity, respectively. The accuracy for individual patients varied from 99.95% to 42.53%, which was correlated with dyskinesia severity and the improvement of the PD symptoms with medication.

Motion sensor dyskinesia assessment during activities of daily living.

BACKGROUND: Dyskinesia throughout the levodopa dose cycle has been previously measured in patients with Parkinson's disease (PD) using a wrist-worn motion sensor during the stationary tasks of arms resting and extended. Quantifying dyskinesia during unconstrained activities poses a unique challenge since these involuntary movements are kinematically similar to voluntary movement. OBJECTIVE: To determine the feasibility of using motion sensors to measure dyskinesia during activities of daily living. METHODS: Fifteen PD subjects performed scripted activities of daily living while wearing motion sensors on bilateral hands, thighs, and ankles over the course of a levodopa dose cycle. Videos were scored by clinicians using the modified Abnormal Involuntary Movement Scale to rate dyskinesia severity in separate body regions, with the total score used as an overall measure. Kinematic features were extracted from the motion data and algorithms were generated to output severity scores. RESULTS: Movements when subjects were experiencing dyskinesia were less smooth than when they were not experiencing dyskinesia. Dyskinesia scores predicted by the model using all sensors were highly correlated with clinician scores, with a correlation coefficient of 0.86 and normalized root-mean-square-error of 7.4%. Accurate predictions were maintained when two sensors on the most affected side of the body (one on the upper extremity and one on the lower extremity) were used. CONCLUSIONS: A system with motion sensors may provide an accurate measure of overall dyskinesia that can be used to monitor patients as they complete typical activities, and thus provide insight on symptom fluctuation in the context of daily life.

Objective motion sensor assessment highly correlated with scores of global levodopa-induced dyskinesia in Parkinson's disease.

BACKGROUND: Chronic use of medication for treating Parkinson's disease (PD) can give rise to peak-dose dyskinesia. Adjustments in medication often sacrifice control of motor symptoms, and thus balancing this trade-off poses a significant challenge for disease management. OBJECTIVE: To determine whether a wrist-worn motion sensor unit could be used to ascertain global dyskinesia severity over a levodopa dose cycle and to develop a severity scoring algorithm highly correlated with clinician ratings. METHODS: Fifteen individuals with PD were instrumented with a wrist-worn motion sensor unit, and data were collected with arms in resting and extended positions once every hour for three hours after taking a levodopa dose. Two neurologists blinded to treatment status viewed subject videos and rated global and upper extremity dyskinesia severity based on the modified Abnormal Involuntary Movement Scale (mAIMS). Linear regression models were developed using kinematic features extracted from motion sensor data and extremity, global, or combined (average of extremity and global) mAIMS scores. RESULTS: Dyskinesia occurring during a levodopa dose cycle was successfully measured using a wrist-worn sensor. The logarithm of the power spectrum area between 0.3-3 Hz and the combined clinician scores resulted in the best model performance, with a correlation coefficient between clinician and model scores of 0.81 and root mean square error of 0.55, both averaged across the arms resting and extended postures. CONCLUSIONS: One sensor unit worn on either hand can effectively predict global dyskinesia severity during the arms resting or extended positions.

Quantitative assessment of levodopa-induced dyskinesia using automated motion sensing technology.

The objective was to capture levodopa-induced dyskinesia (LID) in patients with Parkinson's disease (PD) using body-worn motion sensors. Dopaminergic treatment in PD can induce abnormal involuntary movements, including choreatic dyskinesia (brief, rapid, irregular movements). Adjustments in medication to reduce LID often sacrifice control of motor symptoms, and balancing this tradeoff poses a significant challenge for management of advanced PD. Fifteen PD subjects with known LID were recruited and instructed to perform two stationary motor tasks while wearing a compact wireless motion sensor unit positioned on each hand over the course of a levodopa dose cycle. Videos of subjects performing the motor tasks were later scored by expert clinicians to assess global dyskinesia using the modified Abnormal Involuntary Rating Scale (m-AIMS). Kinematic features were extracted from motion data in different frequency bands (1-3Hz and 3-8Hz) to quantify LID severity and to distinguish between LID and PD tremor. Receiver operator characteristic analysis was used to determine thresholds for individual features to detect the presence of LID. A sensitivity of 0.73 and specificity of 1.00 were achieved. A neural network was also trained to output dyskinesia severity on a 0 to 4 scale, similar to the m-AIMS. The model generalized well to new data (coefficient of determination= 0.85 and mean squared error= 0.3). This study demonstrated that hand-worn motion sensors can be used to assess global dyskinesia severity independent of PD tremor over the levodopa dose cycle.