Feasibility of telemedicine research visits in people with Parkinson's disease residing in medically underserved areas

Pre-trained convolutional neural networks identify Parkinson's disease from spectrogram images of voice samples

Machine learning approaches including deep learning models have shown promising performance in the automatic detection of Parkinson's disease. These approaches rely on different types of data with voice recordings being the most used due to the convenient and non-invasive nature of data acquisition. Our group has successfully developed a novel approach that uses convolutional neural network with transfer learning to analyze spectrogram images of the sustained vowel /a/ to identify people with Parkinson's disease. We tested this approach by collecting a dataset of voice recordings via analog telephone lines, which support limited bandwidth. The convolutional neural network with transfer learning approach showed superior performance against conventional machine learning methods that collapse measurements across time to generate feature vectors. This study builds upon our prior results and presents two novel contributions: First, we tested the performance of our approach on a larger voice dataset recorded using smartphones with wide bandwidth. Our results show comparable performance between two datasets generated using different recording platforms despite the differences in most important features resulting from the limited bandwidth of analog telephonic lines. Second, we compared the classification performance achieved using linear-scale and mel-scale spectrogram images and showed a small but statistically significant gain using mel-scale spectrograms.

A systematic review of the effectiveness of digital cognitive assessments of cognitive impairment in Parkinson's disease

Background: Digitalization in healthcare has been extended to how we examine and manage Parkinson's Disease Mild Cognitive Impairment (PD-MCI). Methods: Moyer Population (those with PD and in some cases control groups), Intervention (digital cognitive test) and Outcome (validity and reliability) (PIO) and Campbell et al. Synthesis Without Meta-analysis (SWiM) methods were employed. A literature search of MEDLINE, PsycINFO, CINAHL, OpenGrey, and ProQuest Theses and Dissertations Sources screened for articles. Results: The digital trail-making test (dTMT) was the most used measure. There was strong validity between the dTMT and pencil-paper TMT, Mini-Mental State Examination (MMSE), and Montreal Cognitive Assessment (MoCA) scores (ranging from r = .55 to .90, p < .001). Validity between the TMT pencil-paper and digital versions were adequate (ranging from r = .51 to 90, p < .001). Reliability was demonstrated between PD and control groups' scores (ranging from r = .71 to .87). One study found excellent inter-rater reliability (ICC = .90 to .95). The dMoCA was the most used screen that assessed more than two cognitive domains. There was a range in the strength of agreement between digital and pencil-paper versions (ICC scores = .37 to .83) and only one study demonstrated adequate validity (r = .59, p < .001). Poor internal consistency (α = .54) and poor test re-test reliability (between PD and control groups' scores, p > .05) were found. Conclusion: This review found that digitalized cognitive tests are valid and reliable methods to assess PD-MCI. Considerations for future research are discussed.

Telemedicine in Alzheimer's disease and other dementias: Where we are?

The prevalence and global health burden of dementia including Alzheimer's disease (AD) is rising, while patients living in remote and underserved areas face significant challenges in reaching specialized care. Telemedicine offers a valuable solution in bridging this widening gap, by providing equal and timely access to tertiary-specialized centers. Accumulating evidence highlights that most parts of the remote neuropsychological and neurological evaluation are feasible, with patients, healthcare professionals and caregivers being generally satisfied with this means of care. Herein, we provide an updated overview of the available evidence on the use of telemedicine for patients with cognitive disorders, focusing on the different applications and settings, the remote, video-based neurological and neuropsychological assessment, current recommendations, non-pharmacological interventions, as well as legal and ethical considerations. Based on the literature review and our three-year experience in the "Specialized Outpatient Clinic of Memory, Dementia and Parkinson's disease through the National Telemedicine Network" in the Aiginition University Hospital of Athens, we propose a brief guide for assessing patients with cognitive impairment via telemedicine and suggest future research directions for the more effective and appropriate use of telemedicine in dementia assessment and care.

Pre-trained Convolutional Neural Networks Identify Parkinson's Disease from Spectrogram Images of Voice Samples

Machine learning approaches including deep learning models have shown promising performance in the automatic detection of Parkinson's disease. These approaches rely on different types of data with voice recordings being the most used due to the convenient and non-invasive nature of data acquisition. Our group has successfully developed a novel approach that uses convolutional neural network with transfer learning to analyze spectrogram images of the sustained vowel /a/ to identify people with Parkinson's disease. We tested this approach by collecting a dataset of voice recordings via telephone lines, which have limited bandwidth. This study builds upon our prior results in two major ways: First, we tested the performance of our approach on a larger voice dataset recorded using smartphones with wide bandwidth. Our results show comparable performance between two datasets generated using different recording platforms where we report differences in most important features resulting from the limited bandwidth of telephonic lines. Second, we compared the classification performance achieved using linear-scale and mel-scale spectrogram images and showed a small but statistically significant gain using mel-scale spectrograms. The convolutional neural network with transfer learning approach showed superior performance against conventional machine learning methods that collapse measurements across time to generate feature vectors.

Neurological Examination via Telemedicine: An Updated Review Focusing on Movement Disorders

Patients with movement disorders such as Parkinson's disease (PD) living in remote and underserved areas often have limited access to specialized healthcare, while the feasibility and reliability of the video-based examination remains unclear. The aim of this narrative review is to examine which parts of remote neurological assessment are feasible and reliable in movement disorders. Clinical studies have demonstrated that most parts of the video-based neurological examination are feasible, even in the absence of a third party, including stance and gait-if an assistive device is not required-bradykinesia, tremor, dystonia, some ocular mobility parts, coordination, and gross muscle power and sensation assessment. Technical issues (video quality, internet connection, camera placement) might affect bradykinesia and tremor evaluation, especially in mild cases, possibly due to their rhythmic nature. Rigidity, postural instability and deep tendon reflexes cannot be remotely performed unless a trained healthcare professional is present. A modified version of incomplete Unified Parkinson's Disease Rating Scale (UPDRS)-III and a related equation lacking rigidity and pull testing items can reliably predict total UPDRS-III. UPDRS-II, -IV, Timed "Up and Go", and non-motor and quality of life scales can be administered remotely, while the remote Movement Disorder Society (MDS)-UPDRS-III requires further investigation. In conclusion, most parts of neurological examination can be performed virtually in PD, except for rigidity and postural instability, while technical issues might affect the assessment of mild bradykinesia and tremor. The combined use of wearable devices may at least partially compensate for these challenges in the future.

Feasibility of regional center telehealth visits utilizing a rural research network in people with Parkinson's disease

Background

Impaired motor and cognitive function can make travel cumbersome for People with Parkinson's disease (PwPD). Over 50% of PwPD cared for at the University of Arkansas for Medical Sciences (UAMS) Movement Disorders Clinic reside over 30 miles from Little Rock. Improving access to clinical care for PwPD is needed.

Objective

To explore the feasibility of remote clinic-to-clinic telehealth research visits for evaluation of multi-modal function in PwPD.

Methods

PwPD residing within 30 miles of a UAMS Regional health center were enrolled and clinic-to-clinic telehealth visits were performed. Motor and non-motor disease assessments were administered and quantified. Results were compared to participants who performed at-home telehealth visits using the same protocols during the height of the COVID pandemic.

Results

Compared to the at-home telehealth visit group (n = 50), the participants from regional centers (n = 13) had similar age and disease duration, but greater disease severity with higher total Unified Parkinson's disease rating scale scores (Z = -2.218, p = 0.027) and lower Montreal Cognitive Assessment scores (Z = -3.350, p < 0.001). Regional center participants had lower incomes (Pearson's chi = 21.3, p < 0.001), higher costs to attend visits (Pearson's chi = 16.1, p = 0.003), and lived in more socioeconomically disadvantaged neighborhoods (Z = -3.120, p = 0.002). Prior research participation was lower in the regional center group (Pearson's chi = 4.5, p = 0.034) but both groups indicated interest in future research participation.

Conclusions

Regional center research visits in PwPD in medically underserved areas are feasible and could help improve access to care and research participation in these traditionally underrepresented populations.

Improving naturalistic neuroscience with patient engagement strategies

Introduction

The clinical implementation of chronic electrophysiology-driven adaptive deep brain stimulation (DBS) algorithms in movement disorders requires reliable representation of motor and non-motor symptoms in electrophysiological biomarkers, throughout normal life (naturalistic). To achieve this, there is the need for high-resolution and -quality chronic objective and subjective symptom monitoring in parallel to biomarker recordings. To realize these recordings, an active participation and engagement of the investigated patients is necessary. To date, there has been little research into patient engagement strategies for DBS patients or chronic electrophysiological recordings.

Concepts and results

We here present our concept and the first results of a patient engagement strategy for a chronic DBS study. After discussing the current state of literature, we present objectives, methodology and consequences of the patient engagement regarding study design, data acquisition, and study infrastructure. Nine patients with Parkinson's disease and their caregivers participated in the meeting, and their input led to changes to our study design. Especially, the patient input helped us designing study-set-up meetings and support structures.

Conclusion

We believe that patient engagement increases compliance and study motivation through scientific empowerment of patients. While considering patient opinion on sensors or questionnaire questions may lead to more precise and reliable data acquisition, there was also a high demand for study support and engagement structures. Hence, we recommend the implementation of patient engagement in planning of chronic studies with complex designs, long recording durations or high demand for individual active study participation.

Self-Management Systems for Patients and Clinicians in Parkinson's Disease Care: A Scoping Review

Background

Digital self-management tools including mobile apps and wearables can enhance personalized care in Parkinson's disease, and incorporating patient and clinician feedback into their evaluation can empower users and nurture patient-clinician relationships, necessitating a review to assess the state of the art and refine their use.

Objective

This review aimed to summarize the state of the art of self-management systems used in Parkinson's disease management, detailing the application of self-management techniques and the integration of clinicians. It also aimed to provide a concise synthesis on the acceptance and usability of these systems from the clinicians' standpoint, reflecting both patient engagement and clinician experience.

Methods

The review was organized following the PRISMA extension for Scoping Reviews and PICOS frameworks. Studies were retrieved from PubMed, CINAHL, Scopus, ACM Digital Library, and IEEE Xplore. Data was collected using a predefined form and then analyzed descriptively.

Results

Of the 15,231 studies retrieved, 33 were included. Five technology types were identified, with systems combining technologies being the most evaluated. Common self-management strategies included educational material and symptom journals. Only 11 studies gathered data from clinicians or reported evidence of clinician integration; out of those, six studies point out the importance of raw data availability, data visualization, and integrated data summaries.

Conclusions

While self-management systems for Parkinson's disease are well-received by patients, the studies underscore the urgency for more research into their usability for clinicians and integration into daily medical workflows to enhance overall care quality.

Development and implementation of the frog-in-maze game to study upper limb movement in people with Parkinson's disease

Upper-limb bradykinesia occurs early in Parkinson's disease (PD) and bradykinesia is required for diagnosis. Our goal was to develop, implement and validate a game "walking" a frog through a maze using bimanual, alternating finger-tapping movements to provide a salient, objective, and remotely monitorable method of tracking disease progression and response to therapy in PD. Twenty-five people with PD and 16 people without PD participated. Responses on 5 different mazes were quantified and compared to spatiotemporal gait parameters and standard disease metrics in these participants. Intertap interval (ITI) on maze 2 & 3, which included turns, was strongly inversely related to stride-length and stride-velocity and directly related to motor UPDRS scores. Levodopa decreased ITI, except in maze 4. PD participants with freezing of gait had longer ITI on all mazes. The responses quantified on maze 2 & 3 were related to disease severity and gait stride-length, were levodopa responsive, and were worse in people with freezing of gait, suggesting that these mazes could be used to quantify motor dysfunction in PD. Programming our frog-in-maze game onto a remotely distributable platform could provide a tool to monitor disease progression and therapeutic response in people with PD, including during clinical trials.

A machine learning method to process voice samples for identification of Parkinson's disease

Machine learning approaches have been used for the automatic detection of Parkinson's disease with voice recordings being the most used data type due to the simple and non-invasive nature of acquiring such data. Although voice recordings captured via telephone or mobile devices allow much easier and wider access for data collection, current conflicting performance results limit their clinical applicability. This study has two novel contributions. First, we show the reliability of personal telephone-collected voice recordings of the sustained vowel /a/ in natural settings by collecting samples from 50 people with specialist-diagnosed Parkinson's disease and 50 healthy controls and applying machine learning classification with voice features related to phonation. Second, we utilize a novel application of a pre-trained convolutional neural network (Inception V3) with transfer learning to analyze the spectrograms of the sustained vowel from these samples. This approach considers speech intensity estimates across time and frequency scales rather than collapsing measurements across time. We show the superiority of our deep learning model for the task of classifying people with Parkinson's disease as distinct from healthy controls.

A First Methodological Development and Validation of ReTap: An Open-Source UPDRS Finger Tapping Assessment Tool Based on Accelerometer-Data

Bradykinesia is a cardinal hallmark of Parkinson's disease (PD). Improvement in bradykinesia is an important signature of effective treatment. Finger tapping is commonly used to index bradykinesia, albeit these approaches largely rely on subjective clinical evaluations. Moreover, recently developed automated bradykinesia scoring tools are proprietary and are not suitable for capturing intraday symptom fluctuation. We assessed finger tapping (i.e., Unified Parkinson's Disease Rating Scale (UPDRS) item 3.4) in 37 people with Parkinson's disease (PwP) during routine treatment follow ups and analyzed their 350 sessions of 10-s tapping using index finger accelerometry. Herein, we developed and validated ReTap, an open-source tool for the automated prediction of finger tapping scores. ReTap successfully detected tapping blocks in over 94% of cases and extracted clinically relevant kinematic features per tap. Importantly, based on the kinematic features, ReTap predicted expert-rated UPDRS scores significantly better than chance in a hold out validation sample (n = 102). Moreover, ReTap-predicted UPDRS scores correlated positively with expert ratings in over 70% of the individual subjects in the holdout dataset. ReTap has the potential to provide accessible and reliable finger tapping scores, either in the clinic or at home, and may contribute to open-source and detailed analyses of bradykinesia.

Rural Research Network to engage rural and minority community members in translational research

Background

To address the high prevalence of health disparities and lack of research opportunities among rural and minority communities, the University of Arkansas for Medical Sciences (UAMS) developed the Rural Research Network in January 2020.

Aim

The aim of this report is to describe our process and progress in developing a rural research network. The Rural Research Network provides a platform to expand research participation opportunities to rural Arkansans, many of whom are older adults, low-income individuals, and underrepresented minority populations.

Methods

The Rural Research Network leverages existing UAMS Regional Programs family medicine residency clinics within an academic medical center.

Results

Since the inception of the Rural Research Network, research infrastructure and processes have been built within the regional sites. Twelve diverse studies have been implemented with recruitment and data collection from 9248 participants, and 32 manuscripts have been published with residents and faculty from the regional sites. Most studies were able to recruit Black/African American participants at or above a representative sample.

Conclusions

As the Rural Research Network matures, the types of research will expand in parallel with the health priorities of Arkansas.

Relevance to Patients

The Rural Research Network demonstrates how Cancer Institutes and sites funded by a Clinical and Translational Science Award can collaborate to expand research capacity and increase opportunities for research among rural and minority communities.