An official website of the United States government
BackgroundStroke therapy is essential to reduce impairments and improve motor movements by engaging autogenous neuroplasticity. Traditionally, stroke rehabilitation occurs in inpatient and outpatient rehabilitation facilities. However, recent literature increasingly explores moving the recovery process into the home and integrating technology-based interventions. This study advances this goal by promoting in-home, autonomous recovery for patients who experienced a stroke through robotics-assisted rehabilitation and classifying stroke residual severity using machine learning methods. ObjectiveOur main objective is to use kinematics data collected during in-home, self-guided therapy sessions to develop supervised machine learning methods, to address a clinician’s autonomous classification of stroke residual severity–labeled data toward improving in-home, robotics-assisted stroke rehabilitation. MethodsIn total, 33 patients who experienced a stroke participated in in-home therapy sessions using Motus Nova robotics rehabilitation technology to capture upper and lower body motion. During each therapy session, the Motus Hand and Motus Foot devices collected movement data, assistance data, and activity-specific data. We then synthesized, processed, and summarized these data. Next, the therapy session data were paired with clinician-informed, discrete stroke residual severity labels: “no range of motion (ROM),” “low ROM,” and “high ROM.” Afterward, an 80%:20% split was performed to divide the dataset into a training set and a holdout test set. We used 4 machine learning algorithms to classify stroke residual severity: light gradient boosting (LGB), extra trees classifier, deep feed-forward neural network, and classical logistic regression. We selected models based on 10-fold cross-validation and measured their performance on a holdout test dataset using F1-score to identify which model maximizes stroke residual severity classification accuracy. ResultsWe demonstrated that the LGB method provides the most reliable autonomous detection of stroke severity. The trained model is a consensus model that consists of 139 decision trees with up to 115 leaves each. This LGB model boasts a 96.70% F1-score compared to logistic regression (55.82%), extra trees classifier (94.81%), and deep feed-forward neural network (70.11%). ConclusionsWe showed how objectively measured rehabilitation training paired with machine learning methods can be used to identify the residual stroke severity class, with efforts to enhance in-home self-guided, individualized stroke rehabilitation. The model we trained relies only on session summary statistics, meaning it can potentially be integrated into similar settings for real-time classification, such as outpatient rehabilitation facilities.
more »
« less
Comparing the Reliability of Virtual and In-Person Post-Stroke Neuropsychological Assessment with Language Tasks
Abstract ObjectiveNeuropsychological testing is essential for both clinical and basic stroke research; however, the in-person nature of this testing is a limitation. Virtual testing overcomes the hurdles of geographic location, mobility issues and permits social distancing, yet its validity has received relatively little investigation, particularly in comparison with in-person testing. MethodWe expand on our prior findings of virtual testing feasibility by assessing virtual versus in-person administration of language and communication tasks with 48 left-hemisphere stroke patients (21 F, 27 M; mean age = 63.4 ± 12; mean years of education = 15.3 ± 3.5) in a quasi-test–retest paradigm. Each participant completed two testing sessions: one in their home and one in the research lab. Participants were assigned to one of the eight groups, with the testing condition (fully in-person, partially virtual), order of home session (first, second) and technology (iPad, Windows tablet) varied across groups. ResultsAcross six speech-language tasks that utilized varying response modalities and interfaces, we found no significant difference in performance between virtual and in-person testing. However, our results reveal key considerations for successful virtual administration of neuropsychological tests, including technology complications and disparities in internet access. ConclusionsVirtual administration of neuropsychological assessments demonstrates comparable reliability with in-person data collection involving stroke survivors, though technology issues must be taken into account.
more »
« less
- Award ID(s):
- 1923129
- PAR ID:
- 10386498
- Publisher / Repository:
- Oxford University Press
- Date Published:
- Journal Name:
- Archives of Clinical Neuropsychology
- Volume:
- 38
- Issue:
- 4
- ISSN:
- 1873-5843
- Page Range / eLocation ID:
- p. 557-569
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
IntroductionEarly informal learning experiences are essential for sparking long-term interest in science, technology, engineering, and math (STEM). In a prior study, we found more promising parent involvement outcomes when families of young children were provided with STEM family education events along with home STEM activity kits compared to providing workshops alone. This study was a conceptual replication using the same program—Teaching Together STEM—to deliver educational workshops plus home activity kits; however, we varied the delivery method by using virtual “funshops” to evaluate if parents perceived this modality as feasible and useful. MethodsMuseum informal science educators introduced four units via virtual video chat sessions linked to 12 hands-on STEM activities that were mailed to families randomly assigned to the treatment group. Half of the families were assigned to a waitlist control group that received a portion of the virtual program after the posttest. Participants included 60 families with children aged 3 to 5 years from diverse linguistic and socioeconomic backgrounds. ResultsOur results indicate no significant group differences in the primary outcome of parents’ involvement in informal STEM but a small, positive effect size (ES = 0.18) that was similar in magnitude to the prior, in-person study. Although parents mostly perceived the remote delivery as convenient and the materials as engaging for their child, there were no significant program impacts on children’s general science interests (ES = −0.19). DiscussionDespite the convenience, parents reported time was a barrier to doing STEM activities at home. Parents with lower education levels were less likely to attend, suggesting virtual approaches are not sufficient for ensuring broad access to family engagement programs for populations underrepresented in STEM.more » « less
-
Abstract Research internships provide students with invaluable experience conducting independent research, contributing to larger research programs, and embedding in a professional scientific setting. These experiences increase student persistence in ecology and other science, technology, engineering, and mathematics (STEM) fields and promote the inclusion of students who lack opportunities at their home institutions and/or are from groups that are underrepresented in STEM. While many ecology internship programs were canceled during the 2020 COVID‐19 pandemic, others successfully adapted to offer virtual internships for the first time. Though different from what many researchers and students envision when they think of internships, virtual ecology internship programs can create more accessible opportunities and be just as valuable as in‐person opportunities when research programs and advisors develop virtual internships with intention and planning. Here, we highlight six ways to structure a virtual intern project, spanning a spectrum from purely computer‐based opportunities (e.g., digital data gathering, data analysis, or synthesis) to fully hands‐on research (e.g., sample processing or home‐based experiments). We illustrate examples of these virtual projects through a case study of the Smithsonian Environmental Research Center's 2020 Virtual Internship Program. Next, we provide 10 recommendations for effectively developing a virtual internship program. Finally, we end with ways that virtual internships can avoid the limitations of in‐person internships, as well as possible solutions to perceived pitfalls of virtual internships. While virtual internships became a necessity in 2020 due to COVID‐19, the development and continuation of virtual internships in future can be a valuable tool to add to the suite of existing internship opportunities, possibly further promoting diversity, equity, and inclusion in ecology and STEM.more » « less
-
Abstract BackgroundPrevious work has shown that ~ 50–60% of individuals have impaired proprioception after stroke. Typically, these studies have identified proprioceptive impairments using a narrow range of reference movements. While this has been important for identifying the prevalence of proprioceptive impairments, it is unknown whether these error responses are consistent for a broad range of reference movements. The objective of this study was to characterize proprioceptive accuracy as function of movement speed and distance in stroke. MethodsStroke (N = 25) and controls (N = 21) completed a robotic proprioception test that varied movement speed and distance. Participants mirror-matched various reference movement speeds (0.1–0.4 m/s) and distances (7.5–17.5 cm). Spatial and temporal parameters known to quantify proprioception were used to determine group differences in proprioceptive accuracy, and whether patterns of proprioceptive error were consistent across testing conditions within and across groups. ResultsOverall, we found that stroke participants had impaired proprioception compared to controls. Proprioceptive errors related to tested reference movement scaled similarly to controls, but some errors showed amplified scaling (e.g., significantly overshooting or undershooting reference speed). Further, interaction effects were present for speed and distance reference combinations at the extremes of the testing distribution. ConclusionsWe found that stroke participants have impaired proprioception and that some proprioceptive errors were dependent on characteristics of the movement (e.g., speed) and that reference movements at the extremes of the testing distribution resulted in significantly larger proprioceptive errors for the stroke group. Understanding how sensory information is utilized across a broad spectrum of movements after stroke may aid design of rehabilitation programs.more » « less
-
Lee, YunJu (Ed.)BackgroundWhile many factors are associated with stepping activity after stroke, there is significant variability across studies. One potential reason to explain this variability is that there are certain characteristics that arenecessaryto achieve greater stepping activity that differ from others thatmayneed to be targeted to improve stepping activity. ObjectiveUsing two step thresholds (2500 steps/day, corresponding to home vs. community ambulation and 5500 steps/day, corresponding to achieving physical activity guidelines through walking), we applied 3 different algorithms to determine which predictors are most important to achieve these thresholds. MethodsWe analyzed data from 268 participants with stroke that included 25 demographic, performance-based and self-report variables. Step 1 of our analysis involved dimensionality reduction using lasso regularization. Step 2 applied drop column feature importance to compute the mean importance of each variable. We then assessed which predictors were important to all 3 mathematically unique algorithms. ResultsThe number of relevant predictors was reduced from 25 to 7 for home vs. community and from 25 to 16 for aerobic thresholds. Drop column feature importance revealed that 6 Minute Walk Test and speed modulation were the only variables found to be important to all 3 algorithms (primary characteristics)for each respective threshold. Other variables related to readiness to change activity behavior and physical health, among others, were found to be important to one or two algorithms (ancillary characteristics). ConclusionsAddressing physical capacity isnecessary but not sufficientto achieve important step thresholds, asancillary characteristics, such as readiness to change activity behavior and physical health may also need to be targeted. This delineation may explain heterogeneity across studies examining predictors of stepping activity in stroke.more » « less
