Convolutional neural network in upper limb functional motion analysis after stroke.
Ontology highlight
ABSTRACT:
SUBMITTER: Szczesna A
PROVIDER: S-EPMC7549467 | biostudies-literature | 2020
REPOSITORIES: biostudies-literature
Publications
Convolutional neural network in upper limb functional motion analysis after stroke.
Szczęsna Agnieszka A Błaszczyszyn Monika M Kawala-Sterniuk Aleksandra A
PeerJ 20201009
In this work, implementation of Convolutional Neural Network (CNN) for the purpose of analysis of functional upper limb movement pattern was applied. The main aim of the study was to compare motion of selected activities of daily living of participants after stroke with the healthy ones (in similar age). The optical, marker-based motion capture system was applied for the purpose of data acquisition. There were some attempts made in order to find the existing differences in the motion pattern of ...[more]
Similar Datasets
Project description:PurposeWe introduce and validate a scalable retrospective motion correction technique for brain imaging that incorporates a machine learning component into a model-based motion minimization.MethodsA convolutional neural network (CNN) trained to remove motion artifacts from 2D T2 -weighted rapid acquisition with refocused echoes (RARE) images is introduced into a model-based data-consistency optimization to jointly search for 2D motion parameters and the uncorrupted image. Our separable motion model allows for efficient intrashot (line-by-line) motion correction of highly corrupted shots, as opposed to previous methods which do not scale well with this refinement of the motion model. Final image generation incorporates the motion parameters within a model-based image reconstruction. The method is tested in simulations and in vivo motion experiments of in-plane motion corruption.ResultsWhile the convolutional neural network alone provides some motion mitigation (at the expense of introduced blurring), allowing it to guide the iterative joint-optimization both improves the search convergence and renders the joint-optimization separable. This enables rapid mitigation within shots in addition to between shots. For 2D in-plane motion correction experiments, the result is a significant reduction of both image space root mean square error in simulations, and a reduction of motion artifacts in the in vivo motion tests.ConclusionThe separability and convergence improvements afforded by the combined convolutional neural network+model-based method shows the potential for meaningful postacquisition motion mitigation in clinical MRI.
| S-EPMC6626557 | biostudies-literature
Project description:Interventions: None
Primary outcome(s): The primary outcome of the study is the accuracy of the CAD-CNN system for predicting histology of diminutive colorectal polyps (1-5mm) compared with the accuracy of the prediction of the endoscopist. Both the CAD-CNN system and the endoscopist will use NBI for their predictions.
Accuracy is defined as the percentage of correctly predicted optical diagnoses of the CAD-CNN system and/or endoscopist compared to the gold standard pathology. For the calculation of the accuracy, adenomas and SSLs will be dichotomised as neoplastic polyps, while HPs and other non-neoplastic histology are considered non-neoplastic.
Study Design: N/A: single arm study, Open (masking not used), N/A , unknown, Other
| 2443187 | ecrin-mdr-crc
Project description:IntroductionThe primary aim of this work was to summarize and compare the effects of active rehabilitation assisted by new technologies (virtual reality [VR], robot-assisted therapy [RAT] and telerehabilitation [TR)) on upper limb motor function and everyday living activity during the subacute and chronic phases of stroke. The secondary aims were to compare the effects of these technologies according to the intervention design (in addition to or in substitution of conventional therapy), the duration of active rehabilitation and the severity of patients' motor impairments.Evidence acquisitionSeveral databases, namely PubMed, Scopus, Embase and Cochrane Library, were searched. Studies were included if they were meta-analyses with a moderate to high level of confidence (assessed with AMSTAR-2) that compared the effects of a new technology promoting active rehabilitation to that of a conventional therapy program among patients with stroke. Network meta-analyses were conducted to compare the effects of the new technologies.Evidence synthesisEighteen different meta-analyses were selected and fifteen included in the quantitative analysis. In total these 15 meta-analyses were based on 189 different randomized controlled trials. VR (SMD≥0.25; P<0.05), RAT (SMD≥0.29; P≤0.29) and TR (SMD≥-0.08; P≤0.64) were found to be at least as effective as conventional therapy. During the subacute phase, RAT's greatest effect was observed for patients with severe-moderate impairments whereas VR and TR's greatest effects were observed for patients with mild impairments. During the chronic phase, the highest effects were observed for patients with mild impairments, for all studies technologies. Network meta-analyses showed that VR and RAT were both significantly superior to TR in improving motor function during the chronic phase but revealed no significant difference between VR, RAT and TR effectiveness on both motor function (during the subacute phase) and activity (during both chronic and subacute phase).ConclusionsThis overview provides low-to-moderate evidence that rehabilitation assisted with technologies are at least as effective as conventional therapy for patients with stroke. While VR and RAT seem to be more efficient during the subacute phase, all technologies seem to be as efficient as one another in the chronic phase.
| S-EPMC9980549 | biostudies-literature
Project description:Dynamic functional connectivity (dFC) networks derived from resting-state functional magnetic resonance imaging (rs-fMRI) help us understand fundamental dynamic characteristics of human brains, thereby providing an efficient solution for automated identification of brain diseases, such as Alzheimer's disease (AD) and its prodromal stage. Existing studies have applied deep learning methods to dFC network analysis and achieved good performance compared with traditional machine learning methods. However, they seldom take advantage of sequential information conveyed in dFC networks that could be informative to improve the diagnosis performance. In this paper, we propose a convolutional recurrent neural network (CRNN) for automated brain disease classification with rs-fMRI data. Specifically, we first construct dFC networks from rs-fMRI data using a sliding window strategy. Then, we employ three convolutional layers and long short-term memory (LSTM) layer to extract high-level features of dFC networks and also preserve the sequential information of extracted features, followed by three fully connected layers for brain disease classification. Experimental results on 174 subjects with 563 rs-fMRI scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI) demonstrate the effectiveness of our proposed method in binary and multi-category classification tasks.
| S-EPMC9298744 | biostudies-literature
Project description:Resting-state functional magnetic resonance imaging (rs-fMRI) based on the blood-oxygen-level-dependent (BOLD) signal has been widely used in healthy individuals and patients to investigate brain functions when the subjects are in a resting or task-negative state. Head motion considerably confounds the interpretation of rs-fMRI data. Nuisance regression is commonly used to reduce motion-related artifacts with six motion parameters estimated from rigid-body realignment as regressors. To further compensate for the effect of head movement, the first-order temporal derivatives of motion parameters and squared motion parameters were proposed previously as possible motion regressors. However, these additional regressors may not be sufficient to model the impact of head motion because of the complexity of motion artifacts. In addition, while using more motion-related regressors could explain more variance in the data, the neural signal may also be removed with increasing number of motion regressors. To better model how in-scanner motion affects rs-fMRI data, a robust and automated convolutional neural network (CNN) model is developed in this study to obtain optimal motion regressors. The CNN network consists of two temporal convolutional layers and the output from the network are the derived motion regressors used in the following nuisance regression. The temporal convolutional layer in the network can non-parametrically model the prolonged effect of head motion. The set of regressors derived from the neural network is compared with the same number of regressors used in a traditional nuisance regression approach. It is demonstrated that the CNN-derived regressors can more effectively reduce motion-related artifacts.
| S-EPMC6482337 | biostudies-literature
Project description:Motor recovery is still limited for people with stroke especially those with greater functional impairments. In order to improve outcome, we need to understand more about the mechanisms underpinning recovery. Task-unbiased, blood flow-independent post-stroke neural activity can be acquired from resting brain electrophysiological recordings and offers substantial promise to investigate physiological mechanisms, but behaviourally relevant features of resting-state sensorimotor network dynamics have not yet been identified. Thirty-seven people with subcortical ischaemic stroke and unilateral hand paresis of any degree were longitudinally evaluated at 3 weeks (early subacute) and 12 weeks (late subacute) after stroke. Resting-state magnetoencephalography and clinical scores of motor function were recorded and compared with matched controls. Magnetoencephalography data were decomposed using a data-driven hidden Markov model into 10 time-varying resting-state networks. People with stroke showed statistically significantly improved Action Research Arm Test and Fugl-Meyer upper extremity scores between 3 weeks and 12 weeks after stroke (both P < 0.001). Hidden Markov model analysis revealed a primarily alpha-band ipsilesional resting-state sensorimotor network which had a significantly increased life-time (the average time elapsed between entering and exiting the network) and fractional occupancy (the occupied percentage among all networks) at 3 weeks after stroke when compared with controls. The life-time of the ipsilesional resting-state sensorimotor network positively correlated with concurrent motor scores in people with stroke who had not fully recovered. Specifically, this relationship was observed only in ipsilesional rather in contralesional sensorimotor network, default mode network or visual network. The ipsilesional sensorimotor network metrics were not significantly different from controls at 12 weeks after stroke. The increased recruitment of alpha-band ipsilesional resting-state sensorimotor network at subacute stroke served as functionally correlated biomarkers exclusively in people with stroke with not fully recovered hand paresis, plausibly reflecting functional motor recovery processes.
| S-EPMC10853981 | biostudies-literature
Project description:BackgroundFollowing a stroke, brain activation reorganisation, movement compensatory strategies, motor performance and their evolution through rehabilitation are matters of importance for clinicians. Two non-invasive neuroimaging methods allow for recording task-related brain activation: functional near-infrared spectroscopy (fNIRS) and electroencephalography (fEEG), respectively based on hemodynamic response and neuronal electrical activity. Their simultaneous measurement during movements could allow a better spatiotemporal mapping of brain activation, and when associated to kinematic parameters could unveil underlying mechanisms of functional upper limb (UL) recovery. This study aims to depict the motor cortical activity patterns using combined fNIRS-fEEG and their relationship to motor performance and strategies during UL functional tasks in chronic post-stroke patients.MethodsTwenty-one healthy old adults and 21 chronic post-stroke patients were recruited and completed two standardised functional tasks of the UL: a paced-reaching task where they had to reach a target in front of them and a circular steering task where they had to displace a target using a hand-held stylus, as fast as possible inside a circular track projected on a computer screen. The activity of the bilateral motor cortices and motor performance were recorded simultaneously utilizing a fNIRS-fEEG and kinematics platform.Results and conclusionsKinematic analysis revealed that post-stroke patients performed worse in the circular steering task and used more trunk compensation in both tasks. Brain analysis of bilateral motor cortices revealed that stroke individuals over-activated during the paretic UL reaching task, which was associated with more trunk usage and a higher level of impairment (clinical scores). This work opens up avenues for using such combined methods to better track and understand brain-movement evolution through stroke rehabilitation.
| S-EPMC11494975 | biostudies-literature
Project description:Introduction: Tremor of the upper limbs is a disabling symptom that is present during several neurological disorders and is currently without treatment. Functional MRI (fMRI) is an essential tool to investigate the pathophysiology of tremor and aid the development of treatment options. However, no adequately or standardized protocols for fMRI exists at present. Here we present a novel, online available fMRI task that could be used to assess the in vivo pathology of tremor. Objective: This study aims to validate the tremor-evoking potential of the fMRI task in a small group of tremor patients outside the scanner and assess the reproducibility of the fMRI task related activation in healthy controls. Methods: Twelve HCs were scanned at two time points (baseline and after 6-weeks). There were two runs of multi-band fMRI and the tasks included a "brick-breaker" joystick game. The game consisted of three conditions designed to control for most of the activation related to performing the task by contrasting the conditions: WATCH (look at the game without moving joystick), MOVE (rhythmic left/right movement of joystick without game), and PLAY (playing the game). Task fMRI was analyzed using FSL FEAT to determine clusters of activation during the different conditions. Maximum activation within the clusters was used to assess the ability to control for task related activation and reproducibility. Four tremor patients have been included to test ecological and construct validity of the joystick task by assessing tremor frequencies captured by the joystick. Results: In HCs the game activated areas corresponding to motor, attention and visual areas. Most areas of activation by our game showed moderate to good reproducibility (intraclass correlation coefficient (ICC) 0.531-0.906) with only inferior parietal lobe activation showing poor reproducibility (ICC 0.446). Furthermore, the joystick captured significantly more tremulous movement in tremor patients compared to HCs (p = 0.01) during PLAY, but not during MOVE. Conclusion: Validation of our novel task confirmed tremor-evoking potential and reproducibility analyses yielded acceptable results to continue further investigations into the pathophysiology of tremor. The use of this technique in studies with tremor patient will no doubt provide significant insights into the treatment options.
| S-EPMC6036145 | biostudies-literature
Project description:As stroke mortality rates decrease, there has been a surge of effort to study poststroke dementia (PSD) to improve long-term quality of life for stroke survivors. Hippocampal volume may be an important neuroimaging biomarker in poststroke dementia, as it has been associated with many other forms of dementia. However, studying hippocampal volume using MRI requires hippocampal segmentation. Advances in automated segmentation methods have allowed for studying the hippocampus on a large scale, which is important for robust results in the heterogeneous stroke population. However, most of these automated methods use a single atlas-based approach and may fail in the presence of severe structural abnormalities common in stroke. Hippodeep, a new convolutional neural network-based hippocampal segmentation method, does not rely solely on a single atlas-based approach and thus may be better suited for stroke populations. Here, we compared quality control and the accuracy of segmentations generated by Hippodeep and two well-accepted hippocampal segmentation methods on stroke MRIs (FreeSurfer 6.0 whole hippocampus and FreeSurfer 6.0 sum of hippocampal subfields). Quality control was performed using a stringent protocol for visual inspection of the segmentations, and accuracy was measured as volumetric correlation with manual segmentations. Hippodeep performed significantly better than both FreeSurfer methods in terms of quality control. All three automated segmentation methods had good correlation with manual segmentations and no one method was significantly more correlated than the others. Overall, this study suggests that both Hippodeep and FreeSurfer may be useful for hippocampal segmentation in stroke rehabilitation research, but Hippodeep may be more robust to stroke lesion anatomy.
| S-EPMC8675423 | biostudies-literature
Project description:BackgroundApproximately two thirds of stroke survivors maintain upper limb (UL) impairments and few among them attain complete UL recovery 6 months after stroke. Technological progress and gamification of interventions aim for better outcomes and constitute opportunities in self- and tele-rehabilitation.ObjectivesOur objective was to assess the efficacy of serious games, implemented on diverse technological systems, targeting UL recovery after stroke. In addition, we investigated whether adherence to neurorehabilitation principles influenced efficacy of games specifically designed for rehabilitation, regardless of the device used.MethodThis systematic review was conducted according to PRISMA guidelines (PROSPERO registration number: 156589). Two independent reviewers searched PubMed, EMBASE, SCOPUS and Cochrane Central Register of Controlled Trials for eligible randomized controlled trials (PEDro score ≥ 5). Meta-analysis, using a random effects model, was performed to compare effects of interventions using serious games, to conventional treatment, for UL rehabilitation in adult stroke patients. In addition, we conducted subgroup analysis, according to adherence of included studies to a consolidated set of 11 neurorehabilitation principles.ResultsMeta-analysis of 42 trials, including 1760 participants, showed better improvements in favor of interventions using serious games when compared to conventional therapies, regarding UL function (SMD = 0.47; 95% CI = 0.24 to 0.70; P < 0.0001), activity (SMD = 0.25; 95% CI = 0.05 to 0.46; P = 0.02) and participation (SMD = 0.66; 95% CI = 0.29 to 1.03; P = 0.0005). Additionally, long term effect retention was observed for UL function (SMD = 0.42; 95% CI = 0.05 to 0.79; P = 0.03). Interventions using serious games that complied with at least 8 neurorehabilitation principles showed better overall effects. Although heterogeneity levels remained moderate, results were little affected by changes in methods or outliers indicating robustness.ConclusionThis meta-analysis showed that rehabilitation through serious games, targeting UL recovery after stroke, leads to better improvements, compared to conventional treatment, in three ICF-WHO components. Irrespective of the technological device used, higher adherence to a consolidated set of neurorehabilitation principles enhances efficacy of serious games. Future development of stroke-specific rehabilitation interventions should further take into consideration the consolidated set of neurorehabilitation principles.
| S-EPMC8204490 | biostudies-literature