Efficient Deep Learning-based Wound-bed Segmentation For Mobile Applications.

This paper proposes a deep learning image segmentation method for the purpose of segmenting wound-bed regions from the background. Our contributions include proposing a fast and efficient convolutional neural networks (CNN)-based segmentation network that has much smaller number of parameters than U-Net (only 18.1% that of U-Net, and hence the trained model has much smaller file size as well). In addition, the training time of our proposed segmentation network (for the base model) is only about 40.2% of that needed to train a U-Net. Furthermore, our proposed base model also achieved better performance compared to that of the U-Net in terms of both pixel accuracy and intersection-over-union segmentation evaluation metrics. We also showed that because of the small footprint of our efficient CNN-based segmentation model, it could be deployed to run in real-time on portable and mobile devices such as an iPad.

Brain-Computer Interface-Based Soft Robotic Glove Rehabilitation for Stroke.

OBJECTIVE: This randomized controlled feasibility study investigates the ability for clinical application of the Brain-Computer Interface-based Soft Robotic Glove (BCI-SRG) incorporating activities of daily living (ADL)-oriented tasks for stroke rehabilitation. METHODS: Eleven recruited chronic stroke patients were randomized into BCI-SRG or Soft Robotic Glove (SRG) group. Each group underwent 120-minute intervention per session comprising 30-minute standard arm therapy and 90-minute experimental therapy (BCI-SRG or SRG). To perform ADL tasks, BCI-SRG group used motor imagery-BCI and SRG, while SRG group used SRG without motor imagery-BCI. Both groups received 18 sessions of intervention over 6 weeks. Fugl-Meyer Motor Assessment (FMA) and Action Research Arm Test (ARAT) scores were measured at baseline (week 0), post- intervention (week 6), and follow-ups (week 12 and 24). In total, 10/11 patients completed the study with 5 in each group and 1 dropped out. RESULTS: Though there were no significant intergroup differences for FMA and ARAT during 6-week intervention, the improvement of FMA and ARAT seemed to sustain beyond 6-week intervention for BCI-SRG group, as compared with SRG control. Incidentally, all BCI-SRG subjects reported a sense of vivid movement of the stroke-impaired upper limb and 3/5 had this phenomenon persisting beyond intervention while none of SRG did. CONCLUSION: BCI-SRG suggested probable trends of sustained functional improvements with peculiar kinesthetic experience outlasting active intervention in chronic stroke despite the dire need for large-scale investigations to verify statistical significance. SIGNIFICANCE: Addition of BCI to soft robotic training for ADL-oriented stroke rehabilitation holds promise for sustained improvements as well as elicited perception of motor movements.

Brain-computer interface-based robotic end effector system for wrist and hand rehabilitation: results of a three-armed randomized controlled trial for chronic stroke.

The objective of this study was to investigate the efficacy of an Electroencephalography (EEG)-based Motor Imagery (MI) Brain-Computer Interface (BCI) coupled with a Haptic Knob (HK) robot for arm rehabilitation in stroke patients. In this three-arm, single-blind, randomized controlled trial; 21 chronic hemiplegic stroke patients (Fugl-Meyer Motor Assessment (FMMA) score 10-50), recruited after pre-screening for MI BCI ability, were randomly allocated to BCI-HK, HK or Standard Arm Therapy (SAT) groups. All groups received 18 sessions of intervention over 6 weeks, 3 sessions per week, 90 min per session. The BCI-HK group received 1 h of BCI coupled with HK intervention, and the HK group received 1 h of HK intervention per session. Both BCI-HK and HK groups received 120 trials of robot-assisted hand grasping and knob manipulation followed by 30 min of therapist-assisted arm mobilization. The SAT group received 1.5 h of therapist-assisted arm mobilization and forearm pronation-supination movements incorporating wrist control and grasp-release functions. In all, 14 males, 7 females, mean age 54.2 years, mean stroke duration 385.1 days, with baseline FMMA score 27.0 were recruited. The primary outcome measure was upper extremity FMMA scores measured mid-intervention at week 3, end-intervention at week 6, and follow-up at weeks 12 and 24. Seven, 8 and 7 subjects underwent BCI-HK, HK and SAT interventions respectively. FMMA score improved in all groups, but no intergroup differences were found at any time points. Significantly larger motor gains were observed in the BCI-HK group compared to the SAT group at weeks 3, 12, and 24, but motor gains in the HK group did not differ from the SAT group at any time point. In conclusion, BCI-HK is effective, safe, and may have the potential for enhancing motor recovery in chronic stroke when combined with therapist-assisted arm mobilization.

A clinical study of motor imagery BCI performance in stroke by including calibration data from passive movement.

Electroencephalogram (EEG) data from performing motor imagery are usually used to calibrate a subject-specific model in Motor Imagery Brain-Computer Interface (MI-BCI). However, the performance of MI is not directly observable by another person. Studies that attempted to address this issue in order to improve subjects with low MI performance had shown that it is feasible to use calibration data from Passive Movement (PM) to detect MI in healthy subjects. This study investigates the feasibility of using calibration data from PM of stroke patients to detect MI. EEG data from 2 calibration runs of MI and PM by a robotic haptic knob, and 1 evaluation run of MI were collected in one session of recording from 34 hemiparetic stroke patients recruited in the clinical study. In each run, 40 trials of MI or PM and 40 trials of the background rest were collected. The off-line run-to-run transfer kappa values from the calibration runs of MI, PM, and combined MI and PM, to the evaluation run of MI were then evaluated and compared. The results showed that calibration using PM (0.392) yielded significantly lower kappa value than the calibration using MI (0.457, p=4.40e-14). The results may be due to a significant disparity between the EEG data from PM and MI in stroke subjects. Nevertheless, the results showed that the calibration using both MI and PM (0.506) yielded significantly higher kappa value than the calibration using MI (0.457, p=9.54e-14). Hence, the results of this study suggest a promising direction to combine calibration data from PM and MI to improve MI detection on stroke.