Scope and Objectives

Scope: It is emerging today that with an increasing number of stroke-affected patients, the number of physiotherapists/neurologists to support the post-stroke rehabilitation is inadequate and therefore, an assistive brain-controlled hand exoskeleton could improve the pace of rehabilitation. By actively engaging stroke patients in repetitive tasks, the brain is able to rewire neurological pathways to motor functions to relearn movement. Providing a stroke patient a mechanism for repetitive hand movements with cerebral feedback could enhance recovery. Recent post-stroke rehabilitation clinical trials have suggested the need for continuous monitoring for compliance of prescription and provision of a proactive approach towards treatment that enhances recovery at various levels.

Objectives: 1) Design a brain-controlled Robotic Exoskeleton that actuates fingers to assist rehabilitation exercises as prescribed by a physician. 2) Develop methods based Electroencephalography (EEG) acquisition, decoding, and generation of action for different finger flexion and extension movement. 3) Design an EMG based bio-feedback control system which will provide information regarding muscle activity and also assess the stiffness level of the stroke patients. 4) Software platform development for assigning exercises to the patient and the assessment of progress during trials on stroke patients with a recording of fMRI data for pre and post exercise conditions.

Deliverables

It is intended to develop a lightweight and cost-effective exoskeleton which will be useful for hand movement rehabilitation. The proposed prototype would be triggered and controlled by biosignals including those emerging from the motor cortex. Further, it is proposed to deliver a robotic exoskeleton for enhancing neuroplasticity to improve motor power in the wrist and fingers by programmed repetitive training. These finger and wrist movements will be triggered by the Electroencephalogram EEG signal recorded from the subject brain. and/or a program prescribed by the doctor/physiotherapist that can be used with a portable handheld system.

Videos

Scientific Output

Three significant outcomes have been achieved so far: 1 A scheme has been evolved to identify Thumb, index and middle finger (of right hand) movements of 6 healthy patients from EEG data of Motor cortex with a 56 to 86 percent accuracy. This is indicative that one can expect to develop a BCI based technique to move the respective fingers through an assistive device like an exoskeleton. This work has been reported at an international conference. 2. An extensive study of 13 first time ischaemic stroke effected patients (10 male and 3 female Stroke Patients at Christian Medical College and Hospital Ludhiana.) was undertaken in which pre- and post-exercise baseline QEEG parameters such as mean Delta Theta Alpha and Beta frequency powers power ratio index delta vs alpha ratio and global brain symmetry index were studied. The progress was measured by comparing these parameters over multiple days 1, 7, 30 and 90. The study was expected to validate the EEG based evaluation of the progress under the prescribed exercise therapies. This has been reported in a conference. 3. A design process has been developed for various types of hand exoskeletons, namely linkage-driven and cable-driven mechanisms. These were studied for kinematic suitability with degrees of freedom, range of motion, flexibility and other parameters. The positioning of the joints and ability to move various phalanges of the fingers along with adduction and abduction capabilities were considered. An extra degree of freedom was introduced at the metacarpophalangeal MCP joint location of the exoskeleton fingers. Simulations have been carried out using MATLAB, Simulink and Adams software. The evaluated models from simulations are being prototyped for tests on real hands and efficacy, The parallel linkage mechanisms show most promising outcomes so far. Subsequent evaluation of forces encountered and its linkage to any pain/discomfort will be looked into. This work was presented at a conference.

Results and outcome till date

Three significant outcomes have been achieved so far: 1 A scheme has been evolved to identify Thumb, index and middle finger (of right hand) movements of 6 healthy patients from EEG data of Motor cortex with a 56 to 86 percent accuracy. This is indicative that one can expect to develop a BCI based technique to move the respective fingers through an assistive device like an exoskeleton. This work has been reported at an international conference.
2. An extensive study of 40 first time ischaemic stroke effected patients (23 male and 17 female Stroke Patients at Christian Medical College and Hospital Ludhiana.) was undertaken in which pre- and post-exercise baseline QEEG parameters such as mean Delta Theta Alpha and Beta frequency powers power ratio index delta vs alpha ratio and global brain symmetry index were studied. The progress was measured by comparing these parameters over multiple days 1, 7, 30 and 90. The study was expected to validate the EEG based evaluation of the progress under the prescribed exercise therapies. This has been reported in a conference.
3. A design process has been developed for various types of hand exoskeletons, namely linkage-driven and cable-driven mechanisms. These were studied for kinematic suitability with degrees of freedom, range of motion, flexibility and other parameters. The positioning of the joints and ability to move various phalanges of the fingers along with adduction and abduction capabilities were considered. An extra degree of freedom was introduced at the metacarpophalangeal MCP joint location of the exoskeleton fingers. Simulations have been carried out using MATLAB, Simulink and Adams software. The evaluated models from simulations are being prototyped for tests on real hands and efficacy, The parallel linkage mechanisms show most promising outcomes so far. Subsequent evaluation of forces encountered and its linkage to any pain/discomfort will be looked into. This work was presented at a conference.A repository of stroke patient data has been developed and is accessible over the internet. This has currently 40 stroke patient EEG data of Day 1 Day 2 Day 7 Day 30 and Day 90 in the form of pre and post exercise baseline data as well as video EEG and EMG data during physiotherapy exercises. DICOM Scan images of the stroke affected patient taken before initiation of treatment are also uploaded. Additionally a set of exercises prescribed by a physiotherapist are also stored in the repository. It is expected to have a large India specific data set over the balance part of the project thus creating a realistic repository for validation and future studies.
4. A CAD model of an exoskeleton hand has been designed and subsequently simulated to analyze finger flexion/extension.
5. All patients had undergone routine physiotherapy for hand for 15 to 30 minutes. The EEG data will be recorded for individual finger flexion/extension movements and different grasping actions. The subject had said to imagine their affected hand movement for grasping action.
6. The proposed exoskeleton model has one extra degree of freedom at the metacarpophalangeal MCP joint location of the exoskeleton fingers which could offer an added constrained motion in a direction parallel to the palm.
7. The dynamic analysis shows that the design can be used even if movement at one or more joints in the finger is restricted.
8. The design is investigated in SolidWorks for internal or external interference and the mechanism is visualized using animation.
9. The simulation and analysis in Adams help in achieving a better design of the hand exoskeleton.
10. Such a design can be fabricated by 3D printing and use of only two servos for actuation would ascertain cost efficiency.
11. We have analyzed five patients Pre & post baseline EEG data and thefollowing features were calculated. Feature calculation (Time, Frequency and Time-frequency domain).
12. We have applied quantitative EEG GUI application for computation the QEEG parameters such as mean Delta, Theta, Alpha, and Beta frequency powers, power ratio index, delta vs alpha ratio and global brain symmetry index and compare the changes observed in the pre & post-baseline exercises.
13. The simulation and analysis in Adams help in achieving a better design of the hand exoskeleton.

Societal benefit and impact anticipated

The rate of stroke is 84-262/100,000 in rural and 334-424/100,000 in urban areas in India. By actively engaging stroke patients in repetitive tasks with the device, the brain is expected to be able to rewire neurological pathways to motor functions to relearn movement. This study and the developed device with protocols can facilitate hand function recovery to enable such patients to return to work.

Next steps

After doing the initial work, multiple new issues and ideas have emerged. These will be addressed in the following ongoing phases: Phase-II: Conducting a safety and feasibility study of using low-cost robotic hand exoskeleton in normal subjects. Phase-III: Conducting a safety and feasibility study using low-cost robotic hand exoskeleton in subacute ischemic stroke patients. Phase-IV: To assess the efficacy of low-cost robotic hand exoskeleton to improve motor activity in the affected cerebral hemisphere using functional MRI.

Publications and reports

1) Sutanu Bera, Rinku Roy, Deebdeep Sikdar and Aupendu Kar; A Randomised Ensemble Learning Approach for Multiclass Motor Imagery Classification Using Error-Correcting Output Coding; 40th International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); July 17-21, 2018
2) Abhishek Sharma, C S Kumar, and Manjunatha Mahadevappa; Modelling and design of a multi-finger exoskeleton for post-stroke rehabilitation of hand motion; IFToMM Asian Mechanism and Machine Science, Asian MMS 2018; December 17-20, 2018.
3) Idha Sood, Bablu Tiwari, Dr Dimple Dawar, Dr Mahesh P Kate, Dr Jeyaraj D Pandian, Prof Manjunatha Mahadevappa, Prof C S Kumar, Quantitative electroencephalography to assess motor recovery in patients with first-ever stroke ,Junior World Congress-2019 (JWC-2019), April 4-8, 2019, at Christian Medical College, Ludhiana in Association with Asian Medical Students Association.
4) Shatakshi Singh, Ashutosh Pradhan, Koushik Bakshi, Bablu Tiwari, Dimple Dawar, Mahesh Kate, Jeyaraj Pandian, C.S. Kumar, and Manjunatha Mahadevappa, monitoring post-stroke Motor rehabilitation using EEG analysis; communicated to the 11th International Conference on Intelligent Human Computer Interaction; December 12-14, 2019.

Patents

None so far. Research activity is in a fundamental stage Patent for exoskeleton system will be filed after assessing the validity of initial tests.

Scholars and Project Staff

Several students and staff of IIT Kharagpur are associated with the project.
1) KoushikBakshi (P.hD Scholar, Since the start of the project)
2) Shatakshi Singh (P.hD Scholar, Since July 2018)
3) Bablu Tiwari (Junior Research Fellow, Since December-2017)
4) Sanjana Mohapatra (Junior Research Fellow, Sine May-2019)
5) MitrajitSamanta and Aditya Kameswara Rao Nandula from other projects at various phases
6) Ashutosh Pradhan (B.Tech. Student from April-2019).

Similarly, there are members of CMC-Ludhiana who are associated with the project:
1) Dr Rajeshwar Prasad (MD DM Neurology, Co- Investigator, from January 2019 onwards)
2) Dr. Dimple Dawar (M.O. Th.Neurology, Occupational Therapist/ Research Coordinator, Since April 2018 onwards )
3) Idha Sood (MBBS Student, from December 2018 onwards)
4) Rajiv Bhatti (Neurophysiology Technician, since July 2018)

Challenges faced

Getting sustained manpower for the project. Obtaining a safety certification of the new device being developed and thereafter getting it approved for clinical trials is challenging. The variety of stroke patients in different stages need different treatment protocols and providing such functionality in the hand exoskeleton is challenging. An additional robotic device is being envisaged for wrist motions.

Other information

A complete ICT paradigm of therapy prescription, monitoring and exercise compliance is taking shape in the form of a web/cloud-based IoT controlled system. This can be accessed and implemented in handheld/app-based setups.