Keywords

1 Introduction

According to a study by the World Health Organization (WHO), stroke has become one of the most serious health problems in the world. Approximately one-third of patients recovering from a stroke may suffer from life-long disabilities, of which motor dysfunction is a main sequela. Many researchers are currently dedicated to research involving stroke rehabilitation. Physiotherapy is one of the most common rehabilitation methods. However, upper-limb motor dysfunction is the most difficult sequela to be cured after stroke. Reducing permanent disability caused by stroke is still the most important goal of neurorehabilitation. Restoring the function of the arm and hand is very important during neurological rehabilitation [1].

The clinical effects of conventional rehabilitation therapies focus on improving walking ability; therefore, conventional rehabilitation is less effective in comprehensive recovery. In particular, hand motor dysfunction is very common in clinical practice. Although it is possible to recover some motor functions through neurorehabilitation, a large number of patients cannot achieve functional recovery of their affected arms after traditional rehabilitation exercises. However, many new interventions have emerged recently and can somewhat improve the motor functions of patients with stroke, including constraint-induced movement therapy (CIMT), bilateral movement therapy (BMT), mirror therapy (MT), robot-assisted therapy, task-oriented training (TOT), virtual reality (VR), motor imagery (MI), and functional electro-stimulation (FES). These approaches may partially improve upper limb functional impairment in patients with paralysis of the upper limbs after stroke and are supported by evidence-based medical efficacy [2, 3]. There is growing evidence that proactive repetitive motor exercises can help in the recovery of patients with stroke [4]. Intensive exercises can more effectively improve the movement ability and function of the upper limbs [4, 5].

In addition, transcranial magnetic stimulation (TMS) has been employed in recent years as an effective aid for the rehabilitation of patients with stroke [6]. Transcranial magnetic stimulation can inhibit ipsilateral brain activity or stimulate contralateral brain activity to help patients in reconstructing cortical function. Therefore, understanding the interaction between treatment and the brain is important in rehabilitation.

To this end, this study aimed to advocate a rehabilitation framework in combination with a visual interface using the concept of spontaneously induced neuromuscular therapy and to explore using a brain-computer interface for the acquisition of brainwave signals and to establish treatment information for the evaluation of a patient’s rehabilitation. We wish to address the possibility of improving a patient’s sense of accomplishment and compliance in rehabilitation.

2 Background

A study has suggested that the golden period of post-stroke rehabilitation span lasts approximately three to six months [7]; therefore, a discussion of how to assist patients with stroke to recovery in this golden period is urgently needed. Traditional upper-limb rehabilitation often applies sensory-motor training to promote the recovery of motor function, such as muscle strength, muscle endurance and range of joint motion. The current mainstream physical therapy for patients with stroke involves a task-oriented approach [8]. Clinically, upper-limb movements after stroke have slower and uncoordinated patterns during reaching. Therefore, reaching motion is often considered a key parameter for the assessment of upper limb function after stroke [9]. In addition, the latest studies suggest that exercise training based on repetitive movements of specific upper-limb muscles results in significant improvements in upper-limb function recovery in patients with hemiplegia [10].

Mechanically assisted treatment has been developed for over a decade to make up for the lack of trainers and to overcome patient movement restrictions in traditional rehabilitation and to more objectively and instantaneously record working data. Studies have demonstrated that clinically proven bilateral upper-limb rehabilitation devices, such as the MIT-MANUS, BI-MANU-TRACK, BATRAC (bilateral arm training with rhythmic auditory cueing), and MIME (mirror image motion enabler), have better efficacy in mechanically aided rehabilitation than traditional therapies [11]. However, the current bilateral upper-limb biomechanical devices are not equipped with auxiliary visual or brainwave interfaces.

In recent years, many brain-computer interfaces have been used in the health care industry and can be divided into proactive and reactive interfaces [12]. The reactive device can record and collect physiological signals, such as brain waves, as a reference for monitoring or care, and the proactive device can provide brainwave control or interactive modes and has certain advantages, including real-time acquisition, non-intrusive monitoring and low cost. Its applications in industry and academic fields are becoming increasingly popular [13]. The brain-computer interface can also be used in motion training programs. For example, motor imagery (MI) training was originally used in the training of athletes. Since 2006, some researchers have discussed the possibility of using MI in rehabilitation after stroke [14], especially in patients with severely impaired motion after stroke. In addition, in terms of assessing the effectiveness of rehabilitation, the brain-computer interface (BCI) has certain advantages, including higher temporal resolution, low cost, long-term monitoring, and the use of noninvasive measurements. Studies have assessed and confirmed that BCI can now distinguish between the intention of moving the patient’s shoulder or elbow [15]. According to the literature, there have been studies on the effects of using reactive/proactive brain-computer interface games to enhance the effectiveness and positive attitudes of patients in rehabilitation [12]. However, no studies have evaluated the possibility of employing brain-computer interface games with various proactive and reactive interactive modes in rehabilitation devices.

Therefore, this study is a scoping study on the development of an interactive upper-limb rehabilitation system controlled by a brain-computer interface and is expected to propose an interactive framework system for upper-limb rehabilitation that is controlled by a brain-computer interface, using the concept of spontaneously induced neuromuscular therapy to monitor the cerebral motor area, hand muscle rehabilitation status, and the interactions between the treatments. The system framework includes hardware and software to assist the user in rehabilitation training and, at the same time, records the movement trajectory of the user using the interactive rehabilitation training system. The data recorded can be used to evaluate the effectiveness of the user’s rehabilitation training and to compare with the EEG and EMG signals at the corresponding time points.

3 Conceptual System Framework

The framework of the interactive upper-limb rehabilitation system with brain-computer interfaces developed in this study is shown in Fig. 1. The system mainly includes an interactive rehabilitation training platform, a rehabilitation database system, and an EEG and EMG acquisition system. The interactive rehabilitation training system platform includes a virtual rehabilitation game system and an interactive upper-limb rehabilitation device by which a user can perform proactive and reactive rehabilitation. In addition, the EEG and EMG acquisition system also includes measuring electrodes and a mechanical design to measure the user’s EEG and EMG signals. The user can first activate the motor-sensory cortex through repeated transcranial magnetic stimulations and can then perform rehabilitation training through the interactive rehabilitation training system platform. At the same time, the interactive rehabilitation training system platform can track the user’s movement, while the EEG and EMG acquisition system can measure the electrical signals of both the user’s brain and their hand muscles during rehabilitation training. These data collected can be used to analyse the effect of the cerebral cortex activation by repeated TMS and the hand muscle condition and thus can be used to analyse the effect of rehabilitation training. The interactive rehabilitation training system platform can guide the patients to correctly complete the pre-set rehabilitation motion through a virtual rehabilitation game and can give feedback about the user’s rehabilitation by analysing data from the EEG and EMG acquisition system and the interactive upper-limb rehabilitation device. These data on the patient’s rehabilitation status are then stored in the rehabilitation database system for the medical care providers’ reference.

Fig. 1.
figure 1

The framework of the interactive upper-limb rehabilitation system with brain-computer interface

Full size image

3.1 Interactive Rehabilitation Training System Platform

The interactive rehabilitation training system platform includes a proactive and reactive rehabilitation training device, a rehabilitation game visual interface, and a brain-computer interface. The rehabilitation training device, which is designed using an electronic control device to drive the gears, can help a stroke patient to move their shoulders and elbows in a symmetrical manner. The speeds of both proactive and reactive motion can be adjusted. During training, the EEG brain-computer interface is used to capture brain signals from the motor area of the cerebral cortex. This approach, using virtual games with a visual interface, can make training become goal-oriented and fun, resulting in recuperation compliance; the electromechanical and mechanical testing is conducted, and the back-end database of the system is designed based on this concept. In addition, this study aimed to utilize user-centred rehabilitation games and a human-computer interaction interface to facilitate researchers in obtaining EEG and EMG response signals through a friendly interface. Then, the biomedical signals and algorithm are further processed to help users in integrating the visual interaction system.

3.2 EEG and EMG Control System

The system framework of this study presents an EEG and EMG control system that includes a miniaturized EEG and EMG acquisition system. The integrated system components and structure of the miniaturized EEG and EMG acquisition system are shown in Fig. 2 and include a front-end amplifier, analogue MUXs, programmable gain amplifiers (PGA), analogue-to-digital converters (ADC), a microcontroller unit (MCU), a power management unit, and a communication circuit. The EEG or EMG signals captured by the electrode will be amplified by the front-end physiological amplifier. Then, the analogue multiplexer will select the EEG or EMG channel to send the signals to the PGA for signal amplification and to the ADC for signal digitalization. The digitized signals are transferred to the microcontroller unit. Finally, the EEG and EMG data are transmitted to an external device for signal analysis. Generally speaking, the size of the brainwave signal is less than 100 µV, and its frequency ranges between 0.1 and 100 Hz. Therefore, the design of this system circuit must include the following features: (1) high differential gain that can amplify small brain wave signals; (2) high CMRR to reduce common mode noise; (3) high input impedance; (4) high SNR; (5) low power consumption; (6) low noise; and (7) small chip area.

Fig. 2.
figure 2

EEG and EMG acquisition system

Full size image

3.3 Rehabilitation Database System

The system framework proposed in this study is a rehabilitative database system, as shown in Fig. 3. The rTMS setting parameters can be recorded asynchronously, and the user’s EEG/EMG signals recorded by the EEG/EMG control system can be stored in the system simultaneously for virtual rehabilitation games to set up the parameter values of the game environment. The user can play the virtual rehabilitation game on an external monitor and use the proactive and reactive upper-limb rehabilitation devices that connect to this system.

Fig. 3.
figure 3

The framework of the rehabilitation database system

Full size image

Most users requiring upper-limb rehabilitation after stroke may need long-term rehabilitation monitoring and EEG signal recording. Therefore, the system framework is better designed to use a cloud database for information storage on the back-end, which can be accessed by authorized medical professionals. In this system framework, a user-directed virtual rehabilitation game is designed according to medical professional recommendations. The system obtains physiological signals from the EEG and EMG control system and quantifies and records signals. Quantified information will be used for real-time fine-tuning and virtual rehabilitation game design based on user needs. The main functions of the rehabilitation database are to retrieve biomedical signals, process digital signals, record usage of the virtual rehabilitation game, and back up the patient’s information. In addition, a log recording information access will also be stored in the database to enable medical professionals to take advantage of the system to follow up patient recovery and perform medical analyses. The system also provides a patient account to allow patients to log in to their personal accounts and to obtain their own personal rehabilitation information. The back-end database will be used to back up patients’ health and personal information.

The main users of this system are caregivers and therapists because many stroke patients are not able to use computers and other digital devices due to their upper-limb disabilities. Caregivers can obtain information on the rehabilitation regimen and efficacy for a patient, basic rehabilitation knowledge, and the therapist opinions using the system.

4 Conclusion

This study has proposed a framework for an interactive upper-limb rehabilitation system with a brain-computer interface and confirmed the integrated system specifications and the configuration of rehabilitation devices according to the framework. The importance of the development of this innovative system is to train patients to perform proactive and reactive rehabilitation exercises after transcranial magnetic stimulation and to use the interactive interface of a visual computer (i.e., to play games) to enhance the rehabilitation motivation of patients after stroke. This system also provides clinicians and therapists with evidence of the patient’s rehabilitation efficacy. In the future, when the new system is developed based on the system framework proposed in this study, the framework will effectively guide the system’s development and clinical validation.