Abstract
BACKGROUND:
Neurofeedback training targets the relevant brain response under minimal stress. It could be a promising approach for the treatment of patients with brain injury.
OBJECTIVE:
This review aimed to examine the existing literature to confirm the effectiveness of applied electroencephalogram (EEG)-based neurofeedback training in the area of occupational therapy for upper limb stroke rehabilitation.
METHOD:
All relevant literature published until July 1, 2020 in five prominent databases (PubMed, CINAHL, PsycINFO, MEDLINE Complete, and Web of Science) was reviewed, based on the five-step review framework proposed by Arksey and O’Malley.
RESULTS:
After a thorough review, a total of 14 studies were included in this review. Almost studies reported significant improvements as a result of EEG-based neurofeedback training, but this had not always account for the differences in effectiveness between groups. However, the results of these studies suggested that neurofeedback training was effective as compared to the traditional treatment and more effective in combination with EEG than that with simple equipment application.
CONCLUSION:
This review demonstrated the effectiveness of the combination of occupational therapy and EEG-based neurofeedback training. Most of these treatments are intended for inpatients, but they may be more effective for outpatients, especially if customized to their requirements. Also, such explorations to assess the suitability of the treatment for patient rehabilitation will help reduce barriers to effective interventions. An analysis of the opinions of participants and experts through satisfaction surveys will be helpful.
Introduction
Occupational therapy aims to reduce motor dysfunction and restore daily life activity, especially in persons with hemiplegic stroke. Conventional treatments to restore the function of the affected limb have been effective for people with mild to moderate paralysis, but an innovative approach is required for those with severe paralysis (Wijekoon et al., 2020).
Recently, due to an increased understanding of brain plasticity, new possibilities have emerged for the rehabilitation of persons with stroke; bio- and neurofeedback therapies for stroke rehabilitation could be new therapeutic paradigms. Neurofeedback training is a promising rehabilitation technique that can reconstruct the motor neurons affected by brain nerve damage; it is an intervention method that detects motor-related neurological activity and provides feedback to patients through vision and/or hearing (Wada et al., 2019). Motor image-induced event-related desynchronization is a brain-machine interface (BMI) applied for neurological rehabilitation in patients. Additionally, non-invasive electroencephalography (EEG) is the most common method for brain signal acquisition used with BMI for rehabilitation purposes; it provides a highly efficient BMI platform. EEG signals for motor rehabilitation are primarily detected during task-based EEG and are used for neural feedback during rehabilitation (Mattia et al., 2020; Zhuralvev et al., 2020). This neural feedback may be suggested to patients with severe stroke because it does not necessarily require active motor control. According to a stroke rehabilitation study, a variety of feedbacks can enhance exercise-related brain signals that restore major motor pathways. These feedbacks have demonstrated moderate success in the motor rehabilitation of patients with severe stroke (Vourvopoulos et al., 2019; Wang et al., 2019). Recent evidence furthers the suggestion that combining conventional rehabilitation and neural feedback can improve exercise outcomes in stroke survivors, which can represent a new approach to functional recovery after stroke (Foong et al., 2019). Real-time neurofeedback allows the patient to participate in neuromodulation, and induces brain plasticity to facilitate functional recovery. In addition, its combination with external devices (i.e. exoskeleton or soft robotic glove) enhances the activation of the cortex and peripheral nerves, and the synchronization of sensory feedback. However, a comprehensive conclusion about the effectiveness of this treatment when applied together with existing therapeutic activities in a realistic rehabilitation environment, is lacking. (Singh et al., 2019). Additionally, most reviews confirming the effectiveness of this therapy for stroke rehabilitation focused on simple neurofeedback therapy to train brain activity directly.
In this review, considering the rehabilitation environment, the aim was to confirm the effect of a combined treatment of brainwave-based neurofeedback training and occupational therapy for upper limb rehabilitation in people with stroke. This was achieved by reviewing the existing literature on EEG-based neural feedback training applied in the rehabilitation of the upper limb impairment after stroke. This scoping review will provide an overview and analysis of previous studies to determine the direction of available studies and to propose further studies.
Methods
This scoping review does not evaluate the entire literature; rather, it answers specific questions. The literature was examined, available quantitative data was collated, and then the literature was summarized and interpreted in a particular field of study. To date, no systematic review has summarized the subdivided topics on the effectiveness of the combination of traditional occupational therapy and neurofeedback training in upper limb stroke rehabilitation in a clinical environment.
The five-step review framework was employed, as proposed by Arksey and O’Malley. The first step was to define a research question. Relevant studies published up until July 1, 2020 were searched in five electronic databases: PubMed, CINAHL, PsycINFO, MEDLINE Complete, and Web of Science.
In the first stage, relevant research was reviewed and three initial exploratory research questions were selected: (1) What are the possible explanations for the therapeutic effects of EEG-based neurofeedback training combined with occupational therapy in persons with stroke over the past 10 years? (2) What are the singularities of past methodologies and results? (3) What is the prevalent direction of the research, and what should be considered when proceeding in the field? (Table 1).
Description of EEG-based neurofeedback training
Description of EEG-based neurofeedback training
EEG: Electroencephalography.
In the second stage, the following eligibility criteria were established: (1) journal article type, (2) articles published from 2010 to present, (3) articles written in English, and (4) studies aimed at confirming the combined therapeutic effectiveness of EEG-based neurofeedback training and occupational therapy for upper limb rehabilitation in persons with stroke. The following search terms were combined: (biofeedback OR neurofeedback OR electroencephalography OR brain waves OR brain biofeedback OR brain-computer interface) AND (occupational therapy OR occupational therapist OR rehabilitation) AND (stroke OR CVA OR cerebrovascular accident OR hemiplegia OR brain attack OR motor impairment OR limb OR extremity). The pertinent articles from the five databases were exported and managed using the RefWorks referencing software program (ProQuest LLC, Ann Arbor, MI, USA).
In the third stage, the titles and abstracts were reviewed for the initial selection. Then, the complete studies were read and evaluated subject to the eligibility criteria. To achieve maximum specificity, this review did not include studies in which the participants had other underlying pathological problems. Studies were also excluded if they had been published as a poster, book, or magazine, or were not written in English. From the final selected articles, data were extracted into six categories that were used to analyze the full reviews: (1) role of occupational therapist, (2) specialty of the first author, (3) study design, (4) stroke type and severity, (5) outcome measure, and (6) main finding of the analyses. Data extraction was performed independently by the reviewer. The process of the study selection was illustrated using a flow diagram (Figure 1), and the findings were tabulated with a narrative description (Table 2, Table 3).
Flowchart of illustrating inclusion process.
Summary of study characteristics from 14 studies
BBD: Both brain damage, Hem: Hemorrhage, Inf.: Infarction, LBD: Left brain damage, QE: Quasi-experimental group, RBD: Right brain damage, RCT: Randomization control trials.
Summary of neurofeedback intervention characteristicsand key results from 14 studies
ADL: Activities of daily living, AF: Auditory feedback, aFA: Asymmetric fractional anisotropy, AOT: Action observational training, ARAT: Action research arm test, CACR: Computer-assisted cognitive rehabilitation, Con.: control group, DMB: Digital mirror box, EEG: Electroencephalography, EMG: electromyography, ERD: Event-related desynchronization, FES: Functional electrical stimulation, FIM: Functional independence measure, FMMA: Fugl-Meyer assessment of motor recovery after strokes, FMMA-A: shoulder/elbow/forearm score, FMMA-C: hand/finger score, fMRI: functional magnetic resonance imaging, HD: Horizontal distance, JHFT: Jebsen hand function test, MAL: Motor activity log, MAL-AOU: Motor activity log-quantity of movement, MAS: Modified ashworth spasticity, MFT: Manual function test, MMSE-K: Mini-mental state examination-Korea, movt: Movement, NF: Neurofeedback, NMES: Neuromuscular electrical stimulation, n.s.: not significant, rBSI: revised brain symmetry index, ROI: Region of interest, SIS: Stroke impact scale, SMR: Sensorimotor rhythm, SPD: Spectral power density, TMS: Transcranial magnetic stimulation, TS: Tactile stimulation, UE: upper extremity, VD: Vertical distance, VF: Visual feedback, VR: Virtual reality, 9-HPT: Nine-hole peg test.
In the final stage of the survey, data were systematically categorized and organized (Table 3), using a data-charting form developed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).
Study selection and characteristics
Figure 1 summarizes the steps involved, the search results, and the reasons for excluding studies. After a thorough review, 466 articles were excluded so a total of 14 studies were included in this review. The results were summarized by topics, and the order of factors did not indicate priority (Table 1, Table 2). In all studies, the role of the occupational therapist correlated with the evaluation and application of the treatment, and the first authors had specialized either in occupational therapy, physical therapy, or rehabilitation engineering. With regards to study design, six were randomized controlled trials (RCTs), two were quasi-experiments, five were single-group studies, and one was a single-subject design study. The number of participants per study ranged from 1–48; the number of patients with left-brain injury was 119, and that with right-brain injury was 99. Eleven studies confirmed the severity of the symptoms of the participants; subjects with moderate to severe effects of stroke participated. Ten studies were conducted with inpatients, and the remaining four were outpatient interventions (Table 2).
Participants and methodology
All 14 studies included in this review were interventional studies with persons with stroke. Ten of the studies applied the EEG-based neurofeedback training in conjunction with external devices. Five of them involved electrical stimulations, such as functional electrical stimulation (FES) and neuromuscular electrical stimulation (NMES). Five were connected with external equipment, such as virtual reality (VR) equipment, exoskeleton, and soft robotic glove. Three studies employed a combination of beta/sensorimotor rhythm (SMR) training and audiovisual feedback; whereas, one study applied only the auditory feedback. In this review, only the EEG-based neurofeedback training applied during occupational therapy was considered. The intervention was applied two to five times a week for 20–90 minutes and lasted more than ten times. Most of the studies (N = 10) applied EEG-based neurofeedback training in addition to the existing occupational therapy, and the remaining studies (N = 4) applied it as part of the occupational therapy. The main purpose of these interventions was to improve the use of upper and lower limbs (N = 12). Two studies applied this therapy to restore upper extremity and cognitive functions (Table 3).
Effectiveness of EEG-based neurofeedback training: Randomized controlled trials
Table 3 describes the six RCTs identified in our review (Ang et al., 2015; Cheng et al., 2020; Cho et al., 2016; Jang et al., 2016; Jung & Shim, 2012; Kim, T. et al., 2016). All six studies reported significant improvements as a result of EEG-based neurofeedback training, but this did not account for the differences in effectiveness between groups. Only two studies confirmed the effectiveness of the groups (Jang et al., 2016; Kim, T. et al., 2016). Three studies compared the differences in the effectiveness of conventional treatments and control groups (Cho et al., 2016; Jung & Shim, 2012; Kim, T. et al., 2016). In addition, three studies confirmed the therapeutic effect of therapy when applied in combination with a BMI-type external device (Ang et al., 2015; Cheng et al., 2020; Jang et al., 2016).
Ang et al. and Cheng et al. employed exoskeleton and robotic gloves in conjunction with EEG and observed recovery in the upper limb function. Both studies reported improvements in the Fugl-Meyer assessment scores before and after treatments; however, they did not observe any difference in effectiveness when compared with the treatment without EEG (Ang et al., 2015; Cheng et al., 2020). On the contrary, Kim et al. and Jang et al. applied a combination of EEG and FES; Kim et al. compared it to the traditional treatment method, and Jang et al. compared the difference between the treatments with or without EEG. The results of these studies suggested that neurofeedback training was effective as compared to the traditional treatment and more effective in combination with EEG than that with simple FES application (Jang et al., 2016; Kim et al., 2016). Two studies applied the beta/SMR training technique with visual and auditory feedback and confirmed the effectiveness of visual perception and cognitive ability. The results of the study aimed only to confirm the changes in EEG rather than to assess functional improvements (Jung & Shim, 2012; Rayegani et al., 2014). No study reported the difference in the number of treatment sessions; however, they applied an average of 20 interventions, with an average time of 0.7 h per session. One of the RCTs reported the maintenance of the effectiveness of long-term treatment, wherein the patient retained the limb function for more than a year after the treatment session (Cheng et al., 2020).
Effectiveness of groups with EEG-based neurofeedback training: Quantitative research
Five pre-post studies of this group’s results have been reported, but the effectiveness of the treatment method could not specifically described (Nakano et al., 2018; Nishimoto et al., 2018; Ono et al., 2018; Vourvopoulos et al., 2019; Wada et al., 2019). Two studies were non-RCT, one of which compared the experimental and control periods, and another directly compared the effects of neurofeedback and biofeedback training (Rayegani et al., 2014; Young et al., 2016). One study was a single-subject design study (Mukaino et al., 2014) (Table 3).
Wada et al. and Ono et al. reported improvement in the muscle tone in hand by connecting a digital mirror box to the EEG (Ono et al., 2018; Wada et al., 2019). In particular, Wada et al. additionally linked an exoskeleton to EEG and confirmed the effectiveness of the functional side of the hand as well as the muscle tone. Vourvopoulos et al. and Nishimoto et al. reported improvement in the upper extremity and hand function by linking EEG with VR equipment and NMES devices. Vourvopoulos et al. especially reported the effectiveness of the additional connection of electromyogram in their treatment (Nishimoto et al., 2018; Vourvopoulos et al., 2019). Mukaino et al. and Nakano et al. reported improvements in upper and lower extremities and hand function using NMES or audiovisual stimulation. However, this study could not confirm the statistical significance (Mukaino et al., 2014; Nakano et al., 2018). Young et al. and Rayegani et al. mixed relatively complex external stimuli in their studies. Young et al. linked the EEG to the FES and tactile stimulation devices, and Rayegani et al. reported improvement of upper and lower extremities and hand function using audiovisual feedback, while executing the neurofeedback game program (Rayegani et al., 2014; Young et al., 2016). No study reported the difference in the number of treatment sessions; however, they conducted an average of 11 sessions with 0.7 h per session.
Discussion
This scoping review was conducted to address three main research questions. The first question that was answered was, “What are the possible explanations for EEG-based neurofeedback training combined with occupational therapy in persons with stroke over the past ten years?”
This review revealed that most of the included studies were quantitative in design and focused on inpatients. Few studies applied EEG-based neurofeedback training to confirm its effectiveness in improving visual perception and cognitive function along with upper limb function. Although not highlighted in these studies, it is noteworthy that improved visual perception and cognitive function can significantly contribute towards the recovery of upper limb function; both these functions markedly influence motor function. However, when considering the treatment area for occupational therapy, the number of studies that investigate the effect of complex therapy for improving these functions to, in turn, improve upper limb and hand functions, is surprisingly insufficient. Persons with stroke may experience a variety of visual perception and cognitive impairments. Considering the advantages of EEG in combination with devices that can apply a variety of feedbacks, it is expected to be a more effective intervention than traditional treatment methods.
In the field of occupational therapy, the therapeutic continuum leading to the completion of occupational participation with functional recovery is important (Kongwattanakul et al., 2020; Radhakrishnan et al., 2019). However, this continuum was not evident as the main result of this review. This suggests that EEG-based neurofeedback training provided with occupational therapy should aim to improve the occupational participation of human life. Most of the research in this field is focused on inpatients, but it can be applied to outpatients as an additional rehabilitation method using self-training. Additionally, it could be easier to focus on functions that fit the actual situation in outpatients. Most interventions are aimed at a specific disability or function. Only one study reported functional interventions that could be applied to real-life situations. Therefore, additional research is needed to link daily activities and traditional occupational therapy interventions.
Regarding our second research question, “What are the singularities of past methodologies and results?” most studies confirmed the effectiveness of neurofeedback training together with EEG and external equipment in improving upper extremity and hand function during occupational therapy (Calabrò et al., 2019; Carino-Escobar & Cantillo-Negrete, 2020; Chang et al., 2017; Fuentes et al., 2018). Furthermore, other research directly compared the interventions either integrated with, or without, EEG. Moreover, Rayegani et al. reported an improvement in hand function by comparing biofeedback-type and neurofeedback-type occupational therapy (Rayegani et al., 2014). In our review, the effectiveness of the combination of occupational therapy and EEG-based neurofeedback was evident. For example, EEG-based neurofeedback training combined with traditional occupational therapy could make a difference in the application and effectiveness of traditional therapy (Kim, T. et al., 2016). They discussed the effectiveness of additional intervention rather than the therapeutic effect of traditional occupational therapy, and evaluated the actual situation. The researchers emphasized that as the characteristics of the subject’s neurological symptoms varied, it was necessary to evaluate the performance of specific activity rather than the effectiveness of simple movements (Haji-Ahmad et al., 2015; Ikeda et al., 2020; Kakui et al., 2018; Kondo et al., 2015; Oblak et al., 2019). Some studies reported the persistence of treatment effects. Effects of treatment were found to last from 4 months to 1 year after treatment was stopped (Ang et al., 2015; Cheng et al., 2020). In particular, this scoping review highlights the limited nature of research on the persistence of therapeutic effects. There is no study that compared the effects of interventions at the time of onset. Such studies will help predict the therapeutic- and cost-effectiveness of treatment (Badawi & El Saddik, 2020; Power et al., 2020).
The third question guiding this review was, “What is the prevalent direction of the research, and what should be considered when proceeding in the field?” Overall, this review barely performed a qualitative study of the adaptability of patients and clinical impacts of EEG-based neurofeedback occupational therapy interventions in the rehabilitation of persons with stroke. The participation of users and service providers is extremely important in the development of rehabilitation services and systems. In addition, their experiences would provide insight to increase the efficiency and effectiveness of the rehabilitation service. Although one study in this review included consumer feedback on interventions (Cheng, 2019), the information was insufficient to evaluate the quality of the service. In the future, it would be valuable to collate the opinions of participants and experts through satisfaction surveys and in-depth analysis (Locke et al., 2020; Singh et al., 2019; Tolin et al., 2020; Yoon & Kim, 2015; Yu et al., 2012).
Neurofeedback training, targeting related brain responses while reducing stimuli that cause brain stress, could be a promising approach for the treatment of patients with brain damage. Additionally, EEG-based neurofeedback is one of the relatively inexpensive and readily applicable techniques. However, the number of studies on the effectiveness of application in the field of occupational therapy is scarce. This review revealed that participants who underwent EEG–FES-linked interventions during occupational therapy were more effective in restoring upper extremity and hand function compared with the control group (Jang et al., 2016; Kim et al., 2016). The authors acknowledge the limitations of this review, including the small sample size and techniques used for measuring the results, highlighting the need for further clinical studies. Future studies should assess the suitability of treatment with the main purpose of patient rehabilitation. In addition, a full understanding of the clinician’s use of the device and further exploration to promote treatment will reduce barriers to effective treatment interventions and help mediators in providing valuable services. Inputs of participants can also be useful in determining the facilitation of processes by clinicians to encourage meaningful participation and interaction during interventions. Additional research that reflects the perspectives of users and clinicians, as well as differences in the benefits of each type of intervention, will be important for service development and improvement, thus suggesting the need for qualitative research. Moreover, it will be necessary to evaluate the cost-effectiveness according to the health and economic conditions of users.
Conclusions
Through this review, it was determined that EEG-based neurofeedback occupational therapy interventions in persons with stroke could have a synergistic effect along with conventional occupational therapy interventions. Combinations of various types of devices were effective in restoring their upper and lower limb and hand functions, and deviceless interventions had relatively lower effectiveness. In order to fully understand the field, understanding the viewpoints of users and clinicians will be necessary for effective intervention. Additional research is imperative to compare each type of treatment.
Author contributions
The author has screened the literature and selected papers for inclusion in the review.
Conflict of interest
None to report.
Declarations
The results have not been published previously and are not under submission elsewhere.
Ethics approval and consent to participate
Not applicable.