Introduction

Stroke is the second leading cause of death after ischemic heart disease, accounting for 9% of all fatalities worldwide [1]. The most common deficit after stroke is hemiparesis of the contralateral upper limb (UL) [2]. Stroke is one of the major reasons for disability [3]. Nearly, 33% of stroke survivors experience disability, particularly upper extremity (UE) disability [3, 4]. UE disability affects 50–80% of stroke patients in the acute phase [5,6,7] and 40–50% in the chronic phase [7, 8]. Paresis, abnormal tone of muscle, diminished somatosensation, and problems with coordination are common UE deficits after stroke [9]. UE motor impairment results in limitations of activities of daily living (ADL) [10] and lowered quality of life [10, 11].

Transcranial direct current stimulation (tDCS) is a non-invasive, safe, and relatively painless brain stimulation technique [12]. Constant weak electrical current is applied for few minutes via electrodes, placed in well-considered locations [13, 14]. Majority of the studies administer tDCS between 2 rubber electrodes 25–35 cm2, placed over the scalp using a low-current intensity, ranging from conventional 1–2 mA up to currents of 4 mA for 10–20 min [15,16,17]. It has been substantially scrutinized in cognitive and clinical neuroscience [18] and, is tolerated well by stroke patients with minimal and temporary adverse effects [19] when the protocol is performed according to the current safety guidelines [20]. Robot-mediated rehabilitation is a cutting-edge exercise-based therapy that uses robotic equipment to deliver extremely repetitive, rigorous, adaptable, and quantified physical training [21]. By offering complicated but controlled multimodal stimulation [22], robotic therapy (RT) has been presented as a potential strategy for the rehabilitation of the UE, as a way to increase the amount and intensity of therapy [22] and standardize the treatment [23]. Furthermore, robotic devices can provide quantitative measures of the user's dexterity due to their built-in technology in terms of sensors and actuators [24].

A considerable number of research papers on robot-assisted stroke rehabilitation have been published, examining the effects of robotics alone [25,26,27,28,29,30] and in combination with traditional therapy [31,32,33]. A Cochrane review published in 2012 [34] found that rehabilitation-assisted electro mechanics and robotics may assist in improving the function of the arm after stroke, so there might be an improvement in daily living activities, muscle strength, and arm function. Although, results of treatment depend on practice intensity, training amount and duration, treatment type, type of device, and characteristics of participants as they are potential drivers for efficacious motor rehabilitation interventions after stroke [35]. Extensive research has been done on the effects of tDCS on arm function and motor learning in stroke patients but conclusions drawn from these studies reported mixed inferences [36]. To this point in time, no systematic review has been conducted which determines the effect of combined tDCS and RT in different types of strokes. This review aims to fill this slot in literature to date. Therefore, the purpose of the present review was to comprehensively and systematically evaluate clinical trials to investigate the effect of tDCS in combination with RT in UE impairments in patients with stroke.

Methods

This review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [37]. The protocol of this review was registered on PROSPERO (International Prospective Register of Systematic Reviews) and can be accessed by registration number CRD42021238334.

Study selection criteria

The inclusion criteria for the selection of studies were: full-text randomized controlled trials published in the English language; people clinically diagnosed with acute, subacute, or chronic stroke (ischemic/hemorrhagic) and having UE impairment; with age > 18 years; studies administering a combination of tDCS and RT; studies having at least two of the validated clinical outcome measure. Trials including patients with cognitive impairment, having preceding epilepsy or intracranial metal implant impairments and studies done on animals were excluded.

Search strategy

MEDLINE (accessed by PubMed), Web of Science (Web of Science Core Collection), ScienceDirect, and PEDro were searched for relevant literature between inception to February 2021. Search terms used were a combination of keywords, “transcranial direct current stimulation, robotics, upper limb, and stroke”. Boolean operators ‘AND’ or ‘OR’ were used. Three authors (AR, SP, and FB) independently searched and then compiled all the records. The titles and abstracts of records identified were analysed. Full-text articles were assessed to meet the eligibility criteria after removing the redundant studies.

Quality assessment of studies

Two reviewers (AR and SP) independently performed the quality assessment of selected studies using PEDro scale which is an 11-point scale to determine the quality of clinical trials. Any disagreement that remains unresolved was taken to the third reviewer (MN). To reduce ambiguity in responses, each criterion was rated either yes (score = 1) or no (score = 0). Summing all of the replies yielded a total score for each included study's methodological quality (maximum score = 10). Based on the total score, studies were classified as poor (score of 4), fair (score of 4 or 5), good (score of 6–8), and excellent (score of > 8) quality [38].

Results

Study selection

A sum of 208 articles was identified after a systematic search of databases. A total of 55 duplicates were removed. The remaining 153 articles were then screened by reading their title and abstract. Of these, 136 articles were excluded after screening. The remaining 17 articles were identified and were assessed for eligibility by three reviewers (AR, SP, and FB). A total of six studies were considered for qualitative analysis (Fig. 1).

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram of the study

Full size image

Characteristics of study

Participants

Six papers included in this review had a total of 213 patients with stroke. Their age ranged from 18 to 90 years. The sample size ranged from 21 participants [39] to 82 participants [40]. Details of sample size calculation were reported in only one study [41], where the sample size was determined using the PS Power and sample size calculation software (version 3.1.2, department of biostatistics, Vanderbilt University, USA). In other studies, the method of calculation was not mentioned [3, 4, 36, 39, 40]. Two studies included patients with subacute stroke [3, 41], two with chronic stroke [39, 40], and two trials included both subacute and chronic stroke patients [4, 36]. All the studies included both males and female in different proportions (Table 1).

Table 1 Characteristics of included trials (n = 6)
Full size table

Intervention

In three studies, a current of 2 mA [3, 40, 41] was used and the rest of the studies used 1 mA [4, 36, 39]. Only one study used a single session of RT and tDCS, while the rest of the trials used multiple sessions of RT and tDCS, which ranged from 2 sessions [4] to 36 sessions [40] per week. The duration of the treatment session varied from 20 min [39] to 1 h and 20 min [4, 40] (Table 2).

Table 2 Characteristics of intervention used in included trials (n = 6)
Full size table

Outcome measures

Four studies used box and block test (BBT) for gross and fine manual dexterity [3, 36, 39, 41]. Four studies used Fugl–Meyer assessment scale (FMAS) [3, 4, 36, 40, 41], two studies used motricity index (MI) [3, 41], and two studies employed motor activity log (MAL) [4, 36]. For the assessment of spasticity, modified Ashworth scale for the wrist (MAS/w) was used in two trials [3, 41]. Only one study used action research arm test (ARAT) [4] (Table 1).

Quality assessment of included trials

The average PEDro score for all the included studies was 8.5 (excellent quality). Four trials scored 8/10 [3, 36, 39, 41], one scored 9/10 [4], and one scored 10/10 [40]. Three studies lacked allocation concealment [36, 39, 41]. All the included trials reported blinding of the participants. Blinding of the therapist was reported in only one trial [40]. There was a blinding of assessors in two studies [3, 41]. All the studies scored well on reporting between-group differences, random allocation, point estimation, and variability reporting. All the studies applied to intention to treat analysis of drop-outs (Table 3).

Table 3 Quality assessment of included trials (n = 6)
Full size table

Effect of RT and tDCS on UE function and spasticity

Five studies found positive improvement in arm motor recovery when measured with FMAS-UE [3, 4, 36, 40, 41]. Two studies showed no significant positive change in MAS/w score [3, 41]. Five studies showed an increase in BBT score, out of which only three studies reported a significant increase in BBT score [3, 36, 41]. No significant change was reported in scores of MAL in two trials [4, 36]. Two studies showed positive significant change in MI [3, 41]. Two studies administered tDCS first followed by RT [4, 40]. Three studies administered tDCS simultaneously with RT [3, 39, 41]. One trial did not make it clear whether RT was administered first or tDCS or both were delivered simultaneously to the patient [36].

Discussion

This is the first systematic review providing information on major findings, characteristics, and quality of studies investigating the combined effect of RT and tDCS on UE motor function and spasticity in people with stroke. Although the direct pooled analysis was limited due to the variability of the outcome measures, the conclusion obtained from the existing evidence implies that stroke patients receiving RT combined with tDCS had a favourable effect on UE motor function of stroke patients having UE impairment as indicated by positive changes in FMAS-UE, MI, and BBT score.

The studies provided considerable motor improvements, evaluated by clinical scales, although, no superiority of real tDCS was seen over sham tDCS. Straudi and her colleagues administered 10 sessions of RT and real tDCS in patients with acute and chronic stroke for 2 weeks. When these patients were compared with stroke patients receiving RT and sham tDCS, no significant difference was reported in measures of hand function [36]. Similar findings were reported by Edward et al. [40], Mazzoleni et al. [3], Triccas et al. [4], and Mazzoleni et al. [41]. Dehem et al. [39] showed a significant difference in scores of BBT, after patients received RT + real tDCS. All studies reported no significant difference in scores of FMAS-UE between groups receiving RT + tDCS and RT + sham tDCS. A possible explanation for the similarity between the groups could be the relative timing of delivering the stimulation or the effectiveness of tDCS overlapped due to the effect of high-intensity RT [41]. The BBT was used in four studies to assess UE gross motor function [3, 36, 39, 41]. It was found that both treatment arms improved BBT score and the patient was able to move the box to a greater distance. The MAL tool was used in 2 trials [4, 36], and researchers found significant differences from baseline to post-intervention without any significant difference between the groups. Two studies used MAS/w, where one study reported only a slight decrease in spasticity in both groups [3]. In another study, the score of the scale remained unchanged in both groups [41].

The effect of lesion location does not affect the intervention outcome. This was examined in a study [40], where the administration of RT plus tDCS in patients with cortical and subcortical lesions does not influence the clinical improvement. Two studies investigated the combined effect of RT and tDCS in subacute and chronic stroke patients [4, 36]. According to Straudi et al., bilateral tDCS along with RT was more effective and suitable in chronic stroke patients [36]. Conversely, Triccas et al. showed that the percentage improvement of motor impairment was more in subacute stroke patients [4]. This can be explained by the spontaneous recovery in the subacute phase which involves cortical reorganization [42, 43].

Only two studies examined the sustained effect of RT + tDCS and evaluated the participants for follow-up [4, 40]. The treatment effect does persist for a longer period of time, as seen in one study, where significant changes were reported in scores of FMAS-UE between baseline and follow-up at 3 months; however, no significant difference was shown in scores of FMAS-UE between groups receiving RT + tDCS and RT + sham tDCS [4]. Similar findings were reported in another study, where a 6-month follow-up was performed. Significant changes were observed in measures of UE motor function between baseline and 6-month follow-up but no significant mean difference between groups was found [40].The current amplitude used by three studies was 1 mA [4, 36, 39], while the rest of the studies included in this review used 2 mA of current [3, 40, 41]. No change was observed in clinical improvement when studies used different stimulation intensities.

The essential process driving improvement in functional outcomes after a stroke is neuroplasticity [44]. As a result, one of the most essential goals of stroke therapy is to make good use of neuroplasticity for functional recovery. Goal planning, high-intensity practice, multidisciplinary team care, and task-specific training are some of the principles of stroke rehabilitation [43]. As a result, for stroke recovery, high-dose rigorous training and repetitive practice of functional tasks are critical [45]. All these requirements make stroke rehabilitation a labor-intensive process. The tDCS [13, 46] and RT [47] are two promising therapies in rehabilitation used for stroke patients. It is unlikely for tDCS to bring any optimal functional change on its own but it augments other rehabilitation strategies by enhancing brain plasticity. The RT has been proven to improve functional recovery after stroke [26]. Both of these treatments aim to influence brain plasticity; therefore, there is a case to be made for seeing if combining the two could provide an additive effect. However, the optimal approach for combined therapy of tDCS and RT, and whether the outcomes of the combination of these two are cooperative or conflicting, is yet to be determined. The tDCS and RT are synergistic therapy approaches which was evident by an increase in corticospinal excitability resulting from the administration of tDCS followed by RT [40].

This systematic review analyzed the effect of integrated therapies using RCTs, has good methodological qualities, and certifies a significant level of evidence, so the results are considered reliable. However, there are certain limitations of this review. First, this review lacks studies administering tDCS and RT in acute cases of stroke. The second limitation of this study is the number of patients in the included studies was quite small to be analyzed with full potential. Third, a follow-up in  four out of six studies was not done. This could provide us with a sustained effect of desired treatment after completing the proposed sessions.

RT can be used as an adjunct in the management of UE impairment in stroke patients. Future research should involve administering tDCS + RT with more complex tasks with different stimulation time. A large sample size in future studies could help generalize data over the population. Trials on examining carryover effects can also be done.

Conclusion

Motor improvement was observed in the affected UE after taking combined tDCS and RT intervention in both groups. However, there is no added benefit of real direct current stimulation compared to sham, justifying robotic therapy as the main influencer of recovery. A probable reason for this could be the administration of robotic training at high intensity. The sample size was too small in the included trials and the effects of the two treatments were overlapping, making it difficult to detect potential benefits solely from stimulation. The treatment was seen to be more effective in sub-acute cases rather than chronic, and the treatment program used in the studies is safe and well-tolerated by patients, with no or minimal adverse events.