From December 2022 to October 2023, 57 patients were qualified at Shanghai Health Rehabilitation Hospital (Shanghai, China). 48 patients (including 18 women, representing 31.6% of the total sample) with a mean age of 67 years (Standard Deviation (SD) = 10.6) were ultimately recruited for participation in the study (Table 1). The inclusion criteria were: (1) age of patient, 60–75 years; (2) right-handed; (3) fulfilled the diagnostic criteria for stroke and cognitive impairment [29, 30, 31]; (4) a clear state of consciousness and stable vital signs; (5) at least one limb should function without any impediments; (6) with a cognitive impairment appearing after stroke, with a scoring range on the MoCA test of 9–26 points. The exclusion criteria were: (1) cognitive impairment stemming from either primary or secondary neurological disorder; (2) a history of epileptic seizures; (3) severe general impairment or concurrent medical conditions; (4) severe aphasia or cognitive impairment that might hinder the provision of informed consent or impede the behavioral measurements conducted during the study; (5) any contraindications to Transcranial Magnetic Stimulation (TMS) [32].

Ethical approval was obtained from the Medical Ethics Committee of Shanghai University of Medicine & Health Sciences on October 27, 2022 (2022-E4-6100-21-201034-03-210302197009090947). The Ethics Committee of Chinese Registered Clinical Trials has reviewed this study and has granted its approval. The registration number is ChiCTR220066184. Each participant provided written informed consent.

The flow chart of the procedures is illustrated in Fig. 1.

Fig. 1.

The CONSORT flowchart visually outlines the recruitment, group allocation, treatment allocation, follow-up procedures, and analytical methods in clinical trials. rTMS, repetitive transcranil magnetic stimulation.

Fig. 2 shows the experimental design. After recruiting PSCI patients, they were randomly assigned to one of three groups for 20 treatments (4 weeks, 5 consecutive days, and 2 days off per week). All patients received routine rehabilitation treatment (40 min/day, 5 days/week, for 4 weeks), with one-on-one physical and occupational therapy provided by qualified rehabilitation therapists. The experimental rTMS groups formed were (a) the dual-target group for real L-DLPFC + real M1 stimulation; (b) single-target group for real L-DLPFC + sham M1 stimulation; and (c) sham-stimulation group for sham L-DLPFC + sham M1 stimulation. Outcome measures were obtained twice: initially for a baseline prior to the intervention, and after 4 weeks of rehabilitation and rTMS intervention.

Fig. 2.

Experimental design. BDNF, brain-derived neurotrophic factor; VEGF, vascular endothelial growth factor; DLPFC, dorsolateral prefrontal cortex; M1, motor cortex.

The general methods of allocation and blinding used previously [32]. The detailed methods were as follows. An investigator (XJW) used a computer-generated random-number table to randomly allocate patients to the three groups. The allocation information was disclosed to rTMS physical therapists and doctors (YL and YWW), via phone call by statisticians who conducted the rTMS stimulation sessions on patients (XJW). The physical therapists involved in rTMS were not engaged in any other research-related tasks. The researchers responsible for evaluating, collecting, and processing outcome data were blinded to the allocation of treatments. Treatment assignments were also concealed from patients [33].

The rTMS treatment uses a magnetic stimulator with an 8-shaped coil (MagProX100, Tonica, Farum, Denmark). The DLPFC region was targeted for stimulation using the “standard 5 cm” method, which involves transitioning 5 cm forward from the left M1 region. The stimulus intensity of each group was 80% of the resting motor threshold. The resting motor threshold was the minimum stimulating power necessary to elicit an overt motor response (motor evoked potential, significantly greater than 50 µV) in the contralateral abductor pollicis brevis, on at least 5 of 10 trials [34]. The dual-site stimulation was performed sequentially. The L-DLPFC is stimulated first, then the ipsilesional M1. During the actual procedure, we used a new and more rigorous method for sham stimulation. The coil was positioned at the identical scalp location and operated with identical parameters as those used for the dual-target group [32]. The handle was oriented downwards, and the coil was tilted by 90° in the axial plane [9]. The edge of the coil was kept in contact with the scalp to produce audible clicking noises [35, 36, 37]. This approach ensured that the sham-stimulation procedure closely resembled the real stimulation, but without producing significant neural effects. Fig. 3 shows rTMS intervention details. Stimulation parameters: L-DLPFC: 10 Hz, 80% resting motor threshold (RMT), 2000 pulses (5 s $\times$ 40 trains, 25 s interval); Ipsilesional M1: 10 Hz, 80% RMT, 1200 pulses (2 s $\times$ 60 trains, 8 s interval).

Fig. 3.

The details of repetitive transcranil magnetic stimulation (rTMS) intervention. RMT, resting motor threshold; L-DLPFC, left dorsolateral prefrontal cortex.

All assessments were done twice, pre-treatment (baseline) and post-treatment. The main evaluation indicator were MoCA scores. The version we chose is Montreal Cognitive Assessment Beijing Version (MoCA-BJ), a tool that evaluates the overall cognitive function [38, 39]. The secondary results included the evaluation of the basic activities of daily life by the Modified Barthel Index (MBI) [40], visual attention, executive function with the Trail-making Test (TMT) [41, 42] and attentional function with the Digital Span Test (DST) [43]. Additionally, serum samples were collected to assess the levels of BDNF and VEGF. At baseline, blood specimens were collected before rTMS stimulation. After 4 weeks, blood specimens were collected the day after completing rTMS stimulation. For the blood draw, the patient started fasting at 18.00; 5 mL of peripheral blood were collected the next morning [44]. The blood specimen was centrifuged, and serum was extracted and then stored at –80 °C. The quantification of BDNF and VEGF serum levels was accomplished with enzyme-linked immunosorbent assay (ELISA) kits (Multi-Sciences, Hangzhou, Zhejiang, China). Strict adherence to the provided instructions was observed throughout the procedure. A single technician, blind to the treatment status of the patients, conducted all measurements. Each sample was duplicated and the results were averaged to ensure precision and consistency.

Statistical analysis was performed using SPSS Software 25 (IBM-SPSS Statistics, Chicago, IL, USA) [45]. A paired t-test or Wilcoxon rank-sum test was used for intragroup comparison. Intergroup comparision was done by one-way analysis of variance (ANOVA) or Kruskal-Wallis H test, then a post-hoc analysis was conducted. If there was a difference at baseline, such as in MoCA scores or DST Inverted sequence scores, pre-treatment data were used in a covariate analysis to eliminate the effect of baseline differences. The level of statistical significance was set at p 0.05.

Sixty right-handed patients underwent qualification screening; 3 did not meet the selection criteria. Fifty-seven patients were randomly assigned to three treatment groups, with n = 19/gp. Nine patients did not complete the research project. Forty-eight, therefore, were included in the analysis as shown in the patient flow chart (Fig. 1). Table 1 lists the demographic and clinical characteristics of the sample.

Table 2 shows that after adjusting for Pre-intervention MoCA scores, the post-Intervention MoCA scores of different groups varied after intervention (F = 4.645, p = 0.015). Compared with the baseline, the MoCA scores of all groups were higher at 4 weeks. Post-hoc analysis revealed that the dual-target rTMS group score was significantly higher than the sham rTMS group score (p = 0.021), and the dual-target rTMS group score was significantly higher than the single-target rTMS group score (p = 0.008), upon the completion of the treatment protocol (Fig. 4).

Fig. 4.

Comparison of Montreal Cognitive Assessment (MoCA) scores before and after treatment for the three groups.

Table 3 shows the MBI, TMT and DST scores after intervention. Compared with the baseline, the MBI, TMT, and MoCA scores of the dual-target rTMS group and single-target rTMS group improved after 4 weeks of rTMS treatment and cognitive therapy.

The comparison of baseline BDNF and VEGF levels in patients in three groups showed no significant differences (p $>$ 0.05). After the treatment however, the BDNF levels of dual-target (p = 0.032) and single-target groups (p = 0.034) were higher than their baselines; the sham rTMS group values did not increase above baseline (p = 0.107) (Fig. 5). The VEGF levels of the dual-target group after treatment were significantly higher than baseline (p = 0.008), and the dual-target rTMS group had higher levels than sham-rTMS group (p = 0.003) (Fig. 6).

Fig. 5.

Comparison of BDNF levels of the three groups before and after treatment. ^#p 0.05 for intragroup comparison.

Fig. 6.

Comparison of VEGF levels of the three groups before and after treatment. ^#⁢#p 0.01for intragroup comparison; ^**p 0.01 for intergroup comparison.

Our research has established the practicality of using a 20-session treatment regimen of 10-Hz rTMS, targeted at dual regions (L-DLPFC and ipsilesional M1), in conjunction with traditional rehabilitation therapy, for the effective treatment of patients with PSCI. Overall cognitive function was significantly enhanced by the cumulative impact of bilateral high-frequency rTMS, surpassing both the single-target rTMS and sham rTMS in efficacy. Dual-target 10-Hz rTMS, when combined with traditional rehabilitation methods, is anticipated to yield more favorable outcomes in the management of PSCI patients.

The improvement effect on specific cognitive domains is not obvious. The findings of research examining the impact of rTMS on executive function remain inconclusive. Certain studies have reported favorable outcomes associated with the use of single-target rTMS. For instance, Wang Lu et al. [7] reported that the combination of high-frequency rTMS with cognitive training was found to effectively enhance executive function among stroke patients. Kim et al. [8] observed that high-frequency rTMS applied to ipsilesional DLPFC resulted in significant improvement in cognitive function among stroke patients. They also observed an increase in both forward and reverse DST scores, suggesting that stimulation of DLPFC with 10-Hz rTMS had a definite therapeutic effect on post-stroke execution impairment [8]. The present study showed that there was significant improvement in TMT-A time and DST inverted sequence in the dual-target rTMS group after treatment.

Cognitive function is closely related to a patient’s ability to function in daily life. Cognitive impairment slows down the processing speed of the central nervous system when patients face external information, which is manifested as decreased executive ability, weakened reaction ability, and poor self-care ability [46, 47]. Our findings show that all three research groups improved over their pre-treatment baselines (p 0.05). However, no significant intergroup differences were noted in MBI scores. Li et al. [48] found that high-frequency single-target rTMS of the DLPFC can significantly improve cognitive function and activities of daily living (ADL), and these modifications are typically accompanied by the activation of local frontal regions and alterations in connectivity within the frontal network [49]. Our results may be related to the generally poor limb mobility of the subjects we included. Compared with other research results, our subjects had relatively low ADL scores before treatment, and the subjects were older and sicker. In a plateau period of the patient’s limb recovery, the contrast between the different intervention measures is not very obvious. Similarly, in some studies, the ADL performance has been observed to remain unchanged [50].

There are several candidate mechanisms by which rTMS can be beneficial in PSCI therapy: (a) rTMS can exert significant influence on various neurobiological processes, including the modulation of metabolic activity in both stimulated and remote brain regions; (b) by the alteration of oxidative stress levels; (c) by activation of neurotransmitter systems; (d) by enhancing the plasticity of synapses; (e) through long-term potentiation; and (f) through the connectivity and functionality of neuronal networks [51, 52, 53, 54]. Additionally, according to recent reports, 20-Hz rTMS has been demonstrated to elevate regional cerebral blood flow in the left DLPFC, in addition to other regions such as the cingulate gyrus, basal ganglia, hippocampus, amygdala, and insula [54, 55]. In our research, an increase in serum BDNF and VEGF suggests that curative strategies may be beneficial through increasing synaptic plasticity and promoting the transmission function between synapses, which facilitates learning and the release of neurotrophic factors.

At present, research on the utilisation of rTMS in the management of PSCI is still in the exploratory stage, and there is little research in China on the application of dual-target rTMS. In the present experiment, a comparison was made between sham-rTMS stimulation, single-target rTMS, and dual-target rTMS. Based on the findings of this study, the potential exists to devise an innovative bilateral rTMS approach, which can serve as a therapeutic tool for neurorehabilitation by enhancing cognitive function among stroke survivors.

The current investigation is subject to several constraints. First, it will be necessary in the future to conduct research on a larger sample size of patients to ascertain the precise impact of bilateral rTMS on various stroke subtypes. Second, there was a lack of post-treatment follow-up. Third, there may be subjective errors in the cognitive assessment scales used to assess the effectiveness of treating cognitive impairment. In the future, we will apply techniques such as electroencephalography, positron emission tomography (PET), and functional magnetic resonance imaging (fMRI) to evaluate the degree of cognitive impairment and further explore the mechanism by which rTMS enhances cognitive function in individuals with cognitive impairment. Fourth, the degree of cognitive impairment of the patients included in this study was not limited to mild or moderate. Doing so would have avoided the cognitive differences observed in the patients’ baseline scores. Fifth, the method of sham-stimulation in our study may not have provided a realistic-enough sensation. Recently we learned that a coil has been especially devised for false stimulation that can make similar sounds, but has no actual stimulation effect. The sham-stimulation method should be more realistic in future studies.