Impact of rTMS and iTBS on Cerebral Hemodynamics and Swallowing in Unilateral Stroke: Insights from fNIRS

Introduction

Any abnormal structural or functional deficits of the swallowing-related muscles and nerves result in swallowing problems, also known as dysphagia [1]. Stroke is one of the most common factors associated with this condition, with 50% to 81% of patients with acute stroke experiencing swallowing problems [2]. Post-stroke dysphagia (PSD) can cause numerous complications, including aspiration pneumonia, dehydration, malnutrition, and asphyxia. These conditions are related to the weakened and abnormal coordination of submental muscle contraction [3], thus reducing the quality of life in patients with stroke [4]. Research shows patients with dysphagia have an 8.5-fold higher risk of death than do those with normal swallowing [5]. Another study from the United States reported that dysphagia after stroke significantly increases medical costs, driven by higher hospital and durable medical equipment expenses [6]. Cabib et al proposed that swallowing function is controlled by a multi-layer neural control model, consisting of the cortical swallowing central network, medullary swallowing central pattern generator, and peripheral nerves [7]. The manifestations of swallowing disorders caused by stroke in different parts of the brain are also different. Cortical stroke mainly affects the processing of oropharyngeal sensory information of reflexive swallowing and voluntary swallowing, affects the execution of swallowing movements, and causes delayed or even inability to start swallowing [8]. These symptoms can be reflected by electrophysiological changes in the swallowing muscles [9]. Suprahyoid muscles are an important muscle group involved in pharyngeal swallowing and are located at the front of the neck. Musculation signals in this region can be sensitively acquired by surface electromyographic (sEMG) [10]. Peak amplitude (PA) and swallowing time (ST) are indicators commonly used in sEMG [11]. A previous study has shown that patients with PSD have significantly reduced PA and prolonged ST than do healthy individuals [12]. This phenomenon may be caused by the reduction or atrophy of muscle strength due to denervation of the suprahyoid muscles in patients with PSD, thus resulting in the delayed initiation of swallowing and prolonged pharyngeal latency [12]. However, its causes and neuropathological physiological mechanisms are still not very clear. The development of neuroimaging has made it possible for us to have a deeper understanding of the relationship between behavior and the cerebral cortex. Functional near-infrared spectroscopy (fNIRS), functional magnetic resonance imaging (fMRI), and diffusion-weighted imaging have been used to study the mechanism of swallowing function and have better spatial resolution [13–15]. Swallowing is the product of a complex interaction between the central nervous system and the musculoskeletal system. Unlike the motor function of the upper limbs, swallowing function is not dominated by constant cortical regions, based on fMRI [16]. fMRI studies of healthy people showed that the sensorimotor area, prefrontal lobe, insula lobe, cingulate gyrus, cerebellum, and occipital lobe are activated to varying degrees during swallowing [17]. In addition, there was a tendency of lateralization in different swallowing stages, the pharyngeal stage activated the right brain, and swallowing in the oral stage activated the left brain [18]; however, this not related to handedness [19]. Seeley et al [20] used fMRI and found that a group of regions, including the bilateral dorsolateral prefrontal cortex, ventrolateral prefrontal cortex (PFC), dorsomedial PFC, and lateral parietal cortices, showed significant activation during the performance of cognitive-related tasks and formed a network connection relationship independently, which was named the “executive-control network”. fMRI studies have also confirmed that the supplementary motor area plays a crucial role in motor planning, preparation, and execution [21]. It is indisputable that the involvement of these brain regions is essential during the swallowing process. Therefore, we should prioritize the functional activation of these regions to better assess the nature and severity of dysphagia.

fNIRS is a noninvasive neuroimaging techniques, similar in principle to fMRI, which can also measure changes in the concentrations of oxyhemoglobin and deoxyhemoglobin in the cerebral cortex associated with neural activity [22]. In 1977, Jöbsis et al [23] introduced near-infrared spectroscopy (NIRS), which demonstrated that brain tissue transparency is high in the near-infrared range (700–900 nm). Furthermore, changes in oxyhemoglobin, deoxyhemoglobin, and total hemoglobin can be detected in real-time and non-invasively with fNIRS [23]. The light source emits near-infrared light into a specific region of the brain through a light-emitting diode or a fiber-optic bundle that is adjusted to the individual’s head size [24]. The light scatters in a banana-shaped path, and a light detector situated at a distance from the beam collects the light that the tissue has scattered back [25]. The blood oxygen level in human tissues varies depending on the physiological condition and activity of the body [26]. Physiological tissues have 2 responses to light: absorption and scattering [27]. As the amount of scattering in various parts of the head cortex does not vary with neural activity, the attenuation caused by scattering in cortical tissues remains constant [27]. Therefore, alterations in attenuation, measured during cognitive activity, are attributed to changes in absorption [27]. The primary constituents of tissues that absorb near-infrared light are water and hemoglobin. In the spectral range of NIRS light from 600 to 900 nm, water has a very low absorption rate, whereas oxyhemoglobin and deoxyhemoglobin have different light sensitivities [28]. Oxyhemoglobin is sensitive to the oxygenated state, while deoxyhemoglobin is sensitive to the deoxygenated state. These 2 compounds have distinguishable light-absorbing properties, which are predominantly manifested through the higher absorption coefficient of oxyhemoglobin in wavelengths larger than 800 nm [28]. Neuronal activity necessitates an energy supply of glucose and oxygen. Consequently, augmented neuronal activity results in concurrent elevation in cerebral oxygen metabolism [28]. In situations in which specific brain regions are active and involved in performing a task, the brain’s need for oxygen and glucose increases [29]. This, in turn, leads to blood oxygen consumption to produce energy, ultimately causing oxyhemoglobin concentration to decline and that of deoxyhemoglobin to increase [30]. From the correlation between light attenuation and chromophore concentration changes in tissues [31], it is possible to use the modified Beer-Lambert law to calculate the changes in relative concentrations of oxyhemoglobin and deoxyhemoglobin in the brain during neural activity. This can allow for speculation on the associated brain regions and their interrelationships [31]. In general, the fNIRS technique showed higher applicability for its use in various environments and activities, including ease of handling, portability, less sensitivity to motion artifacts and use during dynamic movements, and relatively lower cost [32]. However, fNIRS has limited penetration ability and can only reach the cerebral cortex 1 to 2 cm below the skull skin; therefore, it cannot monitor and evaluate the function of the whole brain and is not suitable for the study of neural pathways related to the cortex [33]. However, sEMG could provide electrophysiological characterization information of nerve-activated muscle groups, thereby indirectly making up for the shortcomings of the above methods [34]. In a previous study, Griffin et al combined fNIRS and submental muscle sEMG to investigate differences in the swallowing of 1 mL of water in healthy participants, under different levels of transcranial direct current stimulation [9]. These authors found that the higher the intensity of the stimulus, the more likely they were to upregulate the sensorimotor cortex excitability of swallowing; in addition, submental muscle contraction increased, although this change was not statistically significant. In another study, Hamdy et al investigated the sEMG response arising from the stimulation created by single transcranial magnetic stimulation (TMS) shocks given several seconds apart [35]. These authors showed that it was possible to posit a relatively direct and rapid conducting pathway from the cortex to the muscles via the brainstem. This provided evidence to support the fact that TMS causes an interaction between the central nervous system and the corresponding swallowing muscles. Therefore, it is feasible to combine fNIRS with sEMG to evaluate the mechanism of swallowing function recovery after stroke.

Currently, noninvasive brain stimulation for post-stroke swallowing disorders have divided into 2 types: “top-down” central control and “bottom-up” peripheral control [36]. This is in line with the current mainstream clinical research trend and corresponds to the multi-level neuromodulation mechanism of swallowing function. Central control methods include repetitive transcranial magnetic stimulation (rTMS), transcranial direct current stimulation, and transcranial focused ultrasound. rTMS is a newly introduced neuroregulatory technique that can induce changes in cortical excitability, promote neuroplastic changes, control neurotransmitter release, and treat dysphagia by regulating neuroplasticity [37]. During rTMS, electricity is passed through a coil placed on the scalp, creating a magnetic field that penetrates the skull up to 1.5 to 2 cm deep and provides a sufficiently strong electric field to depolarize surface axons and activate cortical neurons [38]. In addition, upon stimulation of the target musculature, an electromyographic response is produced, which is called a motor-evoked potential. Subsequently, corticospinal transmission from the cortex to peripheral muscles is stimulated to adjust the cerebral cortex’s excitability [39]. There are 2 types of rTMS: high-frequency rTMS (≥1 Hz, HF-rTMS) and low-frequency rTMS (≤1 Hz, LF-rTMS) processes [40]. The cerebral cortex becomes more excitable with high frequency, while it becomes less excitable with low frequency [41]. The mechanism by which TMS promotes the recovery of swallowing function after stroke is currently based on the following 2 neural regulation models. First, the interhemispheric inhibition model posits that in unilateral cortical stimulation schemes, the recovery of swallowing function can rely on downregulating the excitability of the contralateral side or activating the ipsilateral cerebral hemisphere to weaken interhemispheric inhibition [42]. Therefore, HF-rTMS produces a long-term enhancement effect on the ipsilateral hemisphere, thereby bringing about rebalancing [43]. Second, based on the compensation model, some scholars have noticed that the recovery of swallowing dysfunction after stroke can be related to the compensatory reorganization of the ipsilateral hemisphere [44], and stimulating the ipsilateral hemisphere with HF-rTMS can promote the occurrence of such compensation. Further studies have found that using HF-rTMS to stimulate bilateral cortex in patients with unilateral PSD can produce better compensatory effects on overall swallowing function. This may be related to the fact that stimulating both cerebral hemispheres helps promote the recovery of the damaged swallowing network [45]. In addition, different TMS stimulation modes are another research direction. Theta burst stimulation (TBS) is a newer form of TMS. Due to the different frequency of electromagnetic stimulation, TBS can be divided into excitatory intermittent TBS (iTBS) and inhibitory continuous TBS [46]. During iTBS, patterned pulses (theta bursts) are delivered at a high frequency, with intermittent pauses occurring between pulse sets [47]. Although both traditional rTMS and iTBS are effective at inducing long-term potentiation, the iTBS pattern enables a shorter administration time (approximately 3–6 min as opposed to 30–38 min for traditional rTMS), allowing for more pulses to be delivered daily and reducing the number of treatment sessions needed [48]. Research has confirmed that, when comparing iTBS with traditional rTMS protocols with regard to treating major depressive disorder indicate, the 2 methods yield similar positive outcomes [49–51], with iTBS notably more efficient and cost effective.

Some studies have found that HF-TMS and iTBS act on the contralateral motor cortex of submental muscles and could improve swallowing function in PSD [52–55]. Studies also hypothesize that the primary sensorimotor cortex, basal ganglia, anterior cingulate gyrus, thalamus, insula, and inner capsule serve as potential targets for stimulation, playing a critical role in improving swallowing function [56,57]; however, it is not yet known which is most effective. However, research on how TMS alters communication and connections between neurons in the cortex remains largely unexplored [58]. Clinical research on iTBS and rTMS mainly focuses on psychological and cognitive disorders, while there are fewer studies on dysphagia. It is still unclear which TMS mode is better for promoting the recovery of swallowing function. Therefore, the main objective of this study was to compare the effects of rTMS with that of iTBS on functional cerebral cortical activity measured by fNIRS in 20 patients with PSD.

Material and Methods

ETHICS APPROVAL AND STUDY DESIGN:

This was a randomized controlled study conducted in the Affiliated Hospital of Guizhou Medical University, Guiyang, Guizhou, China. The study was approved by the

Ethics Committee of this hospital (registration No. 366, 2022 LUNSHEN) and was registered in the Chinese Clinical Trial Registry (registration No. ChiCTR2200065153, https://www.chictr.org.cn/bin/project/edit?pid=184263). A total of 25 patients in this hospital between November 2022 to March 2023 were selected as the study participants. Written informed consent was obtained from each patient before the study began. All assessment and interventions were provided by a same speech therapist, to standardize the study. The speech therapist had over 10 years of clinical experience in the field of swallowing therapy and postgraduate competence in the field of swallowing rehabilitation.

PARTICIPANTS:

Patients were selected according to the following criteria: (1) patients with stroke diagnosed according to the updated definition for the 21st century [59] and confirmed to have had unilateral stroke by computed tomography (CT) or magnetic resonance imaging (MRI); (2) patients with swallowing disorder >grade 3 (food entered the airway and remained above the vocal cords without being cleared out of the airway), which was assessed by the penetration-aspiration scale (PAS) via a video fluoroscopic swallow study; (3) patients with stable vital signs whose disease duration was less than 6 months and age range was from 30 to 75 years; (4) patients who did not take medicine that could affect swallowing function; and (5) patients were right-handed as assessed by the Edinburgh handedness inventory [60]. Patients were excluded for the following reasons: (1) a history of any other neurogenic diseases, such as epilepsy and tumors; (2) patients with severe cognitive impairment or aphasia who could not continue to cooperate with treatment; (3) patients who had intracranial metal implants, pacemakers, skull defects, a history of seizures, or other contraindications to rTMS treatment; and (4) patients who had a history of sedation, antidepressants, or other drugs that could alter the excitability of the cortex. Twenty-five patients met screening criteria and underwent intervention. In the end, 5 of the 25 patients (expulsion rate of 20%) dropped out of the study for the following reasons: 1 patient in the rTMS group dropped out with headache after several interventions, 2 patients dropped out of the following intervention due to communication disorders and loss of contact, and 3 of 13 those patients dropped out (expulsion rate 23.08%). In the iTBS group, 1 patient had a change in their condition and was removed from the study, 1 patient refused to participate in the follow-up assessment, and 2 of 12 patients dropped out (expulsion rate 16.67%). The participant flow diagram is shown in Figure 1.

SAMPLE SIZE:

Based on our preliminary experimental results, the PAS score results were used as the primary outcome indicator. The mean PAS scores of the 2 groups of patients were 5.17 and 4.83, respectively, and the standard deviations were 1.61 and 1.31, respectively. Further, α=0.05 was used as the significant test standard, and 0.9 was used as the test power. According to the following sample size calculation formula, it was estimated that the sample size of the 2 groups of patients would be approximately 8.14. Assuming that 20% of the patients in each group dropped out, it was finally determined that at least 10 patients were required for each group.

RANDOMIZATION AND BLINDING METHOD:

A simple random number table was generated by computer, and the patients were randomly divided into the rTMS group and iTBS group at a ratio of 1: 1 by simple randomization method. The randomization sequence was sealed and preserved by a researcher who was not involved in the training and evaluation. Speech therapists who were unaware of the participants’ group assignment were assigned to provide swallowing rehabilitation to the individuals, according to the different treatment protocols given by the study leader. Also, the participants were blind to their specific group assignment.

INTERVENTIONS:

All patients received the same traditional swallowing rehabilitation, for a total of 12 treatment sessions over a period of 2 weeks, 6 times a week, and 30 min each time. The rTMS group and iTBS group received the rTMS program or iTBS program after traditional swallowing rehabilitation, respectively.

TRADITIONAL SWALLOWING REHABILITATION: The rTMS and iTBS groups both received traditional swallowing rehabilitation, which included oropharyngeal muscle strengthening exercises, sensory stimulation, and tongue retraction exercises. For 10 min, oropharyngeal muscle strengthening exercises focused on muscle strength, and resistance or skill training, or both, such as the following tongue exercises. Patients were asked to hold their tongue gently between their teeth and then swallow their saliva with their tongue in a straight position [61]. For 10 min, effortful swallow required that the patients squeeze their throat muscles as hard as they could, and then swallow [62]. Other methods included the Mendelsohn maneuver and Shaker exercise [63]. Oral taste training, temperature training, and proprioception stimulation were performed using food additives, ice cubes, and air pulses, respectively, for 10 min [64].

DETERMINATION OF THE RESTING MOTION THRESHOLD: The resting motion threshold (RMT) was determined for patients in the rTMS group and iTBS group. A comfortable chair was positioned in a comfortable position for patients to sit on, and the patients wore matching positioning caps. To prevent positional changes, we requested that patients avoided moving their heads. In this experiment, we connected an external magnetic super-rapid stimulator to a hand-held figure-of-eight coil (outer diameter, 70 mm, Yiruide CCY-IA, Wuhan, China) over the cortical region of the suprahyoid on the unaffected hemisphere of the scalp, which is located 2 to 4 cm in front of the skull apex and then moves 4 to 6 cm toward the contralateral hemisphere [65]. Single pulse outputs were manually controlled to obtain motor-evoked potential and RMT on each patient. The location of the maximum motor-evoked potential recording produced in the cortical region of the suprahyoid muscles as determined as the optional coil position (hot spot). The RMT was defined as a stimulus intensity that evoked an motor-evoked potential >50 UV in at least 5 of 10 successive stimulations across the first dorsal interosseous muscle on the contralateral side [65].

RTMS PROGRAM OR ITBS PROGRAM: The rTMS and iTBS programs used in our study were based on the safety guidelines for TMS applications [66], setting of treatment parameters based on previous literature [53]. In both groups, each patient was instructed to sit on a chair or lie on a bed with a comfortable position, and they wore matching positioning caps. To prevent positional changes, we requested that patients avoid moving their heads. We connected an external magnetic super-rapid stimulator to a hand-held figure-of-eight coil (outer diameter, 70 mm, Yiruide CCY-IA, Wuhan, China) over the hot spot to each patient. A 10-Hz rTMS model was applied to the cortical region of the suprahyoid muscles in the rTMS group. Other stimulus parameters included intensity: 100% RMT; stimulation time: 1 s; stimulation interval: 9 s; and treatment duration: 10 min, with a total of 600 pulses. The iTBS group received 3 pulses of 50-Hz bursts of repeated stimulation at 5 Hz (2 s on and 8 s off) for a total of 3 min 12 s (600 pulses). The RMT intensity was 100% for the iTBS model.

OUTCOME MEASUREMENTS:

Outcomes were measured at baseline and 2 weeks after treatment in all participants. The primary outcome was the PAS scale, using sEMG to determine the PA and ST of the bilateral suprahyoid muscles. Additionally, fNIRS was used to monitor changes in cerebral hemodynamics during task performance.

THE PAS: The PAS was performed using a remote-controlled, twin-bed PLD5000 X-ray machine (Perlead, Zhuhai, China). The video examination area was from the top of the nasal cavity down to the seventh cervical vertebra. A thin bolus (viscosity 5 to 10 mPa/s) was mixed with iohexol solution and 50 mL pure water, and the viscosity value was measured with an NDJ-5S Rotational Viscometer (Shanghai Fangrui, Shanghai, China) at a temperature of 25°C. All patients were examined in both a 90-degree sitting and a lateral position. The evaluator utilized a spoon to extract 5 mL of a thin bolus, administering it into the patient’s oral cavity and instructing them to swallow in accordance with their scheduled swallowing routine, according to the evaluation criteria for the PAS [67]. The PAS was composed of 8 grades, with lower scores indicating better swallowing ability [67]. Grade 1 represented no penetration or aspiration; grade 2: the bolus enters the airway, remains above the vocal folds, and is ejected from the airway; grade 3: the bolus enters the airway, remains above the vocal folds, and is not ejected from the airway; grade 4: the bolus enters the airway, contacts the vocal folds, and is ejected from the airway; grade 5: the bolus enters the airway, contacts the vocal folds, and is not ejected from the airway; grade 6: the bolus enters the airway, passes below the vocal folds and is ejected into the larynx or out of the airway; grade 7: the bolus enters the airway, passes below the vocal folds, and is not ejected from the trachea despite effort; and grade 8: the bolus enters the airway, passes below the vocal folds, and no effort is made to eject.

SEMG: We used the MegaWin3.0 data acquisition system (ME6000 sEMG System, Mega Electronics Ltd, Kuopio, Finland) to record sEMG signals (electrical activity) from the bilateral suprahyoid muscles, The technical parameters of the equipment include the measurement range of ±8192 UV, accuracy of ±2% when input <4000 UV, accuracy of ±5% when input < 8192 UV, sampling rate of 1000 Hz, sensitivity of 1 μV, resolution of 1 μV, and input impedance of 20 GΩ. Two pairs of disposable Ag/AgCl electrodes with a diameter of 1 cm were used for the sEMG acquisition electrodes, which were placed along the posterior one-third of the suprahyoid muscles at midline and were 22±1 mm apart from one another; the reference electrode was attached to the ipsilateral mastoid [68]. Before placing the electrodes over the targeted muscle fibers, excessive hair was removed by shaving, and the skin was cleaned with an alcohol preparation pad. After the signal quality was confirmed, each patient was instructed to swallow saliva and repeat this procedure 5 times. The following data of the suprahyoid muscles were analyzed with the MegaWin3.0 data acquisition system: ST: the interval from the onset of swallowing (defined as an increase in sEMG value exceeding the mean baseline value by more than 2 standard deviations within 20 ms) to the offset of swallowing (defined as a decrease in sEMG value below the mean baseline value by more than 2 standard deviations within 20ms), whose unit of measurement was seconds (s); PA: the maximum value of the sEMG signal during the above interval, whose unit of measurement was microvolts (μV) [69]. The mean values of the ST and PA of the bilateral suprahyoid muscles were then determined.

FNIRS DATA ACQUISITION AND SWALLOWING TASK: During voluntary resting and swallowing, we measured changes in oxyhemoglobin and deoxyhemoglobin using fNIRS devices at 730 and 850 nm (NirScan Danyang Huichuang Medical Equipment Co Ltd, Danyang, China). The emitted signals were received by adjacent detectors 30 mm apart and collected at a sampling rate of 10 Hz. Multiple cortical areas are involved in swallowing activity. The 14 source and 14 detector probes formed 35 channels based on the 10–20 system, targeting the following region of interest (ROI) of bilateral cerebral hemisphere: PFC, primary motor cortex (M1), and primary sensory cortex (S1). Figure 2 shows the fNIRS probe configuration. Optodes were placed according to 4 reference points (nasion, central zero, left, and right preauricular points) using the 10–20 system. Spatial registration was done using a 3D localization system, to match probe positions to the Broadmann areas. The patients were in a quiet fNIRS evaluation room to minimize noise interference. Each optical signal was secured with a custom hard plastic cap and covered with a black cloth to prevent ambient light interference. The swallowing procedure blocking included 30-s resting and 15-s swallowing periods, repeated 5 times by a computer voice. The completion of 1 swallowing action was indicated by the observation of a full laryngeal elevation. To minimize oral movement artifacts, participants rehearsed the swallowing motion before the study. The entire procedure took approximately 10 min (Figure 3).

FNIRS DATA ANALYSIS: fNIRS data were automatically preprocessed using the NirSpark software (NirScan Danyang Huichuang Medical), which runs in MATLAB (Mathworks, USA). Previous data processing in the literature [70,71] was referenced, according to the steps for fNIRS data analysis, as follows: setting the signal standard deviation threshold to 6 and the peak threshold to 0.5; removing motion artifacts via spline interpolation; filtering physiological noise (heartbeat, respiration, Mayer waves) with a 0.01–0.1 Hz band-pass filter; converting filtered optical density data to oxyhemoglobin concentration using the modified Beer-Lambert law [28]; and setting the initial time of the hemodynamic response function to 0 s and the end time to 15 s (the time for a single blocking paradigm). Oxyhemoglobin concentrations from the 5 blocking paradigms were superimposed and averaged, yielding a 15-s swallowing period blocking average. The blocking differential was calculated as the average oxyhemoglobin concentration during swallowing minus the rest block oxyhemoglobin concentration. We analyzed the pretreated oxyhemoglobin time-series data for each channel and patient, using a generalized linear model [72], which calculates the match between experimental and ideal hemodynamic response function values [73]. The oxyhemoglobin beta values (OBV), representing cortical activation, were used to estimate the hemodynamic response function prediction of the oxyhemoglobin signal [30]. The fNIRS cortical maps were used to analyze the OBV in each ROI with the 10–20 system. A 2-sample t test was used to assess differences between the rTMS group and iTBS group at baseline and at the conclusion of the study.

STATISTICAL ANALYSIS:

All statistical analyses were conducted with SPSS version 25.0 (IBM Corp, Armonk, NY, USA). Shapiro-Wilk tests were conducted to determine the normality of the data distribution. Continuous variables are presented as mean ± standard deviation (SD), otherwise, represented by median (M) and interquartile range (Q), and categorical variables were represented by counting ratio. Two-sample t tests were adopted to analyze data (age, PAS, ST, and PA) with normal distribution, and the non-parametric Mann-Whitney U test was used for analysis of non-normally distributed data (sex, type of stroke, affected hemisphere, and duration of onset of stroke). Two-sample t tests were used to compare all outcomes (PAS, ST, PA, and OBV in each ROI) between groups before and after intervention. To find the differences between the rTMS and iTBS groups after intervention, paired t tests were used to identify the OBV in each ROI and other outcome data. A 2-sided P value <0.05 was used as the level of significance.

Results

BASELINE CHARACTERISTICS OF PARTICIPANTS:

Table 1 shows the baseline demographic characteristics of the 2 groups of patients. Intergroup comparisons of baseline characteristics did not reveal any significant differences (P>0.05, Table 1). Function assessment at baseline found no significant differences in PAS, ST, and PA between the 2 groups (P>0.05, Table 1).

PAS AND SEMG:

There was no significant difference in the PAS scores, ST, and PA after intervention between the rTMS group and iTBS group (P>0.05, Table 2). By contrast, in comparison with the baseline data, the 2 groups exhibited significant improvement in PAS scores, ST, and PA after 2 weeks of treatment (all P<0.001, Table 2). The results of the PAS showed that the penetration and aspiration during swallowing were improved in the 2 groups after treatment, while the ST was shortened, and the PA was significantly increased, compared with that before treatment.

CORTICAL ACTIVATION ANALYSIS OF FNIRS MEASUREMENTS:

After treatment, a comparison with baseline data revealed no significant difference of OBV in each affected ROI, including affected PFC, affected M1, and affected S1, respectively (all P>0.05, Figures 4–6), in the 2 groups. Nevertheless, in the iTBS group and in the rTMS group, the OBV showed significant increase in each unaffected ROI, including the unaffected PFC (U-PFC) (P<0.001, Figure 7), unaffected M1 (U-M1), and unaffected S1 (U-S1), respectively (all P<0.05, Figures 8, 9). The fNIRS cortical maps showed the OBV of the U-PFC region was significantly higher than that of the U-S1 or U-M1 regions (P<0.05, Figure 4) after treatment. After comparing the post-treatment results, we also found no significant differences between the 2 groups in all affected ROIs (all P>0.05, Figures 4–6) or the unaffected ROIs (P>0.05, Figures 7–10). The above results indicated that both the rTMS and iTBS treatments targeting the motor projection area of suprahyoid muscles on the ipsilateral side resulted in significant increases in OBV in the unaffected hemisphere cortex, suggesting significant improvements in activation levels in corresponding brain regions. This phenomenon was observed in all unaffected ROIs, particularly in the PFC region. However, no significant activation was observed in any of the ROIs in the affected hemisphere.

Discussion

STUDY LIMITATIONS:

There are some limitations to this study that need to be considered. First, the patients were all from the same hospital, and the sample size was small; therefore, future studies need to expand the sample size and be multi-centered, in order to reduce random errors and selective bias, improve the statistical power of the research, and make the research results more reliable and convincing. Second, the observation and follow-up time was short in both groups, and the lack of long-term effects from iTBS or rTMS treatment could not be better illustrated without the advantages of long-term enhancement effects. Third, this study did not further analyze the treatment effect divided into different brain injury position groups and discuss the differences in cortical activation. Hence, further studies must analyze these potential influencing factors with larger sample sizes to demonstrate the potential benefits of fNIRS and noninvasive brain stimulation intervention for patients with PSD and use more accurate experimental equipment to reduce experimental error and bias.

Conclusions

The present results showed that TMS and intermittent theta burst stimulation on cerebral cortical activity have similar effects in improving swallowing function of patients with PSD. fNIRS revealed that the improving of swallowing function depended on compensatory mechanisms in the U-PFC, U-S1, and U-M1, with the U-PFC region being particularly important. Our findings provide new supporting evidence for the compensatory model theory of brain neural regulation. Although we did not identify significant differences in terms of assessment measures between the rTMS and iTBS groups, iTBS may have more potential to increase capacity, improve utilization rate, and reduce waiting times than does rTMS.