Stroke is defined as an episode of acute neurological impairment caused by an ischemic infarction or hemorrhage resulting in focal injury to the brain (Sacco et al., 2013). With an estimated ≈ 13 million cases reported annually, stroke is considered one of the major causes of neurological disability worldwide (Saini et al., 2021). Despite advances in stroke management, many stroke survivors suffer from long-term residual effects (Chohan et al., 2019).

Transcranial magnetic stimulation (TMS) has emerged as a painless, non-invasive technique that can stimulate the human brain and aid stroke survivors in rehabilitation after stroke (Hallett, 2007). Therapeutically, TMS is delivered in a repetitive manner with various stimulation parameters in terms of intensity, frequency, and number of sessions. The mechanism of action of repetitive TMS (rTMS) is believed to rely on principles of long-term potentiation/depression (LTP/LTD)- synaptic plasticity that has been extensively studied at the cellular level (Huang et al., 2017). Excitatory protocols consist of high-frequency (5–20 Hz) rTMS and ultra-high frequency (50 Hz) rTMS applied in the form of a patterned protocol called intermittent theta-burst stimulation (iTBS), while inhibitory protocols consist of low-frequency (≤ 1 Hz) rTMS and continuous TBS (cTBS) (Hallett, 2007). As stroke survivors frequently encounter motor impairments with slow recovery, rTMS stimulation of the primary motor cortex (M1) has been widely investigated to promote the motor rehabilitation process with encouraging results.

However, M1 stimulation has shown therapeutic efficacy beyond the conventional motor symptoms to involve dysphagia, speech impairments, central post-stroke pain, depression, and cognitive dysfunction. This mini-review aims to summarize the latest knowledge about M1 stimulation in post-stroke rehabilitation and discuss the mechanistic rationale behind the management of motor and non-motor symptoms.

The motor impairment after stroke was shown to result from the damaged corticospinal output from the affected M1 along with transcallosal inhibitory drive from the unaffected M1 (Ward and Cohen, 2004). The latter effect is based on the concept of dysbalanced interhemispheric interaction where the unaffected (contralesional) hemisphere becomes “overactive” and slows the recovery process of the affected (ipsilesional) hemisphere (Ward and Cohen, 2004). Therefore, excitatory (high-frequency) rTMS protocols are usually applied over the ipsilesional hemisphere, whereas inhibitory (low-frequency) rTMS protocols are applied over the contralesional hemisphere to restore the interhemispheric balance. A recent meta-analysis revealed that the high-frequency rTMS over the ipsilesional M1 enhanced its excitability without affecting that of the contralesional M1. On the other hand, the low-frequency rTMS protocol not only decreased the contralesional M1 excitability but also enhanced that of the ipsilesional M1, further supporting the bimodal balance recovery model (Bai et al., 2022).

Dysphagia is a common complication in stroke patients and could result in aspiration pneumonia, malnutrition, and dehydration (Takizawa et al., 2016). Different neuromodulation techniques have been investigated to promote swallowing recovery after stroke, including transcranial direct current stimulation, surface neuromuscular electrical stimulation, and pharyngeal electrical stimulation. However, rTMS stimulation was superior to these techniques in the swallowing recovery based on the results of two recent network meta-analyses (Chiang et al., 2019; Li et al., 2021). The mechanism of action of M1-rTMS on dysphagia rehabilitation depends on enhancing neuroplasticity of the disrupted corticobulbar neural pathways that project to swallowing muscles (Gow et al., 2004).

Stimulation targets of the M1 have varied between the tongue, mylohyoid, pharyngeal, and esophageal representations at the M1 (Yang et al., 2021). In addition, no significant difference in the swallowing recovery was noted between low- and high-frequency rTMS protocols (Yang et al., 2021). Nonetheless, bihemispheric M1-rTMS has shown better effects on the swallowing function compared to unilateral stimulation (Cheng et al., 2020). In addition, it’s recommended to combine rTMS with traditional swallowing rehabilitation training to achieve better outcomes (Dziewas et al., 2021).

Concerning the optimal time window, the effect of therapeutic rTMS was most beneficial when applied within the first 2 weeks of stroke, and the effects were most substantial during the first 2 months following application (Cheng et al., 2020).

Based on the cortical region affected, stroke can result in impaired speech production in the form of aphasia or dysarthria (Schindel et al., 2022). As a therapeutic approach, rTMS was mainly applied at the inferior frontal gyrus corresponding to Broca’s area in aphasia rehabilitation (Arheix-Parras et al., 2021). Recently, however, a concurrent use of iTBS and functional magnetic resonance imaging (fMRI) has proposed a therapeutic potential for the iTBS protocol targeting the affected M1 at the hand representation in patients with post-stroke aphasia (Xu S. et al., 2021). These results were in accordance with previous reports showing functional connectivity between the cortical language network and the M1 hand representation (Meister et al., 2003, 2006; Meinzer et al., 2016). While in patients with post-stroke dysarthria, applying low-frequency rTMS to the unaffected mouth representation of M1 in combination with speech therapy improved the articulation functionality significantly in comparison to speech therapy alone (Kwon et al., 2015).

Central post-stroke pain (CPSP) is defined as a pain occurring after a cerebrovascular lesion of the brain or brainstem (Scholz et al., 2019). This pain is felt in the body region corresponding to the central nervous structure affected by stroke (Scholz et al., 2019). It is estimated that more than 50% of patients with a stroke affecting the somatosensory tract will develop CPSP (Liampas et al., 2020). The pathophysiology of CPSP is related to a defective gamma-aminobutyric acid (GABA)-ergic inhibition inside the brain, leading to maladaptive plasticity in multiple cortical regions, including M1 (Tang et al., 2019). Therefore, many studies have targeted the M1 cortex with rTMS to restore the intracortical inhibition in patients with CPSP. As a result, a direct relationship was noted between the analgesic efficacy and the modulation of M1 cortical excitability (Lefaucheur et al., 2006; Hosomi et al., 2013). In addition, the analgesic efficacy of M1 stimulation is believed to stem from the interconnection between M1 and the endogenous opioid system. This relation was evidenced by positron emission tomography scans (Maarrawi et al., 2007; Lamusuo et al., 2017) and pharmacological blocking of the μ-opioid receptors, which minimized the analgesic efficacy of M1 stimulation (de Andrade et al., 2011). On the neural network level, M1 stimulation was found to modulate the excitability of other cortical and subcortical areas related to sensory, cognitive, and emotional components of pain, such as the thalamus, insular cortex, and anterior cingulate gyrus (García-Larrea et al., 1999; Hasan et al., 2014). Recent systematic reviews have shown that 5–10 sessions of high-frequency rTMS over M1 of the affected hemisphere resulted in a significant reduction in the CPSP intensity which lasted for at least 3 weeks post-treatment (Ramger et al., 2019; Liampas et al., 2020).

Depression is a common complaint after stroke, affecting approximately 30% of stroke survivors (Towfighi et al., 2017). Patients with post-stroke depression are at higher risk of recurrent strokes and mortality (Towfighi et al., 2017). Previous rTMS studies have mainly targeted the dorsolateral prefrontal cortex as a therapeutic approach in patients with post-stroke depression (Shen et al., 2017). However, preliminary evidence has shown that low-frequency stimulation of the contralesional M1 cortex might have an antidepressant effect in depressed stroke patients (Carey et al., 2008; Niimi et al., 2020). The mechanistic rationale for the antidepressant efficacy of M1 stimulation might stem from its effect on the kynurenine levels (Kepplinger et al., 2014; Niimi et al., 2020), a tryptophan metabolite and one implicated pathway in depression (Ogyu et al., 2018). In addition, a recent meta-analysis combined with functional magnetic resonance imaging studies has identified M1 as a region of interest (ROI) for rTMS in depressive disorders (Zhang et al., 2020). Notably, some individual studies have reported an antidepressant efficacy following high-frequency rTMS at unilateral or bilateral M1 in Parkinson’s disease and chronic pain conditions (Khedr et al., 2015; Makkos et al., 2016; Li et al., 2020; Bursali et al., 2021). On the other hand, the antidepressant efficacy was not found in other studies with similar conditions and stimulation parameters (Brys et al., 2016; Lindholm et al., 2016; Hosomi et al., 2020). However, these findings need to be further explored in larger randomized controlled studies. In addition, it is worth mentioning that the emotional improvement following rTMS at M1 might be an indirect effect of the concomitant improvement in other symptoms, such as motor symptoms and pain.

It is estimated that up to 83% of stroke survivors suffer from cognitive impairment in at least one cognitive domain (Jokinen et al., 2015). Similar to depression, the dorsolateral prefrontal cortex has been a conventional target for rTMS in patients with post-stroke cognitive impairments (Liu M. et al., 2021). However, mounting evidence suggests that the M1 is directly involved in higher cognitive processes including, but not limited to, attention, memory, motor imagery, and language comprehension (Tomasino and Gremese, 2016; Vukovic et al., 2017; Bhattacharjee et al., 2021; Vitale et al., 2021). In this context, two recent studies have shown that low-frequency rTMS stimulation over the contralesional M1 improved the measures of global cognition and visuospatial recall memory in stroke patients (D’Agata et al., 2016; Askin et al., 2017). Furthermore, this improvement was associated with reduced latency in the N200 and P300 markers of event-related potentials, indicating an increasing speed in the perceptual and cognitive processes (D’Agata et al., 2016). Noteworthy, the improvement in measures of global cognition along with the reduced latency in P300 following high-frequency bilateral M1-rTMS has also been reported in patients with Parkinson’s disease-related dementia (Khedr et al., 2020). This, in turn, might suggest a consistent procognitive effect of the M1-rTMS regardless of the underlying pathology.

With the slow and often incomplete recovery from stroke sequelae, rTMS is becoming an increasingly used technique in post-stroke rehabilitation. Stimulation of the M1 cortex has long been employed to treat residual motor symptoms of stroke incidents. However, accumulating evidence suggests that M1 stimulation can ameliorate a variety of non-motor symptoms that stroke patients frequently experience. The mechanistic rationale behind the management of these symptoms varies depending on the neural networks involved in their pathophysiology. Nonetheless, whether applied in the acute or chronic phase, alone or in combination with other interventions, rTMS stimulation over M1 can yield a therapeutic efficacy that extends beyond the movement execution to involve swallowing, speech, pain, mood, and cognition. Therefore, well-conducted randomized controlled trials, in particular studies combining TMS with EEG and neuroimaging techniques, will help expand our knowledge of the cortical and subcortical connections of M1 in health and disease and ultimately tailor the therapeutic use of rTMS based on the constellation of symptoms in each patient. In addition, as we approach the personalized medicine era, accumulating evidence highlights the influence of genetic polymorphisms on the rate of stroke recovery (Math et al., 2019). This might explain, in part, the variability in responses to the TMS in post-stroke rehabilitation. Therefore, future studies should further explore the potential genetic polymorphisms that interact with TMS responses in stroke, bearing in mind that the most investigated one, the brain-derived neurotrophic factor, has not yet proven as a decisive factor in the M1-rTMS literature (Sasaki et al., 2021).

AT devised the idea for this review and prepared the initial manuscript draft. AT, AHKYK, and WAWS wrote, edited, and revised the final manuscript. All authors approved the final version of the manuscript before submission.