Keywords

25.1 Introduction

Epilepsy affects over 50 million people worldwide, and about 30% of these patients have resistance to treatment with antiseizure medications (ASMs) (Kuzan-Fischer et al. 2020). Namely, drug resistance or pharmacoresistant epilepsy is a health problem with a significant burden in comorbidity and an increased risk of premature death (Brodie 2005; Casadei et al. 2020; Ryvlin et al. 2011). Therefore, efforts to improve therapeutic options, both pharmacological and nonpharmacological treatments, which are currently in the pipeline, have as their main objective the best possible control of seizures (Abdullahi et al. 2022; Bex et al. 2022; Hilz 2022; Riva et al. 2021).

Despite the development of new and safer ASMs, there is no solid evidence to indicate the improvement of the efficacy of these new drugs in patients with drug-resistant epilepsy (Brodie 2005; Enia et al. 2021). In fact, it is known that, when two ASMs fail controlling seizures, the association of another one could have a success less than 5% (Perucca et al. 2011; Wiebe 2004; Wirrell 2013).

Although epilepsy surgery is not the first line of treatment, it is considered a therapeutic option in patients with drug-resistant epilepsy (DRE). It has been shown that surgical resection allows control of epileptic seizures in about 40–80% of patients who fulfill criteria for epilepsy surgery (60–80% in temporal lobe epilepsy, and 40–60% in extratemporal lobe epilepsy) (Morales Chacón et al. 2018, 2021b). Even so, there is a considerable number of patients who still present seizures after epilepsy surgery. Thus, there is a growing interest in applying neuromodulatory techniques such as vagus nerve stimulation, deep brain stimulation, and nonInvasive brain stimulation (NIBS) as a potential strategy to control DRE (Brodie 2005; Hachem et al. 2019; Kwon et al. 2018; Watrous et al. 2015).

At present, repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS) are the most popular noninvasive brain stimulation (NIBS) technologies. These methods use electrical fields generated noninvasively in the brain to long-lastingly modulate the excitability/activity of brain regions contributing to relevant processes (Sanches et al. 2020). Originally, pulsed transcranial magnetic stimulation (TMS) was used for research and functional diagnosis exploring corticospinal pathways. In the 90s, Pascual-Leone et al. found evidence supporting its potential therapeutic use (Pascual-Leone et al. 1999). Likewise, the development of TMS and rTMS allowed the rediscovery of tDCS and its use with therapeutic purposes. However, it was not until the end of the 90s and the early 2000s that tDCS gained attention in the scientific community, resulting in a vast increase in the number of studies examining its potential clinical applications (Been et al. 2007; DeMarse and Carney 2014; Gómez et al. 2017, 2014; Nitsche et al. 2012; San-Juan et al. 2015).

Equally important, noninvasive brain stimulation (NIBS) has been seen common in rehabilitation settings for the treatment of stroke, spinal cord injury, traumatic brain injury, and multiple sclerosis, as well as for some diagnostic neurophysiological measurements (Kesikburun 2022). Correspondingly, neurological disorders such as epilepsy, Parkinson’s disease, Alzheimer’s disease, and depression are conditions that can benefit from these emerging technologies (Eastin and Lopez-Gonzalez 2017).

There are currently some studies suggesting that both methodologies could be effective as adjuvant therapeutic approaches to control pharmacoresistant epilepsy in patients (Carvalho et al. 2018; Chan et al. 2018; Izadi et al. 2018; Lin et al. 2018; Ng et al. 2018). Explicitly, specific modulatory effects can be obtained with different NIBS techniques, improving activity in excitatory networks or increasing intracortical inhibition, and then, reducing the cortical excitability (Chervyakov et al. 2015; Muller-Dahlhaus and Vlachos 2013; Roche et al. 2015). Thus, this chapter will review the potential clinical use of NIBS in drug-resistant epilepsy therapy.

25.2 Non-invasive Brain Stimulation: Basic Principles and Protocols

Even if rTMS and tDCS share effects at the physiological level, their physical properties are not the same. These noninvasive techniques are applied directly through electrodes or magnetic fields on the scalp of the patient to produce electrical currents for the stimulation of brain cells (Camacho-Conde et al. 2022).

NIBS application may result in modulatory cortical inhibition/excitation balance with behavioral and physiological consequences that outlast the stimulation duration. Indeed, these effects may well last weeks or months when rTMS and tDCS are repetitively applied (He et al. 2020; Sever et al. 2022). Based on this information, NIBS might be useful as potential therapy for neurological and psychiatric disease such as pain, movement disorders, stroke, amyotrophic lateral sclerosis, multiple sclerosis, epilepsy, consciousness disorders, tinnitus, depression, anxiety disorders, obsessive–compulsive disorder, schizophrenia, craving/addiction, and conversion (Elyamany et al. 2021; Sanches et al. 2020).

25.2.1 Repetitive Transcranial Magnetic Stimulation

TMS is a neuromodulation technique that uses large transient magnetic fields to induce focal electrical fields in a specific brain area, and the availability of sophisticated equipment has made it possible to employ repetitive TMS (rTMS). The effects of rTMS vary depending on the shape of the coil (figure of eight, H coil, double cone coil), pacing pattern (high frequency, low frequency, theta-burst), and stimulation site (Camacho-Conde et al. 2022; Fregni et al. 2006a).

In addition, the effects induced by rTMS correlate with long-term depression (LTD) and long-term potentiation (LTP), two forms of synaptic plasticity elicited in animal models of cortical circuitry by low- and high-frequency electrical stimulation, respectively (Bliss and Cooke 2011). Besides, rTMS applied at low frequency may exert antiepileptic effects by inducing LTD whereas at high-frequency stimulation, it may facilitate proconvulsant effects (Ziemann et al. 2015). Nevertheless, both phenomena by themselves are insufficient to elucidate the early and long-term changes that take place after short NIBS episodes. Other mechanisms, including enhancement of GABAergic inhibition (Pascual-Leone et al. 1999), may also be involved in the anticonvulsant effects caused by rTMS.

Previous studies using preclinical models indicate that rTMS applied at low frequency (0.5 Hz) reduces the occurrence of status epilepticus and increases the latency of pentylenetetrazole-induced seizures (Akamatsu et al. 2001). It has also been demonstrated in hippocampal and neocortical rat slices that low-frequency (1 Hz) electrical stimulation is able to prevent interictal epileptic discharges and epilepsy-like events in an intensity-frequency and distance-dependent manner. These effects persist after the end of stimulation and are NMDA-receptor dependent, thus indicating that LTD-inducing protocols might have antiepileptic properties (Lanza et al. 2022; Lefaucheur et al. 2014).

The findings reported in animal models are consistent with the potential therapeutic use of LF-rTMS in patients with epilepsy (Ben-Menachem and French 2005; Fregni et al. 2006b). However, further studies are required to analyze the effects of rTMS in experimental models of drug-resistant seizures.

Even though underlying mechanisms of the therapeutic outcomes of rTMS application have not been fully explained, rTMS can induce changes in cerebral blood flow, oxygen consumption, cortical activity, and release of neurotransmitters. As a result, it has been argued that these functional changes might be associated with positive clinical results (Camacho-Conde et al. 2022; George et al. 2003).

25.2.2 Transcranial Direct Current Stimulation (tDCS)

Transcranial direct current stimulation (tDCS) represents a re-emerging noninvasive brain stimulation technique that has been used in animal models and human trials aimed to elucidate neurophysiology and behavior interactions (Sudbrack-Oliveira et al. 2021). When using tDCS, continuous but low-intensity current is applied through electrodes (anode and cathode) placed on the scalp. High-definition tDCS (HD-tDCS) is a variant of this technique, and, in contrast to tDCS where distribution of electrical current in a target area is relatively diffused, HD-tDCS devices are used to increase focal stimulation of a target area. In comparison with rTMS, tDCS is not as powerful and generates weak stimulus; however, it is relatively easy to use and transport, lot less costly, and it has low incidence of side effects (He et al. 2020).

It has been shown that tDCS modulates spontaneous neuronal activity through changes in the resting membrane potential. Moreover, the effect of tDCS varies according to the type of current (direct, alternating, pulsed, and random noise), polarity (anodal or cathodal), current intensity, and stimulation site (Guleyupoglu et al. 2013).

Concerning changes of tDCS in the motor cortex excitability, previous research findings have shown that they are polarity dependent, i.e., anodal tDCS increases motor cortex excitability, whereas cathodal tDCS decreases it. Equally, anodal tDCS increases motor-evoked potential amplitude by about 40%, while cathodal tDCS decreases it (Nitsche et al. 2012; Sudbrack-Oliveira et al. 2021).

A number of studies corroborate that tDCS effects are mediated by D2 and NMDA receptors, regardless of the polarity used (Auvichayapat and Auvichayapat 2011; Basavaraju et al. 2019). For example, using magnetic resonance spectroscopy, Stagg et al. 2010, observed that anodal tDCS decreased GABA brain concentration, with no changes in glutamate concentration, whereas catodal tDCS reduced both glutamate and GABA concentrations (Stagg et al. 2010). Other tDCS effects include changes in dopamine, acetylcholine, and serotonin (Beuthien-Baumann et al. 2005; Cirillo et al. 2017).

Furthermore, preliminary experimental studies indicate a multifaceted scenario potentially relevant to the therapeutic effects of NIBS, including gene activation/regulation, de novo protein expression, morphological changes, alterations in intrinsic firing properties, and modified network properties resulting from changed inhibition, homeostatic processes, and glial function (Cirillo et al. 2017).

25.3 Safety and Tolerability of Noninvasive Brain Stimulation in Patients with Drug-Resistant Epilepsy

During the past two decades, there has been an increasing use of NIBS techniques to treat drug-resistant epilepsy in patients (Acerbo et al. 2022; Alicart et al. 2021; Bermpohl et al. 2006; Nourski et al. 2015). Nevertheless, it is clear that further research should be done to support the effectiveness of these therapeutic strategies. An important issue is that they induce few adverse effects (Been et al. 2007; Begemann et al. 2020; Caldwell et al. 2019). It has commonly been assumed that adverse effects are more associated with rTMS than with tDCS, comprising transient headaches and scalp discomfort as a consequence of activation of scalp pericranial muscles. It has also been demonstrated that it induces tenderness and neck pain. Conversely, TMS seldom induces seizures. Unusual adverse effects also involve syncope, fainting, nausea, vomiting, auditory change, and hypomania. (Muller et al. 2012; Pereira et al. 2016; Zewdie et al. 2020).

In general, the most common adverse effects induced by tDCS encompass local pain, tingling, itching or skin irritation on the stimulation area, fatigue, drowsiness and headache (Brandt et al. 1997; Cortes et al. 2017; Ille et al. 2016; Zewdie et al. 2020).

It is a widely held view that the most feared adverse effect induced by NIBS in patients with epilepsy is the induction of seizures. In this respect, the occurrence of focal seizures was described as a result of applying tDCS with the anode placed on the paracentral region (1.2 mA, 20 min), in a four-year patient with history of epileptic spasms 2 years prior to the intervention (Ekici 2015). On the other hand, no changes in the seizure frequency were detected in a randomized control study including 37 patients with focal temporal lobe epilepsy, in whom the anode was placed on the projection of the dorsolateral prefrontal cortex in order to reduce depression-related symptoms. Moreover, none of the patients showed an increase in the frequency of seizures during the four-week follow-up period after the intervention (Liu et al. 2016).

In a review of 172 studies using tDCS, 56% of them reported adverse effects in both groups of patients: those who received stimulation and those who had placebo stimulation. In this framework, the adverse effects were tingling, sensation of burning, redness, and headache in all cases of low intensity and short duration (Brunoni et al. 2012).

Similarly, in a study performed by Fregni et al. (2006c) to examine the effects of cathode tDCS (one session, 20 min, 1 mA) in ten patients with drug-resistant epilepsy and cortical dysplasia, no adverse effects were observed (Fregni et al. 2006c). Further to this, in an investigation carried out by Auvichayapat et al. (2013) in pediatric patients with DRE, one of the 27 children who received active cathodic tDCS experienced erythematous rash below the reference electrode, that disappeared within 2 h after stimulation (Auvichayapat et al. 2013).

With regard to rTMS, the existence of seizures has been more frequently reported. After the establishment of the safety guide for the use of TMS in 1998 and until 2008, the occurrence of seizures as a consequence of this type of NIBS was indicated in nine reports. Four of them apparently occurred in patients in whom safe stimulation parameters were used, and three of them in patients who received proconvulsant drugs. The remaining seizures happened under the use of stimulation protocols that were not considered to be safe (Liu et al. 2016). As a whole, the presence of seizures as a result of rTMS has been described in few patients with different neurological pathologies, most of them using high-frequency protocols (Bermpohl et al. 2006).

In 2007, Bae et al. analyzed 280 people with epilepsy in order to evaluate the safety and tolerance to rTMS. In this study, in a total of 152 epileptic patients who received rTMS ≤1 Hz sessions, no epileptic seizures related to stimulation were observed. The most frequent adverse effect was headache, occurring in 9.6% of the patients. On the other hand, epileptic seizure occurrence was confirmed in four patients. In 3 of them, typical seizures, considering duration and semiology, were reported. So, it is likely that the seizures did not have a causal relationship with the stimulation. Conversely, in the other patient, atypical seizures from the stimulation region during high-frequency rTMS were recognized, suggesting a causal connection between the stimulation and the occurrence of seizures. No rTMS-related episodes of status epilepticus were described in this review. All together, the authors in this study concluded that the risk of developing seizures was about 1.4% (Bae et al. 2007). Interestingly, findings generated in a recent review suggest that if seizures occurred, they are usually self-limiting, and the risk of TMS-related seizures is <1% overall. The rate of TMS-related seizures is comparable to that of most psychotropic medications. Thus, most treatment recommendations for TMS-related seizures are supportive in nature (Stultz et al. 2020).

In sum, after more than three decades using tDCS, and more than 20 years applying TMS in human as well as in experimental models, along with the safety guidelines established by the International Federation of Clinical Neurophysiology, there is no evidence for irreversible injury produced by conventional stimulating protocols used to apply NIBS (Giustiniani et al. 2022; Rossi et al. 2021). These facts promote the idea that both rTMS and tDCS can be considered safe NIBS.

25.4 Noninvasive Brain Stimulation as Therapeutic Procedure: Effects on Seizures and Interictal Epileptiform Discharges in Drug-Resistant Epilepsy

Traditionally, studying clinical outcomes in epilepsy patients includes seizure frequency assessment. Yet, one of the main techniques to quantitatively measure the benefit of NIBS in these patients is focusing on change in the count of interictal epileptiform discharges (IEDs) using electrophysiological recordings.

Most of the studies show that tDCS results in a decrease of IEDs frequency. However, with regard to seizure frequency (SF), the findings are varied (Gschwind and Seeck 2016; Kwon et al. 2018; San-Juan et al. 2015). In this sense, Fregni et al. (2006a, b, c) used a cathodic tDCS protocol (one session, 20 min, 1 mA) in 19 patients with DRE due to dysplasia, reporting a decrease of 64.3% in IED, along with 44% of seizure reduction (Fregni et al. 2006c). Relatedly, in a control study involving 36 children with DRE using the same stimulation protocol, Auvichayapat et al. (2013) described that patients receiving tDCS showed a decrease in discharges by 45.3% immediately after the intervention, and 57.6% at 48 h; however, there were no significant changes in the frequency of seizures when compared with the control group (Auvichayapat et al. 2013). Two other studies reported significant (>50%) decrease in seizure frequency in patients suffering from Rasmussen’s encephalitis (San-Juan et al. 2011; Tekturk et al. 2016). Another clinical pediatric trial conducted by Auvichayapat et al. (2016) has shown reduction in SF of 55.9% in patients with epileptic spasms and Lennox Gastaut syndrome (LGS) compared to sham group 1 month after 5 consecutive days of 20 min tDCS at 2 mA (Auvichayapat et al. 2016).

Later, 28 patients suffering from mesial temporal lobe epilepsy with hippocampal sclerosis were enrolled in a randomized placebo-controlled, double-blinded clinical trial where they received one session of tDCS at 2 mA for 3 or 5 days (San-Juan et al. 2017). Two months after the cathodal tDCS session, they obtained a decrease of −43% in SF for the group with a three-day stimulation and a decrease of −55% in SF for the group receiving tDCS for 5 days compared to baseline. So that, the heterogeneity of epilepsy types among studies demonstrates the potential efficacy of cathodal tDCS for treating several etiologies of refractory focal epilepsy.

It is recognized that the effect of the TMS is mediated by variables in terms of the stimulation frequency and the intensity of the stimulus. In this regard, most of the studies carried out to date in patients with DRE have applied rTMS at low frequency, and used different protocols including more than two treatment sessions (Erőss et al. 2015; San-Juan et al. 2017; Schulze-Bonhage 2019). Similar to the studies in which cathodic tDCS has been used, not all the investigations related to rTMS have reported how the frequency of IED behaves after the stimulation sessions have been applied (Lefaucheur et al. 2014, 2017). Nevertheless, some authors provide evidence of the significant inhibitory effect of rTMS on IEDs without clinical changes in the seizure frequency (Cantello et al. 2007; Joo 2012). In a control study developed by Sun et al. (2012), TMS was used at low frequency (0.5 Hz) in 60 patients with DRE. In the group where active stimulation was applied, the frequency of seizures decreased from 8.9 ± 11.1 per week to 1.8 ± 3.7, while in the control group, no changes were observed (Sun et al. 2012).

Several studies have demonstrated that LF-rTMS may reduce seizure frequency in patients with refractory epilepsy (Fregni et al. 2005; Joo 2012; Lefaucheur et al. 2014). Noticeably, the findings derived from controlled trials are mixed in relation to antiepileptic rTMS efficacy (Cantello et al. 2007; Fregni et al. 2006b; Kwon et al. 2018; Theodore and Fisher 2004), and the field would benefit from further carefully randomized-controlled trials.

The variations in the results described in the literature regarding IEDs frequency and SF may be due to the great heterogeneity that exists in terms of clinical characteristics and treatment lines with ASD of the samples and the corresponding parameters of the stimulation protocols. Previous research findings indicate that IEDs are not necessarily generated in the ictal onset zone, but can be generated in a wider area that is called irritative zone (Hilz 2022).

In an experimental study in animals using deep brain stimulation, Sobayo and Mogul (2016) found that the stimulation was more effective when the stimulation parameters corresponded to the seizure characteristics of each animal in terms of location and frequency of termination (Sobayo and Mogul 2016). That said, it is probable that the personalization of NIBS protocols is the way to increase the effectiveness of these techniques, taking into account that the physiopathological and clinical characteristics, the location, and extension of the epileptogenic zone (EZ) and treatment with ASM are very fluctuating among patients with DRE.

Interestingly, brain modeling and human studies highlight the influence of individual brain anatomy and physiology on the electric field distribution (Simula et al. 2022). Recent advancements in vivo electric field characterization may enable clinical researchers to derive better relationships between the electric field strength and the clinical results. Also, subject-specific electric field simulations could lead to improved electrode placement and more efficient treatments. Accordingly, processing methods result in personalized NIBS based on metrics like focality and field strength, which allow for correlation with clinical outcomes (Beumer et al. 2022).

Unquestionably, there is a lack of clinical studies investigating changes in intracranial epileptiform discharges during NIBS application, which could make clear the nature of TMS and DCS-related local and network dynamics in epilepsy.

25.5 Evaluation of Non-invasive Brain Stimulation Effects on Electroencephalogram Functional Connectivity

There are well-known effects on functional connectivity in both normal and clinical populations as demonstrated in functional magnetic resonance imaging (fMRI) and Electroencephalogram (EEG) studies. Functional imaging techniques such as positron emission tomography, fMRI, and EEG mapping enable assessment of TMS-related functional brain activation (Pascual-Leone et al. 2011; Ruffini et al. 2014; Shafi et al. 2016). Studies combining resting-state functional magnetic resonance imaging with NIBS allow delineating how stimulation of different brain regions induces complex network modifications, both at the local and distal level. More recently, some studies involving magnetic resonance spectroscopy and NIBS have demonstrated how microscale changes are related to modifications of large-scale networks (Pini et al. 2018).

Altogether, a combination of TMS and functional imaging can be useful in three principal ways: (1) brain imaging before TMS is helpful in defining the accurate coil position over a distinct cortical area targeted by TMS; (2) imaging the brain during TMS is a promising approach for assessing cortical excitability and intracerebral functional connectivity; and (3) brain imaging after TMS can be employed to study the plasticity of the human cortex by evaluating lasting effects of TMS. Undoubtedly, this approach will help to advance our understanding of the therapeutical effects related to TMS (Siebner et al. 2009). Likewise, TMS has been also used in conjunction with EEG (TMS-EEG) to evaluate neurophysiology for a variety of indications. TMS-EEG has significant potential for exploring brain connectivity using focal TMS-evoked potentials and oscillations, which may allow for the system-specific delineation of neural recovery patterns after neurological diseases (Keser et al. 2022).

In relation with tDCS, Stagg et al. (2010) assessed the modulation of cerebral perfusion during and after tDCS application to the left dorsolateral prefrontal cortex. These authors described an increase in the perfusion of the primary sensory-motor cortex, cingulate cortex, and left parietal cortex as compared with baseline perfusion during stimulation with anodal tDCS. On the other hand, cathodal tDCS decreased perfusion in thalamus and in temporal lobe compared with baseline state (Stagg et al. 2010). Recent articles have reported changes of functional connectivity in epileptic patients after tDCS. Further to that, there is evidence that tDCS may act by affecting brain networks, rather than simply modifying local activity in the targeted area (Simula et al. 2022).

Nowadays, the analysis of functional connectivity and the application of graph theory to evaluate the behavior of neuronal networks in patients with DRE who have undergone some neuromodulation techniques constitute a new variant for the approach of the predictive value of the therapeutic response of the intervention, as well as the evaluation of its effect. Moreover, graph theory research has been progressively used to analyze brain networks in different structural and functional modalities (Chiang and Yang 2019; Pedersen et al. 2019). It is important to highpoint that functional and structural connectivity clarifies not only that but also the extent to which different brain zones are connected, whereas network analysis using graph theory provides a framework to characterize the topological organization of functional and structural networks, before, during, and after the stimulation.

The most common parameters utilized in neuronal network analyses using graph theory are the clustering coefficient and the characteristic path length. The clustering coefficient allows to define the local segregation property of the network, and it is used to assess the network capability to share specialized data, while the path length and global efficiency are used to evaluate the capacity of the network as a whole for inner-exchange information. A short path length, a low clustering, and a high global efficiency/local efficiency generally represent a small world topology of the network and characterizes an optimal organization for communication efficiency (Morales Chacón et al. 2021a).

In a control study in patients with extratemporal focal epilepsy using functional magnetic resonance imaging (fMRI), Pedersen et al. (2015) observed increased segregation (clustering coefficient and local efficiency) compared to healthy subjects (Pedersen et al. 2015). In another investigation carried out by Antony et al. (2013), where an analysis of the functional connectivity derived from EEG was made to predict the post-surgical clinical evolution in patients with drug-resistant temporal lobe epilepsy, an increase in the clustering index and the average length was found precisely for the frequency bands slow (Pedersen et al. 2015).

Also, Tecchio et al. (2018) assessed the cathodic tDCS-induced changes of electroencephalography-derived brain functional connectivity using low-resolution electromagnetic tomography (eLORETA) in patients with temporal lobe epilepsy. The findings indicated that about 73% of the changes in functional connectivity involved the epileptogenic zone and that the reduction of seizures were correlated with the increase in functional connectivity (Tecchio et al. 2018). This study also supports the hypothesis that functional connectivity changes may contribute to explain the effects of tDCS in epilepsy, offering a new scenario in the personalization of neuromodulation interventions in epileptic people.

All things considered, research in patients with DRE using fMRI, EEG, magnetoencephalography, and electrocorticography has shown an increase in connectivity patterns around the epileptic zone. On the other hand, the connectivity of the epileptic zone with distant neural networks is diminished, and this is related to the duration and severity of the disease (Morales Chacón et al. 2021a). However, it is not entirely clear how these characteristics of neural networks could be modulated by NIBS.

25.6 Conclusions

As has been demonstrated in this chapter, NIBS can be considered a potential therapeutic alternative for patients with drug-resistant epilepsy who are not eligible for epilepsy surgery. However, NIBS antiepileptic efficacy will have to be determined in future randomized placebo-controlled trials. Equally, further multicenter clinical studies will need to be undertaken to elucidate the effects of these strategies when combined with ASMs, measures of outcome assessment, and eligibility criteria of patients, including novel methods such as personalized stimulation protocols based on computational modeling.