Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive method to temporarily modulate neural activity (Silvanto and Muggleton, 2008; Miniussi et al., 2013) with long-lasting changes of cortical excitability (Robertson et al., 2003). Importantly, rTMS not only acts locally on interneural circuits underneath the coil but its effects also spread to functionally connected brain regions along cortico-cortical connections (Bortoletto et al., 2015; Siviero et al., 2023). When TMS is coupled with EEG, it is possible to map the network of cortical areas engaged in a specific cognitive function and give information on the functional coupling among brain areas with a very high temporal resolution (Ilmoniemi et al., 1997; Miniussi and Thut, 2010) thus providing information on the causal role of TMS-induced cortical changes on distant, but functionally interconnected, areas. The advantage of recording EEG before and after the delivery of a rTMS protocol is that of potentially providing information on the causal relationship between functionally connected areas in terms of both the temporal dynamics (pre- vs. post-stimulation) and the nature (excitatory vs. inhibitory) of the signal that is spread. That is, if an area A is active prior to area B, it can be assumed that area A and B are functionally connected, because a change in area A has an effect on area B, and that activity in area A causes a change in the activity of area B. Moreover, if area A, as an example, is inhibited by rTMS and a coherent reduction of activity is found in area B, the two areas are positively connected (i.e., excitatory connection), conversely if a reduction of activity in area A is followed by an enhancement of activity in area B, the two areas are negatively (i.e., inhibitory connection) connected (Miniussi and Thut, 2010).

Importantly, EEG can provide useful information in terms of brain oscillations which are a notable feature of brain activity and they are thought to play a central role in cognition (Wang, 2010; Tafuro et al., 2023). EEG signal can indeed be decomposed into frequency bands (Berger, 1929), each of which showing a specific power reflecting the amount of synchronized oscillatory activity expressed by neuronal ensembles oscillating in that specific frequency band. Importantly, the power of each frequency band can be used to calculate coherence, i.e., a measure reflecting the strength and the sign of the interplay between patterns of oscillating brain activity recorded at different locations on the scalp, and it is suggested to sustain effective communication among groups of neurons (Fries, 2005). Thus, high coherence between EEG signals recorded at different electrodes suggests high functional coupling between the underlying neuronal networks (i.e., functional connectivity). Coherence can be measured while participants are not performing any task (resting state) and it can, thus, reveal functional connectivity of areas in the resting state. Moreover, if resting state coherence is recorded prior and after the application of a rTMS protocol, it can give important insights on rTMS-induced perturbation effects on the state of the cortex in different areas of the brain, thus revealing both the network(s) of areas functionally connected with the stimulated site and the nature (excitatory vs. inhibitory) of this functional coupling. That is, if coherence among two sites is increased after inhibition of one of the two, it reflects an excitatory connection between the two sites (i.e., both sites are inhibited), conversely if coherence is decreased after the inhibition of one of the two, it reflects an inhibitory connection between the two sites (one site is inhibited and the other is excited).

From a translational perspective, the possibility to detect changes at the network level by means of measures of coherence in brain networks can pave the way for the development of effective interventions aimed at restoring cognitive deficits in stroke patients. Indeed, low-frequency rTMS of the right (contralesional) frontal cortex has been shown to reduce non-fluent aphasia symptoms (Shah et al., 2013), whereas low-frequency rTMS of the left (contralesional) parietal cortex has been shown to reduce symptoms (Hesse et al., 2011; Oliveri, 2011; Müri et al., 2013) in spatial neglect, i.e., a failure to report, respond, or orient to stimuli presented to the contralesional space (Heilman and Valenstein, 1979; Schenkenberg et al., 1980; Savazzi et al., 2007). The logic underlying the application of rTMS on the contralesional hemisphere stands on the assumptions of the interhemispheric rivalry model (Kinsbourne, 1977) stating that, in normal conditions, the two hemispheres inhibit each other. Accordingly, after a lesion to one hemisphere, the contralesional hemisphere is released from the inhibition exerted by the damaged one and becomes hyperactive. However, despite several pieces of evidence showing the efficacy of rTMS in restoring cognitive functions, little is known about the neural underpinnings allowing this to occur and different mechanisms are proposed for the amelioration of symptoms in aphasia and neglect patients. In this respect, amelioration of neglect symptoms, by applying rTMS to the contralesional hemisphere, has been invariably explained in terms of a reduction of the hyperactivity of the undamaged hemisphere and a restoration of the interhemispheric balance (Jacquin-Courtois, 2015). Conversely, amelioration of aphasia symptoms has been interpreted both as a restoration of the interhemispheric balance and as a reduction of maladaptive activation of areas within the right hemisphere (Gainotti, 2015). Importantly, most of the studies showing hyperactivity of the contralesional hemisphere in patients with neglect refers to sub-acute or chronic patients and, thus, leaving open the possibility, as proposed in aphasia literature, that the hyperactivity of the contralesional hemisphere might be the result of maladaptive plasticity (Umarova et al., 2011; Pia et al., 2012; Ricci et al., 2012; Bagattini et al., 2015b) as opposed to a release from inhibition caused by a lesion. Thus, understanding the role of the non-lesioned hemisphere in the genesis and maintenance of the deficits is important not only for a better comprehension of the neural plasticity and reorganization after stroke, but it will be crucial for devising more effective rehabilitation protocols (Bartolomeo, 2014; Lunven et al., 2015).

Several attempts to understand the interplay between the hemispheres to uncover the effects of hypo-activity (due to a lesion or caused by low frequency, inhibitory, rTMS) in one hemisphere on the functionality of the other (i.e., induced hyper-activation or hypo-activation) have been made (e.g., Koch et al., 2008, 2011, 2013; Salatino et al., 2014). However, none of them have used a direct approach by investigating coherence among brain areas with inhibitory TMS approaches for areas different from motor ones (e.g., Strens et al., 2002; Chen et al., 2003). To the best of our knowledge, such an approach to directly test the hypothesis that a reduction of activity in one hemisphere (right parietal lobe) would result in an enhancement of activity in homologous areas in the other hemisphere was present only in a paper by Bagattini et al. (2015b). In this paper, the authors applied 1 Hz rTMS in healthy participants to reduce cortical reactivity in one hemisphere and concurrently measured, with EEG, the effects of this reduction in functionally connected areas. They found that the sites contralateral to the stimulation exhibited a reduction in activity similar to that induced in the stimulated hemisphere. Similarly, adopting an interleaved TMS-fMRI approach, Ricci et al. (2012) showed that low frequency rTMS over right posterior parietal cortex (rPPC) caused decreased neural activity of parieto-frontal regions and of the homologous parietal regions of the left hemisphere. These results were, thus, in direct contrast to the hypothesis of a reciprocal inhibition of the two hemispheres suggesting that what is proposed for aphasia, i.e., maladaptive plasticity, could also account for spatial neglect.

Here, we extend previous investigations by means of an inhibitory rTMS protocol applied over rPPC and by recording resting state EEG before and after the application of rTMS. Coherence between the stimulated site and other cortical areas is used as a measure of functional connectivity to reveal both the functionally interconnected networks engaged by the stimulation and the nature of this connection. The aim of the present paper is, thus, that of revealing whether the reduction of cortical activity induced by 1 Hz rTMS in one hemisphere leads to a reduction (high coherence) of neural activity in other areas within the same hemisphere or on the contralateral hemisphere, thus providing evidence on cortico-cortical connectivity between rPPC and other brain areas. Implications for possible recruitment of intra- and inter-hemispheric neural pathways in spatial neglect are discussed.

The present study investigated the effects of repetitive transcranial magnetic stimulation (rTMS) on functional connectivity through coherence in the brain using EEG. The results contribute to our understanding of the neural mechanisms underlying the therapeutic effects of rTMS and its potential application in stroke rehabilitation. Specifically, we investigated intra- and inter-hemispheric coherence changes induced by rTMS of rPPC. Healthy participants were administered with a 30 min of low frequency (1Hz) rTMS protocol, which has been shown to reduce neural excitability (Chen et al., 1997; Boroojerdi et al., 2000), while resting-state EEG signal was recorded before and after rTMS. The main goal of the present investigation was to uncover the networks of areas functionally interconnected with the stimulated (rPPC) area by measuring EEG coherence. The logic was the following: higher coherence between rPPC and other distant areas would imply that these areas are functionally connected with rPPC and that their activity after rTMS is changed in the same manner (functional coupling) as the activity in rPPC. Specifically, we found that the sites showing a functional coupling (increased coherence) with rPPC after rTMS are the homologous (parietal) sites on the contralateral (left) hemisphere, in line with previous findings using TMS-evoked potentials (Bagattini et al., 2015b), and fronto-central sites within the stimulated (right) hemisphere. Accordingly, and increase in power was found at the stimulated sites and at the contralateral homologous and frontal sites. Since the effects of this rTMS protocol applied to rPPC have been assessed to be inhibitory by directly measuring cortical reactivity (Bagattini et al., 2015b), we can infer that, after rTMS, sites showing high coherence with rPPC are inhibited as well, i.e., they have a reduced cortical reactivity.

These results have implications both for basic knowledge about how different areas are functionally connected and for the possibility of applying this knowledge to the field of rehabilitation after brain damage. Firstly, they reveal a network of functionally coupled areas, both within and between hemispheres, oscillating coherently. Specifically, the present results revealed a functionally coupled fronto-parietal network within the right hemisphere, a network that has been found to be relevant for several cognitive functions (Marek and Dosenbach, 2018), such as for example attentional performance (Fellrath et al., 2016; Rogala et al., 2020), visuo-motor processes (Naranjo et al., 2006; Iturrate et al., 2018), visuospatial judgement (Guidali et al., 2023), and working memory and cognitive control (Gulbinaite et al., 2014; Bertaccini et al., 2022). Moreover, results have shown a functional coupling among homologous sites in the two hemispheres which are, thus, found to be connected in an excitatory manner, in line with previous research with different neuroimaging techniques in both healthy participants and neglect patients (Ricci et al., 2012; Bagattini et al., 2015b; Lunven et al., 2015; Killington et al., 2016; Ptak et al., 2020; Schintu et al., 2020, 2021). Importantly, cortico-cortical connectivity changes induced by TMS on resting-state connectivity MRI (Fox et al., 2012) have also been investigated. These studies have shown to be a powerful tool to reveal complex patterns of network modulation both within and across hemispheres. Unfortunately though, very few TMS protocols have been applied to areas other than the motor cortex (Han et al., 2023). Since network changes induced by TMS are specific for the stimulated area (Castrillon et al., 2020), a direct comparison with the present data lacks conclusive power. To our knowledge, the study most similar to the one presented here relates to a recently published TMS-EEG and MRI integrated approach (Esposito et al., 2022). In this study, the authors applied TMS, among other areas, to the right parietal cortex and analyzed structural and functional connectivity of the default mode network. The results have shown strong functional coupling of the right parietal sites with the homologous left parietal sites and with the fronto-central sites, in agreement with the present data. More research is needed to obtain a clearer picture of the complexity of cortico-cortical connectivity changes induced by TMS.

Interestingly, results also show higher inter-hemispheric coherence for low-frequency bands and higher intra-hemispheric coherence for high-frequency bands which can be explained by the underlying neural mechanisms and functional connectivity patterns in the brain (Varela et al., 2001; Buzsáki and Draguhn, 2004; Fries, 2015). Indeed, frequency power spectrum is topographically organized as a gradient along the antero-posterior axis: lower frequency bands dominate at back of the brain while higher frequency bands dominate at frontal sites (Niedermeyer, 1999). It is, thus, more likely that, by applying rTMS over the rPPC, more inter-hemispheric coherence can be found between posterior pairs of electrodes at low frequency bands. In the same vein, an increase of intra-hemispheric coherence for higher frequency bands is expected along the fronto-parietal network within the stimulated (right) hemisphere. Moreover, low-frequency bands, such as delta and theta waves, are associated with slower oscillations and are believed to reflect long-range connectivity and coordination between brain regions. These slow waves facilitate communication and synchronization between distant brain areas, including inter-hemispheric connections. Therefore, higher inter-hemispheric coherence in low-frequency bands suggests stronger coordination and information exchange between the two hemispheres. Consistently, a recent study showed that theta band effective inter-hemispheric connectivity between parietal regions sustained the leftward visuospatial advantage typically observed in neurologically healthy individuals (Bagattini et al., 2022). On the other hand, high-frequency bands, such as beta and gamma waves, are associated with faster oscillations and are believed to reflect local processing and information integration within specific brain regions. These fast waves are important for intra-hemispheric communication and are involved in various cognitive functions. Therefore, higher intra-hemispheric coherence in high-frequency bands indicates enhanced local processing and functional integration within each hemisphere.

Importantly, the present results provide implications for stroke rehabilitation, particularly addressing cognitive deficits such as neglect. Indeed, previous research has demonstrated the efficacy of contralesional low-frequency rTMS in reducing symptoms associated with these conditions. However, the neural mechanisms underlying these improvements have not been fully understood. The present study, by revealing the network of functionally connected areas engaged by rTMS and the nature of this connection, provided insights into the interplay between the hemispheres and the potential mechanisms underlying the genesis of the chronicity of symptoms. Specifically, the notion of the hyperactivity of the left hemisphere as a consequence of the release from the inhibition exerted by the right hemisphere (Kinsbourne, 1977) may not be the primary cause of the genesis of neglect symptoms. Rehabilitation protocols based on this notion, thus, may not be effective in reducing neglect symptoms in the long term (Müri et al., 2013; Carter and Barrett, 2023). Conversely, these data and those already present in literature on neglect patients could serve in devising ah-hoc protocols taking into account the complex interplay of intra- and inter-hemispheric networks abnormalities in neglect patients (Bartolomeo, 2014, 2019, 2021; Bartolomeo and Thiebaut de Schotten, 2016). Indeed, the hyperactivity in the left hemisphere of neglect patients could be the result of maladaptive plasticity (Nava and Röder, 2011; Altman et al., 2019) following brain lesion and not as the cause of the breakdown of the reciprocal inhibition exerts by the two hemispheres, which seems to be an oversimplification of brain dynamics (Berlucchi, 1983; Ptak et al., 2020). Accordingly, contralesional hyperactivity may not be adequate as the only determinant to implement rehabilitation protocols. Indeed, imaging studies have found no signs of hyperactivity in the left hemisphere of neglect patients tested in the acute phase (Vallar et al., 1988; Fiorelli et al., 1991; Perani et al., 1993; Umarova et al., 2011, 2014) whereas right hemisphere fronto-parietal network activity has been shown to predict neglect's severity (Corbetta and Shulman, 2002; He et al., 2007; Carter et al., 2010; Machner et al., 2020). These pieces of evidence, thus, support the conclusion that the reinstatement of neural functionality in both the left and right hemispheres contributes as a predictor of functional recovery from neglect (Corbetta et al., 2005; Cramer, 2008; Lunven et al., 2015; Umarova et al., 2016).

In conclusion, the study demonstrated that inhibitory rTMS applied to the right posterior parietal cortex modulated functional connectivity and coherence in the brain. The findings challenge the notion of reciprocal inhibition between the hemispheres and suggest a role for maladaptive plasticity in spatial neglect. The study's approach of investigating non-motor areas provides valuable insights into the intra- and inter-hemispheric effects of rTMS and its potential as a therapeutic intervention in stroke rehabilitation. Further research with larger sample sizes and stroke patients is warranted to confirm and expand upon these findings.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://doi.org/10.17605/OSF.IO/4VHM7.

The studies involving humans were approved by Ethical Committee of the Department of Neuroscience, Biomedicine and Movement Sciences. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

CM: Data curation, Formal analysis, Methodology, Software, Supervision, Validation, Visualization, Writing – review & editing, Investigation, Writing – original draft. SM: Data curation, Methodology, Supervision, Validation, Visualization, Writing – review & editing, Investigation, Software. CB: Supervision, Validation, Visualization, Writing – review & editing, Data curation, Investigation, Methodology, Software. JS-L: Supervision, Validation, Visualization, Writing – review & editing, Data curation, Investigation, Methodology, Software. SS: Methodology, Software, Supervision, Validation, Visualization, Writing – review & editing, Conceptualization, Funding acquisition, Project administration, Writing – original draft.