Removing artifacts from TMS-evoked EEG: A methods review and a unifying theoretical framework

A simulation study: comparing independent component analysis and signal-space projection - source-informed reconstruction for rejecting muscle artifacts evoked by transcranial magnetic stimulation

Introduction

The combination of transcranial magnetic stimulation (TMS) and electroencephalography (EEG) allows researchers to explore cortico-cortical connections. To study effective connections, the first few tens of milliseconds of the TMS-evoked potentials are the most critical. Yet, TMS-evoked artifacts complicate the interpretation of early-latency data. Data-processing strategies like independent component analysis (ICA) and the combined signal-space projection-source-informed reconstruction approach (SSP-SIR) are designed to mitigate artifacts, but their objective assessment is challenging because the true neuronal EEG responses under large-amplitude artifacts are generally unknown. Through simulations, we quantified how the spatiotemporal properties of the artifacts affect the cleaning performances of ICA and SSP-SIR.

Methods

We simulated TMS-induced muscle artifacts and superposed them on pre-processed TMS-EEG data, serving as the ground truth. The simulated muscle artifacts were varied both in terms of their topography and temporal profiles. The signals were then cleaned using ICA and SSP-SIR, and subsequent comparisons were made with the ground truth data.

Results

ICA performed better when the artifact time courses were highly variable across the trials, whereas the effectiveness of SSP-SIR depended on the congruence between the artifact and neuronal topographies, with the performance of SSP-SIR being better when difference between topographies was larger. Overall, SSP-SIR performed better than ICA across the tested conditions. Based on these simulations, SSP-SIR appears to be more effective in suppressing TMS-evoked muscle artifacts. These artifacts are shown to be highly time-locked to the TMS pulse and manifest in topographies that differ substantially from the patterns of neuronal potentials.

Discussion

Selecting between ICA and SSP-SIR should be guided by the characteristics of the artifacts. SSP-SIR might be better equipped for suppressing time-locked artifacts, provided that their topographies are sufficiently different from the neuronal potential patterns of interest, and that the SSP-SIR algorithm can successfully find those artifact topographies from the high-pass-filtered data. ICA remains a powerful tool for rejecting artifacts that are not strongly time locked to the TMS pulse.

Adapted Beamforming: A Robust and Flexible Approach for Removing Various Types of Artifacts from TMS-EEG Data

Electroencephalogram (EEG) recorded as response to transcranial magnetic stimulation (TMS) can be highly informative of cortical reactivity and connectivity. Reliable EEG interpretation requires artifact removal as the TMS-evoked EEG can contain high-amplitude artifacts. Several methods have been proposed to uncover clean neuronal EEG responses. In practice, determining which method to select for different types of artifacts is often difficult. Here, we used a unified data cleaning framework based on beamforming to improve the algorithm selection and adaptation to the recorded signals. Beamforming properties are well understood, so they can be used to yield customized methods for EEG cleaning based on prior knowledge of the artifacts and the data. The beamforming implementations also cover, but are not limited to, the popular TMS-EEG cleaning methods: independent component analysis (ICA), signal-space projection (SSP), signal-space-projection-source-informed-reconstruction method (SSP-SIR), the source-estimate-utilizing noise-discarding algorithm (SOUND), data-driven Wiener filter (DDWiener), and the multiple-source approach. In addition to these established methods, beamforming provides a flexible way to derive novel artifact suppression algorithms by considering the properties of the recorded data. With simulated and measured TMS-EEG data, we show how to adapt the beamforming-based cleaning to different data and artifact types, namely TMS-evoked muscle artifacts, ocular artifacts, TMS-related peripheral responses, and channel noise. Importantly, beamforming implementations are fast to execute: We demonstrate how the SOUND algorithm becomes orders of magnitudes faster via beamforming. Overall, the beamforming-based spatial filtering framework can greatly enhance the selection, adaptability, and speed of EEG artifact removal.

Evoked EEG Responses to TMS Targeting Regions Outside the Primary Motor Cortex and Their Test-Retest Reliability

Transcranial magnetic stimulation (TMS)-evoked electroencephalography (EEG) potentials (TEPs) provide unique insights into cortical excitability and connectivity. However, confounding EEG signals from auditory and somatosensory co-stimulation complicate TEP interpretation. Our optimized sham procedure established with TMS of primary motor cortex (Gordon in JAMA 245:118708, 2021) differentiates direct cortical EEG responses to TMS from those caused by peripheral sensory inputs. Using this approach, this study aimed to investigate TEPs and their test-retest reliability when targeting regions outside the primary motor cortex, specifically the left angular gyrus, supplementary motor area, and medial prefrontal cortex. We conducted three identical TMS-EEG sessions one week apart involving 24 healthy participants. In each session, we targeted the three areas separately using a figure-of-eight TMS coil for active TMS, while a second coil away from the head produced auditory input for sham TMS. Masking noise and electric scalp stimulation were applied in both conditions to achieve matched EEG responses to peripheral sensory inputs. High test-retest reliability was observed in both conditions. However, reliability declined for the 'cleaned' TEPs, resulting from the subtraction of evoked EEG response to the sham TMS from those to the active, particularly for latencies > 100 ms following the TMS pulse. Significant EEG differences were found between active and sham TMS at latencies < 90 ms for all targeted areas, exhibiting distinct spatiotemporal characteristics specific to each target. In conclusion, our optimized sham procedure effectively reveals EEG responses to direct cortical activation by TMS in brain areas outside primary motor cortex. Moreover, we demonstrate the impact of peripheral sensory inputs on test-retest reliability of TMS-EEG responses.

Measuring the accuracy of ICA-based artifact removal from TMS-evoked potentials

BACKGROUND

The analysis and interpretation of transcranial magnetic stimulation (TMS)-evoked potentials (TEPs) relies on successful cleaning of the artifacts, which typically mask the early (0-30 ms) TEPs. Independent component analysis (ICA) is possibly the single most utilized methodology to clean these signals.

OBJECTIVE

ICA-based cleaning is reliable provided that the input data are composed of independent components. Differently, in case the underlying components are to some extent dependent, ICA algorithms may yield erroneous estimates of the components, resulting in incorrectly cleaned data. We aim to ascertain whether TEP signals are suited for ICA.

METHODS

We present a systematic analysis of how the properties of simulated artifacts imposed on measured artifact-free TEPs affect the ICA results. The variability of the artifact waveform over the recorded trials is varied from deterministic to stochastic. We measure the accuracy of ICA-based cleaning for each level of variability.

RESULTS

Our findings indicate that, when the trial-to-trial variability of an artifact component is small, which can result in dependencies between underlying components, ICA-based cleaning biases towards eliminating also non-artifactual TEP data. We also show that the variability can be measured using the ICA-derived components, which in turn allows us to estimate the cleaning accuracy.

CONCLUSION

As TEP artifacts tend to have small trial-to-trial variability, one should be aware of the possibility of eliminating brain-derived EEG when applying ICA-based cleaning strategies. In practice, after ICA, the artifact component variability can be measured, and it predicts to some extent the cleaning reliability, even when not knowing the clean ground-truth data.

Causal evidence for a coordinated temporal interplay within the language network

Recent neurobiological models on language suggest that auditory sentence comprehension is supported by a coordinated temporal interplay within a left-dominant brain network, including the posterior inferior frontal gyrus (pIFG), posterior superior temporal gyrus and sulcus (pSTG/STS), and angular gyrus (AG). Here, we probed the timing and causal relevance of the interplay between these regions by means of concurrent transcranial magnetic stimulation and electroencephalography (TMS-EEG). Our TMS-EEG experiments reveal region- and time-specific causal evidence for a bidirectional information flow from left pSTG/STS to left pIFG and back during auditory sentence processing. Adapting a condition-and-perturb approach, our findings further suggest that the left pSTG/STS can be supported by the left AG in a state-dependent manner.

Effects of transcranial magnetic stimulation (TMS) current direction and pulse waveform on cortico-cortical connectivity: A registered report TMS-EEG study

Transcranial magnetic stimulation (TMS)-evoked potentials (TEPs) are a promising proxy for measuring effective connectivity, that is, the directed transmission of physiological signals along cortico-cortical tracts, and for developing connectivity-based biomarkers. A crucial point is how stimulation parameters may affect TEPs, as they may contribute to the general variability of findings across studies. Here, we manipulated two TMS parameters (i.e. current direction and pulse waveform) while measuring (a) an early TEP component reflecting contralateral inhibition of motor areas, namely, M1-P15, as an operative model of interhemispheric cortico-cortical connectivity, and (b) motor-evoked potentials (MEP) for the corticospinal pathway. Our results showed that these two TMS parameters are crucial to evoke the M1-P15, influencing its amplitude, latency, and replicability. Specifically, (a) M1-P15 amplitude was strongly affected by current direction in monophasic stimulation; (b) M1-P15 latency was significantly modulated by current direction for monophasic and biphasic pulses. The replicability of M1-P15 was substantial for the same stimulation condition. At the same time, it was poor when stimulation parameters were changed, suggesting that these factors must be controlled to obtain stable single-subject measures. Finally, MEP latency was modulated by current direction, whereas non-statistically significant changes were evident for amplitude. Overall, our study highlights the importance of TMS parameters for early TEP responses recording and suggests controlling their impact in developing connectivity biomarkers from TEPs. Moreover, these results point out that the excitability of the corticospinal tract, which is commonly used as a reference to set TMS intensity, may not correspond to the excitability of cortico-cortical pathways.

The Cerebellum and the Motor Cortex: Multiple Networks Controlling Multiple Aspects of Behavior

The cerebellum and its thalamic projections to the primary motor cortex (M1) are well known to play an essential role in executing daily actions. Anatomic investigations in animals and postmortem humans have established the reciprocal connections between these regions; however, how these pathways can shape cortical activity in behavioral contexts and help promote recovery in neuropathological conditions remains not well understood. The present review aims to provide a comprehensive description of these pathways in animals and humans and discuss how novel noninvasive brain stimulation (NIBS) methods can be used to gain a deeper understanding of the cerebellar-M1 connections. In the first section, we focus on recent animal literature that details how information sent from the cerebellum and thalamus is integrated into an broad network of cortical motor neurons. We then discuss how NIBS approaches in humans can be used to reliably assess the connectivity between the cerebellum and M1. Moreover, we provide the latest perspectives on using advanced NIBS approaches to investigate and modulate multiple cerebellar-cortical networks involved in movement behavior and plasticity. Finally, we discuss how these emerging methods have been used in translation research to produce long-lasting modifications of cerebellar-thalamic-M1 to restore cortical activity and motor function in neurologic patients.

Rhythmic TMS as a Feasible Tool to Uncover the Oscillatory Signatures of Audiovisual Integration

Multisensory integration is quintessential to adaptive behavior, with clinical populations showing significant impairments in this domain, most notably hallucinatory reports. Interestingly, altered cross-modal interactions have also been reported in healthy individuals when engaged in tasks such as the Sound-Induced Flash-Illusion (SIFI). The temporal dynamics of the SIFI have been recently tied to the speed of occipital alpha rhythms (IAF), with faster oscillations entailing reduced temporal windows within which the illusion is experienced. In this regard, entrainment-based protocols have not yet implemented rhythmic transcranial magnetic stimulation (rhTMS) to causally test for this relationship. It thus remains to be evaluated whether rhTMS-induced acoustic and somatosensory sensations may not specifically interfere with the illusion. Here, we addressed this issue by asking 27 volunteers to perform a SIFI paradigm under different Sham and active rhTMS protocols, delivered over the occipital pole at the IAF. Although TMS has been proven to act upon brain tissues excitability, results show that the SIFI occurred for both Sham and active rhTMS, with the illusory rate not being significantly different between baseline and stimulation conditions. This aligns with the discrete sampling hypothesis, for which alpha amplitude modulation, known to reflect changes in cortical excitability, should not account for changes in the illusory rate. Moreover, these findings highlight the viability of rhTMS-based interventions as a means to probe the neuroelectric signatures of illusory and hallucinatory audiovisual experiences, in healthy and neuropsychiatric populations.

Targeted Modulation of Human Brain Interregional Effective Connectivity With Spike-Timing Dependent Plasticity

OBJECTIVE

The ability to selectively up- or downregulate interregional brain connectivity would be useful for research and clinical purposes. Toward this aim, cortico-cortical paired associative stimulation (ccPAS) protocols have been developed in which two areas are repeatedly stimulated with a millisecond-level asynchrony. However, ccPAS results in humans using bifocal transcranial magnetic stimulation (TMS) have been variable, and the mechanisms remain unproven. In this study, our goal was to test whether ccPAS mechanism is spike-timing-dependent plasticity (STDP).

MATERIALS AND METHODS

Eleven healthy participants received ccPAS to the left primary motor cortex (M1) → right M1 with three different asynchronies (5 milliseconds shorter, equal to, or 5 milliseconds longer than the 9-millisecond transcallosal conduction delay) in separate sessions. To observe the neurophysiological effects, single-pulse TMS was delivered to the left M1 before and after ccPAS while cortico-cortical evoked responses were extracted from the contralateral M1 using source-resolved electroencephalography.

RESULTS

Consistent with STDP mechanisms, the effects on synaptic strengths flipped depending on the asynchrony. Further implicating STDP, control experiments suggested that the effects were unidirectional and selective to the targeted connection.

CONCLUSION

The results support the idea that ccPAS induces STDP and may selectively up- or downregulate effective connectivity between targeted regions in the human brain.

Measurement of the N170 during facial neuromuscular electrical stimulation (fNMES)

BACKGROUND

Studies on facial feedback effects typically employ props or posed facial expressions, which often lack temporal precision and muscle specificity.

NEW METHOD

Facial Neuromuscular Electrical Stimulation (fNMES) allows for a controlled influence of contractions of facial muscles, and may be used to advance our understanding of facial feedback effects, especially when combined with Electroencephalography (EEG). However, electrical stimulation introduces significant interference that can mask underlying brain dynamics. Whether established signal processing methods can allow for a reduction of said interference whilst retaining effects of interest, remains unexplored.

RESULTS

We addressed these questions focusing on the classic N170 visual evoked potential, a face-sensitive brain component: 20 participants viewed images of houses, and of sad, happy, and neutral faces. On half of the trials, fNMES was delivered to bilateral lower-face muscles during the presentation of visual stimuli. A larger N170 amplitude was found for faces relative to houses. Interestingly, this was the case both without and during fNMES, regardless of whether the fNMES artefact was removed or not. Moreover, sad facial expressions elicited a larger N170 amplitude relative to neutral facial expressions, both with and without fNMES.

COMPARISON WITH EXISTING METHODS

fNMES offers a more precise way of manipulating proprioceptive feedback from facial muscles, which affords greater diversity in experimental design for studies on facial feedback effects.

CONCLUSIONS

We show that the combining of fNMES and EEG can be achieved and may serve as a powerful means of exploring the impact of controlled proprioceptive inputs on various types of cognitive processing.

Transferability of cathodal tDCS effects from the primary motor to the prefrontal cortex: A multimodal TMS-EEG study

Neurophysiological effects of transcranial direct current stimulation (tDCS) have been extensively studied over the primary motor cortex (M1). Much less is however known about its effects over non-motor areas, such as the prefrontal cortex (PFC), which is the neuronal foundation for many high-level cognitive functions and involved in neuropsychiatric disorders. In this study, we, therefore, explored the transferability of cathodal tDCS effects over M1 to the PFC. Eighteen healthy human participants (11 males and 8 females) were involved in eight randomized sessions per participant, in which four cathodal tDCS dosages, low, medium, and high, as well as sham stimulation, were applied over the left M1 and left PFC. After-effects of tDCS were evaluated via transcranial magnetic stimulation (TMS)-electroencephalography (EEG), and TMS-elicited motor evoked potentials (MEP), for the outcome parameters TMS-evoked potentials (TEP), TMS-evoked oscillations, and MEP amplitude alterations. TEPs were studied both at the regional and global scalp levels. The results indicate a regional dosage-dependent nonlinear neurophysiological effect of M1 tDCS, which is not one-to-one transferable to PFC tDCS. Low and high dosages of M1 tDCS reduced early positive TEP peaks (P30, P60), and MEP amplitudes, while an enhancement was observed for medium dosage M1 tDCS (P30). In contrast, prefrontal low, medium and high dosage tDCS uniformly reduced the early positive TEP peak amplitudes. Furthermore, for both cortical areas, regional tDCS-induced modulatory effects were not observed for late TEP peaks, nor TMS-evoked oscillations. However, at the global scalp level, widespread effects of tDCS were observed for both, TMS-evoked potentials and oscillations. This study provides the first direct physiological comparison of tDCS effects applied over different brain areas and therefore delivers crucial information for future tDCS applications.

TMS combined with EEG: Recommendations and open issues for data collection and analysis

Transcranial magnetic stimulation (TMS) evokes neuronal activity in the targeted cortex and connected brain regions. The evoked brain response can be measured with electroencephalography (EEG). TMS combined with simultaneous EEG (TMS-EEG) is widely used for studying cortical reactivity and connectivity at high spatiotemporal resolution. Methodologically, the combination of TMS with EEG is challenging, and there are many open questions in the field. Different TMS-EEG equipment and approaches for data collection and analysis are used. The lack of standardization may affect reproducibility and limit the comparability of results produced in different research laboratories. In addition, there is controversy about the extent to which auditory and somatosensory inputs contribute to transcranially evoked EEG. This review provides a guide for researchers who wish to use TMS-EEG to study the reactivity of the human cortex. A worldwide panel of experts working on TMS-EEG covered all aspects that should be considered in TMS-EEG experiments, providing methodological recommendations (when possible) for effective TMS-EEG recordings and analysis. The panel identified and discussed the challenges of the technique, particularly regarding recording procedures, artifact correction, analysis, and interpretation of the transcranial evoked potentials (TEPs). Therefore, this work offers an extensive overview of TMS-EEG methodology and thus may promote standardization of experimental and computational procedures across groups.

Source-based artifact-rejection techniques for TMS-EEG

Neuronal electroencephalography (EEG) signals arise from the cortical postsynaptic currents. Due to the conductive properties of the head, these neuronal sources produce relatively smeared spatial patterns in EEG. We can model these topographies to deduce which signals reflect genuine TMS-evoked cortical activity and which data components are merely noise and artifacts. This review will concentrate on two source-based artifact-rejection techniques developed for TMS-EEG data analysis, signal-space-projection-source-informed reconstruction (SSP-SIR), and the source-estimate-utilizing noise-discarding algorithm (SOUND). The former method was designed for rejecting TMS-evoked muscle artifacts, while the latter was developed to suppress noise signals from EEG and magnetoencephalography (MEG) in general. We shall cover the theoretical background for both methods, but most importantly, we will describe some essential practical perspectives for using these techniques effectively. We demonstrate and explain what approaches produce the most reliable inverse estimates after cleaning the data or how to perform non-biased comparisons between cleaned datasets. All noise-cleaning algorithms compromise the signals of interest to a degree. We elaborate on how the source-based methods allow objective quantification of the overcorrection. Finally, we consider possible future directions. While this article concentrates on TMS-EEG data analysis, many theoretical and practical aspects, presented here, can be readily applied in other EEG/MEG applications. Overall, the source-based cleaning methods provide a valuable set of TMS-EEG preprocessing tools. We can objectively evaluate their performance regarding possible overcorrection. Furthermore, the overcorrection can always be taken into account to compare cleaned datasets reliably. The described methods are based on current electrophysiological and anatomical understanding of the head and the EEG generators; strong assumptions of the statistical properties of the noise and artifact signals, such as independence, are not needed.

Activation Patterns of Functional Brain Network in Response to Action Observation-Induced and Non-Induced Motor Imagery of Swallowing: A Pilot Study

Action observation (AO) combined with motor imagery (MI) was verified as more effective in improving limb function than AO or MI alone, while the underlying mechanism of swallowing was ambiguous. The study aimed at exploring the efficacy of AO combined with MI in swallowing. In this study, twelve subjects performed the motor imagery of swallowing (MI-SW) during magnetoencephalography (MEG) scanning, and trials were divided into three groups: the non-induced group (control group, CG), male AO-induced group (M-AIG), and female AO-induced group (F-AIG). We used event-related spectral perturbations (ERSPs) and phase locking value (PLV) to assess the degree of activation and connectivity of the brain regions during MI-SW in the three groups. The results showed that compared to CG, F-AIG and M-AIG significantly activated more brain regions in the frontoparietal, attention, visual, and cinguloopercular systems. In addition, M-AIG significantly activated the sensorimotor cortex compared to CG and F-AIG. For the brain network, F-AIG and M-AIG increased the diffusion of non-hub hot spots and cold hubs to the bilateral hemispheres which enhanced interhemispheric functional connectivity and information transmission efficiency in the MI-SW task. This study provided supporting evidence that AO induction could enhance the effect of MI-SW and supported the application of AO-induced MI-SW in clinical rehabilitation.

Identifying novel biomarkers with TMS-EEG - Methodological possibilities and challenges

Biomarkers are essential for understanding the underlying pathologies in brain disorders and for developing effective treatments. Combined transcranial magnetic stimulation and electroencephalography (TMS-EEG) is an emerging neurophysiological tool that can be used for biomarker development. This method can identify biomarkers associated with the function and dynamics of the inhibitory and excitatory neurotransmitter systems and effective connectivity between brain areas. In this review, we outline the current state of the TMS-EEG biomarker field by summarizing the existing protocols and the possibilities and challenges associated with this methodology.