Name: Neuroimaging-Guided TMS–EEG for Real-Time Cortical Network Mapping (Video) | JoVE Journal
Uploaded: 2025-06-13T13:35:00.000+00:00
Duration: 9 min 55 s
Description: Learn personalized transcranial magnetic stimulation combined with electroencephalography for real-time cortical network mapping using MRI-guided targeting. This protocol enhances precision in brain stimulation interventions.

ai_agent_IP

ai_agent

maintenance

instantTrialDurationMinutes

isBiopharma

enableNewMyJoveSectionDesign

educationFeatures

seoMetaApi

ai_copilot_institution_blacklist

ai_chat_institution_blacklist

freeAccessWindow

ssoRoleCapture

scheduledMaintenanceBanner

joveLabsFeatures

disable_autoplay_for_bots

extended_trial_duration

5 minutes

xWe would like to thank Profs. Silvia Casarotto, Marcello Massimini, and Mario Rosanova from the University of Milan for pioneering and promoting TMS&#8211;EEG mapping methodology. We also would like to acknowledge the contribution of the dozens of research participants that led to the development of our experimental protocol. Work on &#34;PlaStim: Plasticity Stimulation in the Treatment of Anhedonia&#34; is supported by Wellcome Leap as part of the Multi-Channel Psych Program. This project has also received funding from the European Research Council (ERC) under the European Union&#39;s Horizon 2020 Research and Innovation Programme (grant agreement No 81037).

PL is a consultant to Nexstim Plc for TMS&#8211;EEG applications and speech cortical mapping. RI has patents on TMS technology and has consulted Nexstim Plc on TMS.

This methodological pipeline introduces a novel approach for personalized brain mapping and targeting using multiple neuroimaging techniques. Individualized cortical maps are constructed by combining anatomical parcellation with functional and structural connectivity, which helps define the initial search region for the optimal stimulation target. Next, TMS&#8211;EEG mapping is employed to refine the target&#39;s location, orientation, and stimulation intensity, aiming to enhance the reactivity of early neuronal responses to TMS and reduce muscle, decay, and other artifacts. Cortical areas within the region of interest are systematically explored, adjusting the stimulation location, electric-field orientation, and intensity, while data quality is continuously monitored through real-time visualization of averaged TEPs. The final stimulation target is selected based on artifact-free responses with clearly identifiable early TEP components. This pipeline is especially useful for studies focused on exploring TMS-EEG-based biomarkers.
The pipeline consists of several essential and demanding steps. The first step involves defining the cortical area boundaries for TMS&#8211;EEG mapping. Without this, the target may be selected from an area with a different cytoarchitecture than the functionally relevant one. To address this, the Human Connectome Project multimodal parcellation36 is utilized, which divides brain regions based on a combination of structural and functional connectivity. The derived mask defines the outer boundaries of the area to be mapped with TMS&#8211;EEG. To further refine the cortical area of interest (e.g., DLPFC, a rather large brain region), voxels demonstrating functional connectivity to the sgACC (area 25)38,68,78, recently incorporated into clinical practices for MDD treatment69,79, are selected. From there, TMS&#8211;EEG mapping begins by targeting cortical regions with the strongest functional connectivity to the sgACC, ensuring the selected area exhibits high cortical reactivity without contamination from muscle or decay artifacts. If raw TEPs recorded after 20 pulses meet quality standards, the target is chosen for data collection. However, if significant artifacts are&#160;observed &#8212; a frequent issue in the prefrontal cortex &#8212; adjacent regions are&#160;systematically explored to identify a suitable alternative.
Real-time tractography may help in selecting TMS targets with large-scale structural connectivity, such as the spatially distributed networks implicated in MDD3,78, which are not limited to a single white matter tract. However, it is particularly valuable in cases involving known tracts, like the frontal aslant tract in the speech network, where it ensures precise targeting80. Although offline tractography can provide connectivity information that is useful for ROI selection, it is less effective for TMS&#8211;EEG mapping. Predefined cortical targets identified through offline methods may fail to elicit relevant or sufficiently strong EEG responses, may not effectively engage the desired network, or could result in discomfort and muscle artifacts, compromising data quality and necessitating parameter adjustments during the experiment. By dynamically offering connectivity insights, real-time tractography addresses these challenges, allowing for on-the-fly optimization of coil placement to ensure high-quality, reliable measurements and robust network engagement.
In this protocol, the intensity of TMS is adjusted based on EEG signals reflecting cortico-cortical excitability, while using RMT primarily as a starting point for intensity calibration. RMT plays a secondary role in determining the final intensity. This approach ensures the activation of a significant population of neurons within the target area, producing a strong and reliable signal. Once artifacts are minimized through coil adjustments and functional and structural validation of the target, stimulation intensity is further adjusted. Since this protocol aims to maximize the data obtained from TMS&#8211;EEG responses for biomarker identification, it is crucial to ensure that TEP amplitudes are adequate. When using RMT to define intensity, early responses typically do not exceed 4 &#181;V81 after preprocessing, which, combined with artifacts, reduces the quality of the neural signal. To overcome this, an approach developed by Casarotto et al.25 and refined by Tervo et al.82 is used, involving averaging the EEG signal in real-time after 20&#8211;30 TMS pulses. The goal is to achieve a clear, non-artifactual response of 6&#8211;10 &#181;V in a 10&#8211;50 ms window. As shown in Figures 3 and 5, closely located targets may require different stimulation intensities, leading to either insufficient or excessive neuronal activation.
Standard non-navigated targets, typically defined using distances from the motor hotspot or the 10&#8211;20 EEG system electrode locations, are widely used in rTMS treatments83,84,85,86. They have also been effective in TMS&#8211;EEG studies targeting the DLPFC, with biomarkers like later responses (e.g., N100) and long-interval intracortical inhibition (LICI) showing potential18,19,23,87,88. These non-personalized targets are simpler, cost-effective, and do not require MRI, neuronavigation, or specialized personnel for target identification, and some have been shown to reliably locate the DLPFC85,86. However, the lack of personalization may contribute to lower remission rates in treatments89,90. Moreover, cranial muscles in the area may obscure early responses due to artifacts that prevent intensity adjustment. Personalized intensity adjustments in TMS may therefore&#160;improve treatment effectiveness.
This protocol advances a more personalized approach to brain stimulation that provides better insight into cortical reactivity and connectivity compared to standard non-personalized approaches. However, it demands substantial resources, including individual MRI scans, specialized TMS and EEG equipment, and personnel with expertise in neuroimaging and neurophysiology. Moreover, the proposed approach heavily depends on the subjective judgment of the TEP quality and optimal stimulation parameters. Despite these limitations, the quality of the TMS&#8211;EEG data is critical for biomarker development, justifying the use of this approach. Once reliable biomarkers are identified, the procedure may be simplified to focus on optimizing relevant time windows and minimizing artifacts. In the future, as TMS technology advances (e.g., multi-locus91 and multi-channel92 TMS), the resource requirements for personalized stimulation are expected to decrease, potentially making it the new standard for clinical use.

The human brain is characterized by distinct large-scale structural and functional networks. The intricate interactions within and between the networks enable the brain's remarkable information-processing power, supporting complex cognitive, sensory, and motor processes. Disruptions in these networks, whether due to structural damage, functional dysregulation, or impaired connectivity, can lead to a wide range of neurological and psychiatric conditions1,2,3,4,5,6, including motor deficits, cognitive impairments, and mood disorders. Accurately monitoring the progression of such network-related disorders and identifying optimal treatment strategies require methods that can selectively perturb specific networks and reliably quantify their functional integrity.
Transcranial magnetic stimulation, when delivered efficiently, can non-invasively perturbs neuronal populations and even entire cortical networks7,8,9,10,11,12,13,14,15, allowing their assessment through measures such as motor responses or behavioral changes. TMS in combination with electroencephalography (EEG) allows one to measure TMS-evoked potentials (TEPs) – trial-averaged EEG responses time-locked to the TMS pulse. TEP waveforms recorded beneath the coil capture the immediate cortical excitability of the stimulated area, while responses from distant regions provide insights into effective connectivity between brain areas, establishing TMS–EEG as a valuable tool for investigating cortical networks. TEPs have already shown significant potential as biomarkers for various neurological9,16,17 and psychiatric conditions18,19,20,21,22,23. Moreover, concurrent TMS–EEG allows for personalized stimulation intensities to be decided based on the local cortical reactivity of the individual as measured by EEG24,25,26,27,28. However, the challenge of this method, especially when used for biomarker development22,24, is two-fold: a) to maximize information gain, the quality of the collected TEPs must be high, requiring mapping of the cortical areas to locate a target with minimal artifacts and good reactivity, and b) the response has to come from the network of interest and not from adjacent cortical sites. Low TEP quality is a common issue, often manifesting as noise, artifacts, multisensory responses, or early responses that fail to exceed the power of ongoing oscillatory activity. This poor signal quality is typically identified only after offline processing, often long after data collection, making it difficult or impossible to address.
Navigated TMS (nTMS)29, utilizing structural magnetic resonance imaging to target specific brain areas, has been available for clinical practice and research for more than 20 years. It has been crucial in mapping motor and speech cortical areas for presurgical evaluation30,31,32 and for targeting areas that do not involve motor, visual, or behavioral responses by combining it with EEG9,33,34,35. However, it can be unspecific neuroanatomically (i.e., the borders between cortical regions are not clear) and functionally, as individual variability in brain function cannot be detected using structural MRI. As such, current TMS targeting approaches, commonly reliant on the location of the motor cortex or other anatomical landmarks, often fail to stimulate functionally relevant cortical areas, limiting both treatment efficacy and the discovery of reliable biomarkers. To overcome these challenges, a comprehensive framework has been developed that integrates the use of other MRI modalities, such as functional and diffusion MRI, with precise cortical mapping for artifact-free TEP acquisition. This approach enables reliable and reproducible measurements, particularly of early TEP components, which are highly sensitive to subtle neurophysiological changes. By enhancing the sensitivity of TEP analysis, this framework facilitates robust correlations with clinical phenotypes, establishing a strong foundation for biomarker exploration in neuropsychiatric disorders.
The proposed pipeline is designed to map entire cortical areas, combining anatomical parcellation36 (Figure 1A), functional MRI (fMRI)-based functional connectivity37,38 (Figure 1B), and real-time tractography39 (Figure 1C). Anatomical parcellation allows the delineation of cortical areas corresponding to a specific brain network for an individual, using a predefined anatomical template or brain atlas nonlinearly registered to individual brain anatomy. Functional connectivity provides statistical metrics that measure the correlation between the physiological activity of different brain regions, as reflected by the blood-oxygen-level-dependent (BOLD) signal40. Lastly, real-time tractography is a method based on diffusion MRI (dMRI) that estimates and displays streamlines describing potential structural brain connections in real-time. With the combination of these methods, cortical areas can be targeted based on their structural and functional connectivity during the TMS–EEG mapping. This enables the refinement of stimulation parameters in real-time to ensure artifact-free TEP acquisition.
The proposed novel methodology addresses the combined challenges of TMS targeting and TEP acquisition, paving the way for more precise biomarker identification and effective neuromodulation strategies in clinical and research settings.
<img alt="Brain activity mapping; images A-C; MRI and tractography; neural connectivity analysis; research use." class="xfigimg" src="/files/ftp_upload/67339/67339fig01.jpg" data-altupdatedat="1788304762000" data-alt="Figure 1"> 
Figure 1: MRI-derived information for a representative subject. (A) Anatomical parcellation outlining the targeted dorsolateral prefrontal cortex (DLPFC). (B) Functional connectivity map showing anti-correlation with the subgenual cingulate cortex (lighter colors indicate stronger anti-correlation). (C) Real-time tractography displaying streamlines seeded from the targeted area. <a href="https://www.jove.com/files/ftp_upload/67339/67339fig01large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>

<table><tbody><tr><td>Abrasive gel</td><td>H + H Medical Devices</td><td>OneStep AbrasivPlus</td><td></td></tr><tr><td>Conductive gel</td><td>ECI</td><td>ElectroGel</td><td></td></tr><tr><td>DESIGNER</td><td></td><td>version 2.0.0</td><td>https://nyu-diffusionmri.github.io/DESIGNER-v2</td></tr><tr><td>EEG amplifiers</td><td>Brain Products GmbH</td><td>BrainAmp DC amplifiers</td><td>5000 Hz sampling frequency</td></tr><tr><td>EEG caps</td><td>EASYCAP GmbH</td><td>64 channel BrainCap for TMS</td><td>Sintered Ag/AgCl-Electrodes</td></tr><tr><td>EMG amplifier</td><td>Nexstim Plc</td><td>Nexstim EMG</td><td></td></tr><tr><td>Figure-of-eight coil</td><td>Nexstim Plc</td><td>Cooled Coil</td><td>70-mm diameter air-cooled coil</td></tr><tr><td>fmriprep</td><td></td><td>Version 22.1.0</td><td>https://fmriprep.org</td></tr><tr><td>FreeSurfer</td><td></td><td>Version 7.3.2</td><td>https://surfer.nmr.mgh.harvard.edu</td></tr><tr><td>MATLAB</td><td>MathWorks</td><td>Version R2021b</td><td>https://www.mathworks.com</td></tr><tr><td>MRI scanner</td><td>Siemens Healthcare</td><td>3T Siemens MAGNETOME Skyra</td><td>T1-weigted structural images: MPRAGE, 1-mm istropic voxels, TR = 2530 ms, and TE = 3.42 ms;&lt;br/&gt; fMRI: EPI, 3-mm isotropic voxels, TR= 1250 ms, TE = 3 ms, multiband acceleration factor 3, flip angle 65;&lt;br/&gt; dMRI: 100 different gradient directions total (b = 1500 and 3000 s/mm2), 12 non-diffusion-weighted (b = 0&amp;nbsp; s/mm&lt;sup&gt;2&lt;/sup&gt;) volumes with anterior&amp;ndash;posterior phase-encoding direction, 1 non-diffusion-weighted volume with the reverse phase-encoding direction (posterior&amp;ndash;anterior), 2-mm isotropic voxels, TR = 4100 ms, TE = 105 ms.</td></tr><tr><td>MRtrix3</td><td></td><td>Version 3.0.4</td><td>https://www.mrtrix.org</td></tr><tr><td>nilearn</td><td></td><td>Version 0.10.3</td><td>https://nilearn.github.io</td></tr><tr><td>Noise-cancelling earphones&amp;nbsp;</td><td>Etymotic Research Inc</td><td>ER3C Insert Earphones</td><td></td></tr><tr><td>Python</td><td></td><td>Version 3.9.19</td><td>https://www.python.org</td></tr><tr><td>realTimeTractogramVisualizer</td><td></td><td>Version 0.1</td><td>https://github.com/baranaydogan/realTimeTractogramVisualizer</td></tr><tr><td>rt-TEP</td><td></td><td></td><td>https://github.com/iTCf/rt-TEP</td></tr><tr><td>Surface EMG elecrodes</td><td>Ambu A/S</td><td>Ambu Neuroline 720</td><td></td></tr><tr><td>Surface EMG ground electrode</td><td>Ambu A/S</td><td>Ambu Neuroline Ground</td><td></td></tr><tr><td>TAAC</td><td></td><td></td><td>www.github.com/iTCf/TAAC</td></tr><tr><td>TMS system</td><td>Nexstim Plc</td><td>Nexstim NBT 2.2.4</td><td>Biphasic pulse, jittered interstimulus interval 2&amp;ndash;2.4 s</td></tr></tbody></table>

neuroimaging-guided tms&#8211;eeg for real-time cortical network mapping

Neuroimaging-Guided TMS&#8211;EEG for Real-Time Cortical Network Mapping

The cerebral cortex is organized into structurally and functionally segregated networks, enabling the human brain to process information highly efficiently. Transcranial magnetic stimulation (TMS), in combination with electroencephalography (EEG), offers a non-invasive approach to probing brain networks, revealing cortical excitability and causal connectivity. However, this method faces two significant challenges: (a) ensuring the quality of TMS-evoked potentials (TEPs) to maximize information gain, often requiring comprehensive cortical mapping, and (b) eliciting the response from the network of interest and not from adjacent cortical sites.
Existing TMS targeting approaches frequently fail to precisely stimulate functionally relevant cortical areas, hindering treatment efficacy and the identification of biomarkers. The presented protocol integrates precise cortical mapping to acquire artifact-free TEPs, enabling reliable and reproducible measurements of early TEP components. This precision improves sensitivity to subtle neurophysiological variations and strengthens correlations with clinical phenotypes, supporting biomarker discovery in neuropsychiatric disorders.
The proposed protocol utilizes structural, functional, and diffusion magnetic resonance imaging (MRI) to identify cortical patches belonging to the network of interest. Anatomical parcellation, functional connectivity, and real-time tractography are applied to locate areas with strong connectivity to other brain regions associated with the target network. The resulting personalized cortical clusters define the initial stimulation targets.
TMS&#8211;EEG mapping is then employed to optimize TMS parameters by localizing cortical areas with high excitability, enhancing neuronal response magnitude while reducing non-neuronal noise, including muscle artifacts, decay, and other confounding factors affecting early TMS&#8211;EEG responses. A systematic exploration of the cortical mantle is conducted, adjusting stimulation location, orientation, and intensity, with continuous data quality monitoring through real-time visualization of averaged TEPs. TMS parameters producing artifact-free responses with clearly discernible early TEP components are selected for data collection.
This article introduces the neuroimaging-guided TMS&#8211;EEG mapping technique and highlights the methodological advancements and benefits achievable through its application.

The described protocol was used to collect TMS–EEG data from the left dorsolateral prefrontal cortex (DLPFC) of a healthy volunteer. The study was approved by The Coordinating Ethics Committee of Helsinki University Hospital and carried out in accordance with the Declaration of Helsinki. The subject signed an informed consent form. For the data acquisition details, refer to Table of Materials.
Left DLPFC is a common area of interest, used in clinical practice for repetitive TMS treatment of major depressive disorder (MDD)65. A growing body of literature highlights the importance of the functional connectivity of the subgenual anterior cingulate cortex (sgACC) with DLPFC in the pathophysiology of MDD3,66,67. In particular, there are indications that the TMS treatment delivered over the areas of the left DLPFC showing strong anti-correlation with the left sgACC might have higher efficacy38,68,69,70. Following these findings, it is hypothesized that TMS–EEG data recorded from the targets functionally anticorrelated with sgACC BOLD activity is going to be most informative for further biomarker development.
The structural MRI mask for DLPFC was defined as a combination of the regions 9a, 9p, 9-46d, 46, a9-46v, and p9-46v from the HCPMMP36, encompassing Brodmann areas 9 and 46. The sgACC mask was region 25, equivalent to Brodmann area 25. Following the described protocol, seed-to-voxel correlation maps were derived using fMRI data and selected only voxels belonging to the DLPFC ROI. The maps were then thresholded to only display voxels with negative correlation; the absolute value of the correlation was taken to simplify the visualization. The resulting clusters were overlaid on the structural MRI in the neuronavigation during the TMS–EEG experiment (Figure 2A).
The initial TMS–EEG mapping targets were selected from regions that exhibited the strongest functional anticorrelation with the sgACC. These targets were then iteratively refined through the TMS–EEG mapping procedure to identify the most promising stimulation sites based on TEP quality and their proximity to cortical patches with the highest anticorrelation to the sgACC. The finalized set of potential stimulation sites was subsequently analyzed using real-time tractography. Since there is no known direct pathway from the sgACC to the DLPFC, indirect white matter pathways had to be considered. In cases where specific direct or indirect pathways are known a priori, these should be prioritized. For the sgACC, the large-scale brain network affected by MDD3 was taken into account during the tractography-driven target selection. The structural connectivity of the TMS target was also evaluated, with areas showing broad connectivity to other brain regions — such as the ventromedial prefrontal cortex, Broca's area, or the parietal lobe — being given higher priority.
Following the mapping protocol, the target, orientation, and intensity of the stimulation were chosen based on the resulting TEPs being larger than 6 µV and minimally affected by artifacts. Each optimization iteration involved recording 20 pulses with minimal preprocessing. The final optimized target was then used to collect 300 pulses for data analysis.
The minimal preprocessing pipeline, designed to simulate real-time rt-TEP use, involved the following steps: 
1. Division of data into epochs 
2. Baseline correction 
3. Removal of the TMS artifact by replacing data from -2 to 10 ms around TMS pulse by zeros 
4. Removal of the bad channels 
5. Re-referencing to the average reference 
6. A low-pass filter at 80 Hz and a notch filter at 48–52 Hz were applied when necessary to address line and high-frequency noise.
The full preprocessing pipeline followed the procedures outlined in Mutanen et al., 202471. For further details on TMS–EEG preprocessing, refer to Hernandez-Pavon et al., 202272, Rogasch et al., 201773, Mutanen et al., 201874, and Mutanen et al., 201675. In summary, the process included: 
1. Division of data into epochs 
2. Baseline correction 
3. TMS artifact removal between -2 and 10 ms around the TMS pulse, with cubic interpolation using 5 ms of data before and after the removed interval 
4. Removal of bad channels (4 in total: Fpz, F1, FT10, TP9, PO4) 
5. Removal of bad trials 
6. Removal of drifts with robust detrending 
7. ICA for removal of ocular artifacts 
8. Baseline correction 
9. Application of the SOUND74 algorithm to suppress extracranial noise, followed by re-referencing to the average reference 
10. SSP-SIR75 algorithm to remove TMS-evoked muscle artifacts 
11. Low-pass filtering at 80 Hz and a notch filter at 48–52 Hz 
12. Snipping the ends of the time window to remove possible edge effects 
13. Additional removal of bad trials to ensure residual high-noise trials are excluded (22 out of 300 rejections total)
The experimental procedure started with the identification of the resting motor threshold. Subject's RMT was 41% MSO.
During the experimental procedure, primarily the electrodes over the region of interest (AF3, F1, F3, F5, FC3) (Figure 2B) were examined to assess the quality of the resulting TEPs. Other electrodes were continuously monitored for low impedance, and the impedance was lowered as needed, following the procedure outlined in step 3.1.10. The early TMS responses within the 10 to 60 ms time window, which reflect the direct activation of the cortex by the induced electric field, were used to evaluate the quality of the produced TEPs.
The mapping procedure (see Figure 3) began with one of the fMRI clusters (Figure 3A) and an investigation of the structural connectivity of the target area (Figure 3B), which primarily showed streamlines to the homologous area in the contralateral hemisphere and frontal pole.
For the stimulation, the coil was oriented at 45 degrees to the midline. The stimulation intensity was set at 49% MSO, corresponding to 96 V/m estimated E-field maximum at the hotspot and 120% RMT. In Figure 3C, a potential TEP-like response is observed in the F3, F1, and FC3 electrodes. The AF3 electrode exhibits a large-amplitude ringing artifact, while F5 is affected by a small muscle artifact recognizable by the high-frequency high-amplitude peak right after the TMS pulse76,77. To minimize the ringing artifact during recording, a net cap pressing the electrodes down to reduce movement or a thin foam layer beneath the coil can be used (see Hernandez-Pavon et al., 202328). Although such artifacts can often be removed through filtering, since only AF3 was contaminated, it was excluded from the average reference to avoid unnecessary filtering, resulting in the data shown in Figure 3D.
Focusing on the F3 electrode (Figure 3F), the signal appears largely unaffected by the muscle artifact observed in F5. The characteristic large-amplitude waveform recovers around 15 ms, and a 5-microvolt deflection between 22 ms and 32 ms is likely a genuine cortical response to TMS. Filtering the signal (Figure 3E, G) confirms that the amplitude remains unaffected by noise in other channels.
To investigate whether the muscle artifact in the F5 channel could be reduced through coil rotation, all other stimulation parameters were kept constant (Figure 4), and the coil was first rotated to a posterior-anterior orientation (Figure 4A), where the E-field maximum was estimated to be 80 V/m and then to a lateral-medial direction (Figure 4E), with 101 V/m E-field maximum. A sharp increase in muscle activation was observed with the posterior-anterior coil orientation (Figure 4B,C), which was reported as uncomfortable by the subject. The lateral-medial orientation produced a signal similar to the one shown in Figure 3, but with a larger amplitude, resulting in an early 12 µV TEP response at the F3 electrode (Figure 4F,G), which allows a reduction in stimulation intensity. The lack of excessive muscle, decay, or ringing artifacts, along with the size of the early TEP response, makes this combination of stimulation parameters a promising candidate for the data collection.
Figure 5 shows the effect of stimulation intensity on TEP responses in low-excitability target. The second fMRI cluster (Figure 5A) and its structural connectivity (Figure 5B) were then investigated. Similarly to the first target, the structural connections seem to be limited to the frontal lobes. It is noteworthy that despite the distance between the two fMRI clusters being 28 mm, the stimulation intensity used for the previous targets, 49% MSO (77 V/m, 120% RMT), produced no discernible TMS response, highlighting the importance of TEP-based selection of stimulation intensity. In Figure 5B,C, the TMS response at 55% MSO (89 V/m, 134% RMT) is shown. It is important to note that the large deflection observed after 10 ms is not a genuine TMS response but rather a continuation of the muscle artifact recovery. Therefore, only signals occurring later than 25 ms after the pulse should be considered, resulting in an amplitude of 4 µV between 26 ms and 60 ms. The intensity was further increased to 60% MSO (97 V/m, 146% RMT) in an attempt to obtain distinct earlier responses (Figure 5F,G). Compared to the 55% MSO intensity, an expected increase in the amplitude of the muscle artifact was observed, but there was little to no increase in the TEP amplitude. Based on the observed waveform and the need for a higher stimulation intensity, this target would be considered to exhibit lower cortical excitability compared to the previous targets, making it less informative for data collection. However, a target with similar properties might be preferable for treatment protocols.
In total, 16 distinct combinations of stimulation parameters were investigated using a 20-trial mapping procedure. The total duration of the procedure, excluding the EEG and EMG preparation (which took approximately 30 min), was 2 h 35 min. The final stimulation target can be seen in Figure 6A. While being very close to the initial target (Figure 3A), this target exhibits much more widespread structural connectivity (Figure 6B), potentially providing more information on the signal propagation from the hotspot to other cortical regions. The stimulation intensity was kept at 49% MSO (102 V/m, 120% RMT).
The minimal preprocessing of the first 20 pulses collected from the final stimulation target is shown in Figure 6C,D. Large amplitude ringing contaminated the F1 electrode, which was subsequently rejected. However, residual noise remained present in other electrodes even after this rejection. Despite the presence of the ringing artifact, the absence of the muscle artifact within the region of interest allowed for considering the latencies after 16 ms as a genuine neural signal. After filtering, the amplitude of the early component between 17 ms and 35 ms was 9 µV.
Figure 6E,F show the result of the full preprocessing pipeline applied to the 300-pulse dataset. The resulting TEP closely resembles the waveform derived from 20 pulses (Figure 6D), demonstrating the relevance of real-time monitoring of the TEPs. The early response between 20 ms and 40 ms is 6 µV, demonstrating the expected reduction in amplitude due to the preprocessing.
<img alt="Brain activity visualization; MRI with neural mapping; EEG channel locations diagram." class="xfigimg" src="/files/ftp_upload/67339/67339fig02.jpg" data-altupdatedat="1788304762000" data-alt="Figure 2"> 
Figure 2: Neuronavigation and electrode setup. (A) Subject's 3D MRI-based head model overlaid with fMRI-derived connectivity. (B) Schematic of EEG electrode placement, with electrodes of interest under the stimulation coil marked in orange. Note that the locations of the electrodes are digitized on the 3D head model (step 3.7.4). <a href="https://www.jove.com/files/ftp_upload/67339/67339fig02large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Brain activity analysis; EEG signals, electrode potential graphs; neural response visualization." class="xfigimg" src="/files/ftp_upload/67339/67339fig03.jpg" data-altupdatedat="1788304762000" data-alt="Figure 3"> 
Figure 3: An example of minimal TMS–EEG data processing. (A) The neuronavigation display shows the stimulation site and coil orientation. The red arrow indicates the stronger direction of the biphasic pulse, and the blue arrow indicates the weaker direction. Stimulation intensity is expressed as a percentage of maximum stimulator output (MSO) and resting motor threshold (RMT). (B) Real-time tractography at the stimulation location. (C) Raw TMS–EEG data from electrodes beneath the stimulation coil. (D) TMS–EEG data after noisy channel rejection, with rejected channels replaced by zeros. (E) TMS–EEG data after filtering. (F) Zoomed-in view of the F3 electrode after bad channel rejection. (G) Zoomed-in view of the F3 electrode after filtering. <a href="https://www.jove.com/files/ftp_upload/67339/67339fig03large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Brain potential analysis; diagrams of MSO at 80 V/m and 101 V/m with EEG signal results." class="xfigimg" src="/files/ftp_upload/67339/67339fig04.jpg" data-altupdatedat="1788304762000" data-alt="Figure 4"> 
Figure 4: Effect of stimulation orientation on the muscle artifact. (A) Stimulation target with posterior-anterior coil orientation and corresponding intensity. (B) Raw data for posterior-anterior orientation, showing muscle artifact contamination. (C) Data from F3 electrode with large muscle artifact. (D) Real-time tractography corresponding to the stimulation location. (E) Stimulation target with lateral-medial orientation and corresponding intensity. (F) Raw data for lateral-medial orientation, showing clean TEPs. (G) TEP from the F3 electrode, exhibiting a large, artifact-free early response. <a href="https://www.jove.com/files/ftp_upload/67339/67339fig04large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Brain stimulation, EEG analysis; graphs show potential changes, 55%/60% MSO, neural pathways." class="xfigimg" src="/files/ftp_upload/67339/67339fig05.jpg" data-altupdatedat="1788304762000" data-alt="Figure 5"> 
Figure 5: Effect of stimulation intensity on TEP responses in low-excitability target. (A) Stimulation target location and intensity. (B) Raw data recorded at 55% MSO. C) F3 electrode data, lacking distinguishable early TEP components. (D) Real-time tractography corresponding to the stimulation location. (E) Stimulation intensity increased while keeping the target constant. (F) Raw data recorded at 60% MSO. (G) F3 electrode data showing muscle artifacts due to higher stimulation intensity. <a href="https://www.jove.com/files/ftp_upload/67339/67339fig05large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Neural stimulation effects on brain: electric field map, fiber tracking, EEG potential graphs." class="xfigimg" src="/files/ftp_upload/67339/67339fig06.jpg" data-altupdatedat="1788304762000" data-alt="Figure 6"> 
Figure 6: Effect of minimal and full preprocessing pipelines on the TMS–EEG data. (A) Final stimulation target and intensity. (B) Real-time tractography for the stimulation location. (C) Raw data averaged over 20 trials. (D) 20-trial average TEP at the F3 electrode. (E) Fully preprocessed TEPs averaged over 300 trials. (F) Fully preprocessed 300-trial average TEP at the F3 electrode. <a href="https://www.jove.com/files/ftp_upload/67339/67339fig06large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>

A novel protocol for personalized transcranial magnetic stimulation in combination with electroencephalography (TMS&#8211;EEG) mapping of cortical networks, informed by multiple magnetic resonance imaging (MRI) data modalities.

The cerebral cortex is organized into structurally and functionally segregated networks, enabling the human brain to process information highly ...

neuroimaging-guided-tmseeg-for-real-time-cortical-network-mapping

Research

JoVE Journal

Neuroscience

We use magnetoencephalography (MEG) and electroencephalography (EEG) to map brain areas involved in the processing of simple sensory stimuli.

Functional Mapping with Simultaneous MEG and EEG

We present a method for collecting electrocorticographic signals for research purposes from humans who are undergoing invasive epilepsy monitoring.  We show how to use the BCI2000 software platform for data collection, signal processing and stimulus presentation.  Specifically, we demonstrate SIGFRIED, a BCI2000-based tool for real-time functional brain mapping.

Recording Human Electrocorticographic (ECoG) Signals for Neuroscientific Research and Real-time Functional Cortical Mapping

We use magneto- and electroencephalography (MEG/EEG), combined with anatomical information captured by magnetic resonance imaging (MRI), to map the dynamics of the cortical network associated with auditory attention.

Mapping Cortical Dynamics Using Simultaneous MEG/EEG and Anatomically-constrained Minimum-norm Estimates: an Auditory Attention Example

Source: Stefanou, M., et.al. Brain State-dependent Brain Stimulation with Real-time Electroencephalography-Triggered Transcranial Magnetic Stimulation. J. Vis. Exp. (2019)This video demonstrates the process of real-time EEG-triggered transcranial magnetic stimulation (TMS) to study cortical excitability. The procedure involves positioning the TMS coil over the motor cortex, synchronizing stimulation with specific phases of the EEG signal, and recording motor-evoked potentials (MEPs) to explore...

Real-Time Electroencephalography-Triggered Transcranial Magnetic Stimulation for Cortical Excitation

Source: Shafi, M. M., et al. A Multimodal Imaging- and Stimulation-based Method of Evaluating Connectivity-related Brain Excitability in Patients with Epilepsy. J. Vis. Exp. (2016)The video demonstrates the procedure for using transcranial magnetic stimulation (TMS) and electroencephalography (EEG) to assess neuronal excitability in patients with epilepsy. It details the steps for aligning the patient’s head with a pre-recorded brain scan, positioning the TMS coil over a functionally connected...

Assessing Brain Hyperexcitability with Transcranial Magnetic Stimulation and Electroencephalography

This paper describes a method for collecting and analyzing electroencephalography (EEG) data during concurrent transcranial magnetic stimulation (TMS) guided by activations revealed with functional magnetic resonance imaging (fMRI). A method for TMS artifact removal and extraction of event related potentials is described as well as considerations in paradigm design and experimental setup.

Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation

This study is designed to test the hypothesis that neuronavigational system-guided transcranial magnetic stimulation has higher accuracy for targeting the intended target as demonstrated by eliciting a greater degree of virtual aphasia in healthy subjects, measured by delay in reaction time to picture naming.

Neuronavigation-guided Repetitive Transcranial Magnetic Stimulation for Aphasia

Resting-state functional-connectivity MRI has identified abnormalities in patients with a wide range of neuropsychiatric disorders, including epilepsy due to malformations of cortical development. Transcranial Magnetic Stimulation in combination with EEG can demonstrate that patients with epilepsy have cortical hyperexcitability in regions with abnormal connectivity.

A Multimodal Imaging- and Stimulation-based Method of Evaluating Connectivity-related Brain Excitability in Patients with Epilepsy

The protocol presented here is for TMS-EEG studies utilizing intracortical excitability test-retest design paradigms. The intent of the protocol is to produce reliable and reproducible cortical excitability measures for assessing neurophysiological functioning related to therapeutic interventions in the treatment of neuropsychiatric diseases such as major depression.

Combined Transcranial Magnetic Stimulation and Electroencephalography of the Dorsolateral Prefrontal Cortex

This paper describes real-time electroencephalography-triggered transcranial magnetic stimulation to study and modulate human brain networks.

Brain State-dependent Brain Stimulation with Real-time Electroencephalography-Triggered Transcranial Magnetic Stimulation

Navigated repetitive transcranial magnetic stimulation is a highly efficient non-invasive tool for mapping speech-related cortical areas. It helps in designing brain surgery and speeds up the direct cortical stimulation conducted during the surgery. This report describes how to perform speech cortical mapping reliably for preoperative evaluation and research.

Study Design for Navigated Repetitive Transcranial Magnetic Stimulation for Speech Cortical Mapping

Learn personalized transcranial magnetic stimulation combined with electroencephalography for real-time cortical network mapping using MRI-guided targeting. This protocol enhances precision in brain stimulation interventions.

Neuroimaging-Guided TMS–EEG for Real-Time Cortical Network Mapping

Explore More Videos