Combining techniques

Temporal vs Spatial Resolution

Here the methods are plotted based on their average spatial (ms) and temporal (mm) resolution. The lower values indicate more precise results.

 I made these graphs using R studio and averaged the values of temporal and spatial resolution of each technique. 

FMRI-EEG

Comparison:

( ++ / -- = positive / negative aspects) 

++ EEG: Offers high temporal resolution (millisecond range) and provides a direct measure of neural activity. 

++ fMRI: Has high spatial resolution, it can reach  down to 1-millimeter, at the cost of whole-brain coverage. 

-- EEG:

1) Sensitive only to post-synaptic potentials generated in the superficial layers of the cortical layers (e.g., won't detect activity from the striatum or hippocampus). 

2) The spatial resolution of EEG is very low.

3) It is nearly impossible to accurately reconstruct unique intracranial current sources from the EEG signal.

4) Currents tangential to the skull contribute minimally to the signal.

-- fMRI: 

1) Low temporal resolution, as it measures the blood oxygenation (BOLD signal), which lag behind neural activity by 6–10 seconds.

2) Provides an indirect measure of neural activity, and the signal can be influenced by physiological noise unrelated to neuronal processes.

Combining fMRI - EEG

Combining these two methods would mean a tool that has both high temporal and spatial resolution. Unfortunately, this isn’t easy because MR materials need to be non-magnetic, whilst EEG contains magnetic amplifiers, batteries and electrodes. 

There are some commercially available no-magnet EEGs, but they will still cause artefact problems:

• In the fMRI signal, the artifacts will be caused primarly by the electrodes, as the form of an electrode is a loop* the MRI's magnetic fields will create induced currents in the loops. Furthermore, the conductive gel under the electrode could cause a distortion in the signal, as the gel contains H20 and fMRI mesures the magnetic susceptibility of hydrogen atoms. 

• The EEG signal will be distorted by: the fMRI magnetic field, by the radio frequency (RF) pulse that is used to measure the magnetization, and also by the  ballistocardiograph artifact (which is induced by the heart rate). 

By independent component analysis (ICA), it could be possible to remove the RF pulse and the magnetic field distortion. By measuring one's heart rate, the ballistocardiograph artifact could also be reduced. 

However, trying to correct so many distortions often leads to having a signal to noise ratio that is worse than if the methods hadn't been combined. 

*note: in an fMRI you can’t cross arms or legs as this would create a loop and an electric current would be generated!

Study where EEG and fMRI were combined 

Combing the two techniques has  been used by Ragazzoni et al., 2019, to find the source of the wave P300. They found that oddball stimuli activated the anterior intraparietal sulcus, while omitted targets activated the medial frontal gyrus and cingulate cortex. Shared activation was observed in the inferior frontal gyrus and supplementary motor area.

more information on study: https://www.nature.com/articles/s41598-019-39812-z

fNIRS-EEG

++ fNIRS: Easier to integrate, with fewer artifacts than fMRI, has no magnets and can be done on a wider range of subjects (as movement is allowed). 

--: fNIRS has poor spatial resolution (~2 cm²) and is limited to superficial cortical regions. Temporal resolution remains poor as it also measures BOLD responses.

Combining fNIRS-EEG

In fNIRS, infrared signals are sent to the skull and loop back to the outside, where they are detected by electrodes. One measures the BOLD level like in fMRI, but it can be directly combined with EEG as no magnets are used. Caps with both integrated are available. 

However, due to the negative aspects of fNIRS, it isn’t often used. The whole point of adding fNIRS to EEG would be to increase the spatial resolution, so fMRI is often preferred.

Study where fNIRS and EEG where combined

Combing the two techniques has  been done by Artemenko et al., 2019, to identify the areas related to arithmetic ability. They compared subjects with high and low mathematical activity, and found a difference in the left inferior frontal gyrus. 

Another reason why fMRI was rejected for this study, is that the study was conducted in children, and it is difficult to keep children still for so long in an fMRI machine. 

more information on study: https://link.springer.com/article/10.1007/s00429-018-1618-0

TMS-EEG

++ TMS: offers spatially focal stimulation and is  effective for motor studies, especially studies related to the M1 area, as the muscle output can be measured. 

-- TMS: its effects on non-motor regions are less understood, and it doesn't provide direct measures of neural activity (i.e., you don’t see what’s changing).

Combining TMS - EEG

Combining a correlational method like EEG with a causal method like TMS has many advantages, as it allows researchers to observe causal links between brain regions and activity.

When one creates a TMS pulse, there will first be a disruption in the signal. In ideal circumstances (if one has a compatible system) the artifact can be reduced after 20ms, but this still obscures the early brain responses to the TMS pulse. Furthermore, the TMS coil can disrupt the electrodes or lead placement, causing further distortions in the EEG signal or discomfort for the participant. Ultimately, the interpretation of TMS-evoked EEG signals can be confounded by non-neuronal artifacts, such as : muscle activity, sensory stimulation, or changes in scalp conductivity.

Study where TMS and EEG where combined

A study from Hill et al., 2021, looked at TMS oscillations in depression. Compared to healthy control participants, the activity of lower frequency alpha waves were increased in participants with major depressive disorders. This was achieved by applying a TMS pulse over the Dorsolateral Prefrontal cortex, which is often associated as the region with depressive symptoms. It was possible to find a causal marker, as TMS was used. 

More information on study: https://www.sciencedirect.com/science/article/abs/pii/S0165032721002354?via%3Dihub

TMS-fMRI

Combination TMS -fMRI

This combination is very difficult to set up. The coil has limited position possibilities due to the space and magnetic field of the fMRI. One has to be careful that the magnetic field of the TMS and the one of the fMRI do not sum up. The TMS should be made with MR-safe materials and have a very long cable (as the machinery is magnetic, it has to be out of the MRI room). 

Furthermore, the participant has to be very still for at lest 1hr, as even small movements can cause the TMS to target a complete different area. 

Study where TMS and EEG where combined

A study conducted by Dowdle et al., 2018, stimulated the DLPFC (Dorsolateral prefrontal cortex), and found that this stimulates and causes activity also in deeper areas of the brain like the anterior cingulate cortex and cardiac nucleus. These two areas lie multiple centmiteres away from the actual TMS target. This demonstrates that if one stimulates with TMS, a whole neural network will be targeted. 

More information on study:  https://www.brainstimjrnl.com/article/S1935-861X(18)30077-9/abstract

Closed loop systems

In open looped systems, which are the systems seen until now, one measures two techniques at the same time, but they do not influence each other. On the other hand, in closed loop systems, one changes the features of the techniques based on their reciprocal response. For example, TMS position can be changed based on the signal the EEG is giving. 

CL, combining  TMS -EEG

One measures the EEG signals, and based on what is measured, a specific pulse with TMS is applied. This technique is very time, frequency and phase specific. Most challenges lie in measuring pre-processing EEG in real time, while using it to apply TMS, as this needs to happen in a few milliseconds. 

Study where TMS and EEG where combined

A Study from Shirinpour et al., 2020, focused on creating a new real time phase detection method. The study consisted in hitting with TMS the alpha and mu waves shown by the EEG, and measuring the MEPs response. The goal of this experiment was to have access to the different wave phases. The results were that the output of the stimulation with TMS on muscle activity was higher if stimulated at the peak phase of the wave. Therefore, applying pulses at different wave movements could have several implications in therapy settings. 

More information on study:  https://iopscience.iop.org/article/10.1088/1741-2552/ab9dba

EEG - fUS 

++ fUS: Has a very fine spatial resolution (around 100ms). It is also non-invasive, portable and affordable compared to fMRI. 

-- fUS: Limited to imaging small regions, for this reason it's still mostly used in animal studies or for therapeutic treatments of specific areas. 

Combining EEG and fUS

This combination could potentially offer real-time, precise spatial and temporal insights into neural activity. However, integration remains challenging due to limited accessibility of fUS in human studies. In future, this techinque could be explored for studies requiring high resolution in small brain regions, such as local neurovascular coupling studies or epilepsy research.

EEG - MEG (SQUID MEG and OPM MEG)

• ++ MEG (SQUID): Offers high temporal resolution (similar to EEG) but provides superior spatial resolution compared to EEG, as magnetic fields are less distorted by the skull or scalp.

• ++ OPM MEG: a newer technology, with the same high temporal and spatial resolution as SQUID-based MEG, but it is wearable, portable, and doesn’t require cryogenic cooling (as it operates at room temperature).

• -- MEG SQUID:

1) Requires bulky, cryogenically cooled equipment, making it less accessible and portable.

2) Sensitive to only the tangential component of the brain's magnetic fields.

3) Expensive and requires magnetically shielded rooms to reduce interference.

• -- MEG OPM:

1) Currently, the technology is less developed than SQUID MEG and suffers from fewer available sensors (covering smaller areas of the brain).

2) Magnetic shielding is still required to reduce interference from environmental noise.

Combining EEG and MEG:

MEG has better spatial resolution for detecting activity from sources in the cortex, as it measures magnetic fields less affected by the skull and scalp. The combination allows analysis of a wider range of neural oscillations and activity patterns. 

Unfortunately, MEG systems are expensive, and integrating them with EEG increases the cost and complexity of the experimental setup. The need for specialised facilities and expertise limits the accessibility of combined EEG and MEG studies.

Study where MEG and EEG where combined

This study (Aydin et al. 2015), explored EEG and MEG combination to enhance source reconstruction for presurgical epilepsy diagnosis. A more accurate source localization at early time points with low signal-to-noise ratio was achieved, outperforming EEG or MEG alone. 

More information on study:  https://pmc.ncbi.nlm.nih.gov/articles/PMC4356563/

TES - EEG

• ++ TES: Provides a method to modulate neural activity non-invasively. With techniques like tDCS (transcranial direct current stimulation) or tACS (transcranial alternating current stimulation), TES can enhance or inhibit brain activity in specific regions.

Negative Aspects:

• -- TES:

1) Modulation is non-specific and often targets large regions of the brain, which may inadvertently affect areas unrelated to the study's focus.

2) Effects can vary significantly across individuals due to differences in: skull thickness, brain anatomy, or electrode placement.

3) The mechanism of action is still under debate; it remains unclear how much TES current actually penetrates the brain.

Combining EEG and TES:

Combining TES with EEG gives the opportunity to study causal relationships between brain activity and behavior. However, TES introduces strong electrical artifacts in EEG recordings, especially during stimulation, and affects broader brain regions, complicating data interpretation. Additionally, distinguishing direct effects of TES from secondary or network effects observed in EEG can be challenging.

Study where TES and EEG where combined

Many articles have been published on the possibility of combining TMS-EEG. Aron T. Hill et. al. 2016, explored this combination as a tool for examining the impact of tES on cortical function. Multimodal approaches which combine tES with TMS-EEG could lead to a deeper understanding of the mechanisms which underlie tES-induced cognitive modulation.

More information on study: https://pubmed.ncbi.nlm.nih.gov/22715493/