Kiara Ruda Neuroimaging Portfolio

Combining a device that could incorporate aspects of EEG, fMRI, and PET would be very optimal in practice, as the strengths in one imaging modality could compensate for the weaknesses in another imaging modality. At the moment, there is no single brain imaging technique that captures electric, hemodynamic, and metabolic functions of the brain at the same time. However, with recent developments in neuroimaging technology, there have been novel approaches to exploring all three aspects at the same time by integrating simultaneous EEG-fMRI’s with fPET-FDG in real world practice (Chen et al., 2025).

Each modality is unique in how it contributes to representing the brain’s function, with the combination of all three modalities allowing for a more comprehensive and multidimensional analysis of neural activity. Electroencelography (EEG) provides millisecond-level resolution of the electrical activity in the brain through recording the summed postsynaptic potentials of many cortical pyramidal neurons. The spatial precision of EEG is limited, though, due to signals being distorted by biological and environmental artifacts as they pass through the skull (Britton et al., 2016). Functional magnetic resonance imaging (fMRI) provides high resolution spatial maps of brain activity through the detection of changes in blood flow. When neurons in the brain fire, more oxygen is consumed. This hemodynamic response causes regional changes in the BOLD signal that the fMRI scan visualizes. Unfortunately, a drawback of fMRI is its limited temporal resolution, as it is much slower than EEG at capturing events, typically capturing events in seconds rather than milliseconds (Cleveland Clinic, 2023). Positron emission tomography (PET) reveals metabolic and biochemical function. This scan does this by injecting a radioactive tracer, which gets accumulated in specific tissues in the brain based on activity, allowing for imaging of brain regions and organs based on the uptake of the tracer. A limitation of PET is the long timescale that it takes for tracer distribution, which is typically minutes. A second limitation of PET is its major reliance on radiation, which may restrict repeated use of this form of imaging. (Mayo Clinic, 2025).

Creating a hybrid system with all 3 of these imaging techniques would complement each other significantly, as each technique captures information about the brain that other techniques cannot capture. EEG contributes to the temporal resolution through its millisecond-level information about the neural timing of events, which neither fMRI or PET can do. fMRI provides high resolution spatial maps of where these specific events occur, which EEG cannot precisely localize. PET contributes metabolic and biochemical information, showing how active certain brain regions are, which cannot be quantified through EEG or fMRI. Their temporal and spatial characteristics naturally complement one another: EEG captures rapid electrical changes, fMRI links those changes to specific anatomical regions, and PET explains the underlying metabolic demands. Each technique offsets the limitations of the others. EEG’s poor spatial resolution is strengthened by fMRI’s detailed localization. fMRI’s slow temporal resolution is compensated by EEG’s fast timing. PET fills the mechanistic gap by identifying metabolic activity that electrical and hemodynamic signals cannot.

EEG, fMRI, and PET can all be used in the same session by using a PET-MRI scanner. These scanners are built so PET detectors can work safely inside the MRI’s strong magnetic field, using non-magnetic parts and shielding to keep the PET and MRI systems from interfering with each other (Vitor et al., 2017). Recent case studies have also shown that EEG can be paired with PET, with researchers aligning 64-channel EEG signals to PET metabolic data to evaluate treatment responses in Alzheimer’s disease (Wang et al., 2025). Adding EEG to a PET-MRI setup is feasible when an MRI-compatible EEG cap is used. EEG can be recorded inside an MRI, but the scanner’s magnetic fields and rapid gradient switching create strong electrical noise in the EEG wires. This interference is minimized by using non-metallic electrodes, arranging cables to avoid loops that pick up magnetic noise, and applying software filters afterward to clean out MRI-related artifacts.

All three techniques can collect data at the same time, but they run on very different time scales. EEG measures activity in milliseconds, fMRI takes whole-brain images in seconds, and PET tracks tracer uptake over minutes. To synchronize them, the MRI scanner sends timing pulses to the EEG system so every EEG sample is linked to the exact moment each MRI image is taken. PET data are grouped into time windows that match longer periods of EEG and fMRI activity, which makes it possible to compare fast electrical changes with the slower blood-flow and metabolic signals. This approach allows quick EEG events to be aligned with delayed fMRI BOLD responses and with the even slower PET metabolic changes, so all three modalities can be interpreted within the same timeline.

To combine all three techniques in the same space, all data are aligned to the structural MRI scan that is collected as part of the fMRI session. This acts as the reference frame for spatial coordinates. The positions of the EEG electrodes on the scalp are digitized and mapped onto this MRI head shape so the electrical activity can be linked to specific brain regions. PET images, which are acquired simultaneously in a PET-MRI scanner, are automatically aligned with the MRI as well. Because EEG has poor spatial resolution, fMRI provides detailed spatial maps, and PET has moderate resolution, the signals are combined by matching all three datasets to the MRI anatomy. This allows electrical activity, blood-flow changes measured by fMRI, and metabolic signals from PET to be displayed together on one consistent brain map.

When all three techniques are combined together, a researcher would see signals layered onto the same brain map. The fMRI data would show detailed regions of the brain that change with blood flow, PET would show slower metabolic changes in those same regions, and EEG activity would appear as time-varied signals linked to those spots on the cortex. On the screen, it would look like rapid EEG spikes occurring in a location that the fMRI highlights as active, while the PET map shows whether that region is using more energy than normal. This fusion creates new features that none of the techniques can show on their own. For example, you can measure how much metabolic demand is associated with specific electrical events, or how strongly an EEG pattern predicts a later fMRI or PET change. You can also track the full chain of activity, such as fast electrical firing, followed by blood-flow increases, followed by slow metabolic demand. That “chain” is the added value. If a researcher sees a spike of EEG activity in the temporal lobe at the same time that the fMRI shows increased BOLD signal in that region, and PET shows higher glucose use there, they can conclude that the area is not only firing but also recruiting blood flow and consuming extra energy to support the activity. This kind of multi-layer finding cannot be identified using any individual imaging technique by itself.

A challenge comes from combining EEG with MRI at higher magnetic field strengths, since stronger fields create larger EEG artifacts, greater heating risks, and more distortion in both EEG and fMRI signals, making simultaneous recordings harder to manage (Neuner et al., 2014). Another challenge comes from fMRI being costly and not readily available, which is an issue for PET scanners as well. The environment that individuals are placed in when using an fMRI is very confined, which may trigger claustrophobia, as well as induce brain responses to the noise that is emitted when obtaining these scans (Crosson et al., 2011).

This hybrid system would be beneficial to epileptic patients when experiencing seizures, as it can show the exact timing of a seizure with EEG, the location of the seizure with fMRI, and the long-term metabolic problems in that area with PET. For patients with early stages of Alzheimer's disease development, this system would be helpful as EEG can show slowed brain activity, fMRI can show weaker brain networks, and PET can detect early metabolic changes, making it easier to catch the disease sooner.