Various neuroimaging techniques are utilized to record brain activity, each varying in their spatial, structural and temporal resolution as well as their invasiveness and specificity. Amongst the most commonly used are EEG, an electrophysiological technique measuring electrical activity and fMRI which measures neuronal activity in the brain. Both are considered real-time neuroimaging techniques with different applications and functionality.

Temporal resolution in neuroimaging refers to how frequently a measurement can be recorded to accurately differentiate between states of a specific quality of brain activity. The higher the temporal resolution, the more ‘snapshots’ can be measured per unit of time for neuronal activity or electrical activity in the case of fMRI and EEG respectively. Temporal resolution is useful for monitoring brain activity in various states of sleep and wakefulness or anomalous brain activity such as in Epilepsy (Burle et al., 2015).

Structural resolution describes the level of anatomical detail, combining tissue density features and spatial resolution which relates to the accuracy of locating a signal in a brain section or region. These measurements can be rendered into an image and are invaluable for diagnosing brain tumours and tissue artefacts such as brain shrinkage in Alzheimer’s.

fMRI images are often a composite of scans of brain activity overlaid with negative control scans i.e. when at rest, to highlight areas which are active against base activity or ‘at rest’ levels. fMRI scans are also often superimposed upon structural scans to pinpoint brain activity relative to brain anatomy. This is not possible with EEG which simply shows spikes in electrical brain activity.

fMRI works by using the BOLD effect to detect the magnetic flux influence of deoxyhaemoglobin on the magnetic resonance orientation of water molecules. Therefore, its temporal resolution, although relatively high compared to some neuroimaging techniques, is significantly lower than with EEG as it is not a direct method of measuring of brain activity, there are in fact several intermediary stages involved to initiate a signal and measure the response, which combine to give an indication of neuronal activity and location (Pernet et al., 2016).

EEG images display real-time fluctuations of electrical brain activity accurate to milliseconds therefore it has superior temporal resolution but poorer spatial and is also lacking structural resolution. The lack of structural and minimal spatial resolution in EEG is attributed to the nature of placing electrodes on the scalp which only penetrates a limited depth. This means deeper brain region activity is less accurately recorded as these rely on a cumulative composite of action potentials covering a larger region in the brain than shallower recordings (Smith, 2005). The EEG is useful for distinguishing different types of electrical brain waves, their amplitude and frequency so lacks any form of structural resolution. However, it has some measure of spatial resolution albeit poor compared to fMRI as the location of electrical activity is computed based on the relative locations of the electrodes on the scalp and tissue density mapping. A 19 electrode setup is spatially accurate to approximately 22-37cm3 and a 129 electrode setup accurate to 6-8cm3 (Ferree et al., 2001). fMRI on the other hand has spatial resolution of approximately 3-6mm3 so specific cognitive activity such as shining a light in the eye, can be pinpointed to a region in the brain with great precision.

fMRI utilizes a powerful magnetic field and radio waves to initiate a signal and receive a response. Both of these easily penetrate through skull and brain so do not suffer from the varying density and conductivity of the four different brain layers, comprising skull, brain tissue, cerebrospinal fluid and scalp which further limits spatial resolution of EEG (Ferree et al., 2001).  Current source density methods utilizing Surface Laplacian Computation, improve spatial resolution in EEG using models of brain region density. This works on the principle that variation in tissue density and type influences the amplitude of the electrical signal and temporal delay (Burle et al., 2015).

The temporal resolution of fMRI is in the region of 1hz whereas EEG is up to 20,000hz. Temporal lag of fMRI is due to reliance on levels of deoxygenated blood in a brain region to relay a signal, employing the paramagnetic nature of deoxyhaemoglobin. As neurones become active they use up oxygenated blood causing an initial rise in deoxygenated blood, which induces surrounding capillaries to dilate and flood the local region with more oxygenated blood than regions with less neuronal activity. This haemodynamic lag is a factor for fMRI lacking the temporal resolution of EEG as blood flow and oxygenation is far slower than neuronal action potentials.

Pernet, C.R., Gorgolewski, K.J., Job, J., Rodriguez, D., Whittle, I., & Joanna Wardlaw. (2016). A structural and functional magnetic resonance imaging dataset of brain tumour patients.

Smith, S.J.M. (2005). EEG in the Diagnosis, Classification, and Management of Patients with Epilepsy. Journal of Neurology Neurosurgery and Psychiatry 2005;76(Suppl II):ii2–ii7.

Burle, B., Spieser, L., Clémence, R., Casinia, L., Hasbroucqa, T., Vidala, F. (2015) Spatial and temporal resolutions of EEG: Is it really black and white? A scalp current density view. International Journal of Psychophysiology Volume 97, Issue 3, September 2015, Pages 210-220

Ferree, T.C., Clay, M.T. & Tucker, D.M. (2001). The Spatial Resolution of Scalp EEG. Journal of Neurocomputing.

Ogawa, S. & Sung, Y., (2007), Functional magnetic resonance imaging. Scholarpedia, 2(10):310