Studying Cognitive Activity at several LEvels with Simultaneous depth and surface recordings
lntracerebral EEG (stereotaxic EEG, SEEG) is an invasive measure of brain activity performed during the presurgical evaluation of epilepsy, involving up to 250 distinct channels that record directly from brain structures, at the mesoscopic scale (a few millimetres). SEEG is performed on purely diagnostic motivations; collaterally, it provides a unique opportunity for investigating cognitive brain networks in humans across multiple frequencies with an exquisite spatial specificity, and matchless signal-to-noise ratio. Recently, the microscopic scale has been obtained thanks to microcontacts added to the SEEG probes. The analysis of such SEEG recordings at the mesa- and micro- scales has fueled advances in cognitive neuroscience, notably in the fields of memory and language. Still, SEEG provides only a partial view of brain activity due to its patient specific limited spatial sampling. Non-invasive methods such as EEG and MEG remain the only way to obtain a large-scale view of brain activity at its natural temporal scale (i.e., at the millisecond level) with a macroscopic spatial scale.
Recent research has started to bridge the gap between the macroscopic and mesoscopic scales in humans, by recording simultaneously invasive (SEEG) and non-invasive (EEG, MEG) neurophysiological signals. Such recordings could be used jointly for characterizing brain networks, surpassing the simple addition of modalities. Crucially, simultaneous recordings provide a view of the exact same brain activity at the different scales, and allows applying powerful single trial analysis which would not be available on separate recordings.
Our goal is to better define the spatio-temporal signature of the brain networks involved in simple cognitive tasks using simultaneous surface and depth recordings. Depth recordings will be instrumental in guiding the exploration of functional task networks, either by allowing analyses at the level of single trials, or by providing high resolution seeds for connectivity measures. In one centre, we will explore the possibility of adding micro recordings, thus reaching for the first time the three scales: micro-, mesa- and macro-. Part of the project will be dedicated to methodological (signal processing) advances. The other part will focus on the application of this unique technique to unravelling multi-scale cognitive networks.
The partners involved have extensive experience in non-invasive electrophysiology (EEG, MEG), on (intracerebral) SEEG recordings, as well as on their simultaneous recording. The consortium includes experts on signal processing for brain neurophysiological data, notably in the service of cognitive questions. The resulting datasets will be made available to the scientific community. Such reference datasets should be particularly useful for i) optimizing signal processing methods on surface data by providing a "ground truth" (SEEG) and ii) developing computational models of brain activity that incorporate knowledge from activity both at local and global scales.
In summary, simultaneous multi-scale recordings expand the view of each modality taken separately, and allow unprecedented developments in signal analysis. This project will thus i) help improving signal processing methods ii) bring a better understanding of cognitive networks, and iii) will provide unique reference datasets for computational neuroscientists.
The SCALES project aims at understanding the links between different spatial scales, micro, meso and macro-scales provided by SEEG and MEG/EEG. Furthermore, the project aims at deploying and testing the hypothesis that simultaneous multi-level recordings and analysis are key for characterizing brain activity and processing with unprecedented granularity.
In the SCALES project, we will use different methods for characterizing brain networks, contrasting various hypothesis based on combinations of putative coupling mechanisms (amplitude-amplitude, phase-phase, phase-amplitude) at different frequencies, and handling the massive multiple comparison issue with dedicated methods. This should result in a better understanding of the most relevant markers of network activity across the many different possible mechanisms.
The output of the SCALES project will include multi-level datasets and software tools made available to the scientific community.
Collaboration with HBP
The SCALES project will bring unique data to the HBP Flagship, namely simultaneous recordings of surface (EEG, MEG) and intracerebral data (stereotaxic EEG, SEEG), both in resting states and during cognitive tasks. This data will enable improved signal processing methods, will help with building multi-scale computational models, and will provide new insights on brain networks involved in cognition.