When performing data analysis, a crucial aspect to deal with is data dimensionality. Indeed, both for high spatial resolution (~16.8 million voxels, ~0.5Hz) and for high temporal resolution (~256 channels up to ~16kHz) data, computations can already be extremely intensive. Performing source reconstruction means transitioning from the sensor space to the source space, which results in multiplying the number of data points by a factor of ~60, and thus to a drastic increase in computational times. To overcome this problem and to improve data interpretability, we can parcel the brain in regions of interest (ROIs) and thus group the data following anatomical or functional constraints.
This second seminar about source reconstruction will be divided into two parts. In the first part, we will see how to choose an atlas and to apply it for cortical parcellation; in the second part, participants will engage in a practical hands-on session during which we will apply a full source reconstruction pipeline to a limited dataset using Python.
Join at: imt.lu/sagrestia