Abstract
Recent studies have highlighted the importance of an accurate individual head model for reliably using high-density electroencephalography (hdEEG) as a brain imaging technique. Correct identification of sensor positions is fundamental for accurately estimating neural activity from hdEEG recordings. We previously introduced a method of automated localization and labelling of hdEEG sensors using an infrared colour-enhanced 3D scanner. Here, we describe an extension of this method, the spatial positioning toolbox for head markers using 3D scans (SPOT3D), which integrates a graphical user interface (GUI). This enables the correction of imprecisions in EEG sensor positioning and the inclusion of additional head markers. The toolbox was validated using 3D scan data collected in four participants wearing a 256-channel hdEEG cap. We quantified the misalignment between the 3D scan and the head shape, and errors in EEG sensor locations. We assessed these parameters after using the automated approach and after manually adjusting its results by means of the GUI. The GUI overcomes the main limitations of the automated method, yielding enhanced precision and reliability of head marker positioning.
Highlights
Electroencephalography (EEG) is a non-invasive neuroimaging technique for measuring dynamic changes in electrical potentials over the scalp that are directly induced by neural activity
These few data points cannot guarantee correct alignment with the whole head shape extracted from the magnetic resonance (MR) image; the commonly used iterative closest point registration algorithm might converge to a local minimum of its cost function that does not correspond to the correct head-surface sensor configuration
Once the required input files are loaded for a new study using SPOT3D, the first step is to align the 3D scanning data with the head shape extracted from the structural MR image
Summary
Electroencephalography (EEG) is a non-invasive neuroimaging technique for measuring dynamic changes in electrical potentials over the scalp that are directly induced by neural activity. Used techniques for collecting EEG sensor coordinates are electromagnetic or ultrasound digitization and photogrammetry[7,8,9], whose spatial localization error ranges between 5 and 15 mm[1,10,11,12,13] These widely used EEG positioning procedures are very time-consuming, when dealing with high-density montages, and require extensive contribution by the operator[4,10,14]. Digitization and photogrammetry approaches provide only a few hundred points, corresponding to the EEG sensors embedded in the montage These few data points cannot guarantee correct alignment with the whole head shape extracted from the MR image; the commonly used iterative closest point registration algorithm might converge to a local minimum of its cost function that does not correspond to the correct head-surface sensor configuration. Since the accuracy of EEG sensor positioning and labelling depends on the number of www.nature.com/scientificreports/
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
