Riemannian classification of single-trial surface EEG and sources during checkerboard and navigational images in humans.

Cédric Simar,Guy Cheron,Nichita Bozga,Ana-Maria Cebolla,Axelle Leroy,Gianluca Bontempi,Mathieu Petieau,Robin Petit

doi:10.1371/journal.pone.0262417

Cédric Simar, Guy Cheron + Show 6 more

Open Access

PDF Available

https://doi.org/10.1371/journal.pone.0262417

Copy DOI

Export

Save

Cite

Abstract
Highlights/Summary
Full-Text PDF
Similar Papers

Abstract

Listen

Different visual stimuli are classically used for triggering visual evoked potentials comprising well-defined components linked to the content of the displayed image. These evoked components result from the average of ongoing EEG signals in which additive and oscillatory mechanisms contribute to the component morphology. The evoked related potentials often resulted from a mixed situation (power variation and phase-locking) making basic and clinical interpretations difficult. Besides, the grand average methodology produced artificial constructs that do not reflect individual peculiarities. This motivated new approaches based on single-trial analysis as recently used in the brain-computer interface field. We hypothesize that EEG signals may include specific information about the visual features of the displayed image and that such distinctive traits can be identified by state-of-the-art classification algorithms based on Riemannian geometry. The same classification algorithms are also applied to the dipole sources estimated by sLORETA. We show that our classification pipeline can effectively discriminate between the display of different visual items (Checkerboard versus 3D navigational image) in single EEG trials throughout multiple subjects. The present methodology reaches a single-trial classification accuracy of about 84% and 93% for inter-subject and intra-subject classification respectively using surface EEG. Interestingly, we note that the classification algorithms trained on sLORETA sources estimation fail to generalize among multiple subjects (63%), which may be due to either the average head model used by sLORETA or the subsequent spatial filtering failing to extract discriminative information, but reach an intra-subject classification accuracy of 82%.

Highlights

IntroductionSince its discovery [1, 2], EEG has increasingly been used in fundamental, clinical, and industrial researches
Since its discovery [1, 2], EEG has increasingly been used in fundamental, clinical, and industrial researches. For each of these domains, specific tools were successively developed. These tools include (i) the intracerebral recording with microelectrodes [3, 4] which allowed the recognition of the neuronal origin of EEG signals and a better understanding of the physiological mechanisms that underlie EEG activity; (ii) the grand averaging method, consisting in the average of a series of trials [5] triggered by a repetitive event, which opened the evoked-related potential (ERP) field studies, more recently enriched by EEG dynamics tools [6, 7] including EEG source generators [8–10]; and (iii) the use of EEG for neurofeedback and brain-computer interfaces (BCI) [11, 12]
As BCI procedures are devoted to work with only one subject at a time, the BCI tools are well appropriate to Riemannian classification of single-trial surface EEG and sources during checkerboard and navigational images decipher individual EEG-evoked responses and to extend our understanding beyond grand average representations of the ERP

Summary

Introduction

Since its discovery [1, 2], EEG has increasingly been used in fundamental, clinical, and industrial researches. For each of these domains, specific tools were successively developed These tools include (i) the intracerebral recording with microelectrodes [3, 4] which allowed the recognition of the neuronal origin of EEG signals and a better understanding of the physiological mechanisms that underlie EEG activity; (ii) the grand averaging method, consisting in the average of a series of trials [5] triggered by a repetitive event VEP are obtained by computing the grand average of numerous trials of ongoing EEG signals (see Eq 1), resulting in well-designed and recognizable potentials that are subsequently used to better understand the successive processing stages of visual inputs These evoked responses result from at least two different mechanisms derived from the additive or the oscillation model [8, 21–24]. The relevance of considering single-trial analyses with neuroimaging data is discussed in [42]

Results

Discussion

Conclusion