User Manual Privacy Policy Disclaimer Contact us
  Advanced SearchBrowse




Conference Paper

Is the Artifact Rejection enhanced if the EOG signals are included in the ICA decomposition?

There are no MPG-Authors available
External Ressource
No external resources are shared
Fulltext (public)
There are no public fulltexts stored in PuRe
Supplementary Material (public)
There is no public supplementary material available

Klados, M., Papadelis, C., Frantzidis, C., & Bamidis, P. (2011). Is the Artifact Rejection enhanced if the EOG signals are included in the ICA decomposition? Neuroscience Letters, 500(Suppl.), e50-e51. doi:10.1016/j.neulet.2011.05.216.

Cite as: http://hdl.handle.net/11858/00-001M-0000-0025-02C9-0
During the last decade, many Artifact Rejection techniques, based on Independent Component Analysis (ICA), were proposed [1] and [2]. One of their major drawbacks is that ICA is not able to completely separate the signals derived from the eye-movements and blinks, with the signals originated from the prefrontal cortex. This leads to the bidirection contamination phenomenon [2]. A lot of discussion has been held into various scientific fora and newsletters, suggesting that if the electrooculographic (EOG) signals are included during the decompositions procedure, then the ICA separates more properly the aforementioned signals. Despite this, until now there is not any scientific evidence supporting this practice. This study comes to shed light on this assumption, by comparing the portion of cerebral activity included in the artifactual independet components(ICs) extracted by ICA, including or not the EOG signals during the decomposition procedure. In more details the artificially contaminated dataset described in [2] was decomposed using the extended INFOMAX ICA algorithm [1] two times, one with and one without using the EOG signals. In order to quantify the portion of neural signals included into the artifactual ICs, Artifact to Signal Ratio (ASR) [2] was adopted. ASR was computed for all artifactual ICs extracted with both procedures. Then one-way ANOVA was used in order to investigate if there is a statistically significance among their mean values. Results suggested that the ASR (−4.40 dB ± 4.58) was enhanced when the EOG signals were included in ICA, while the mean ASR for the artifactual ICs extracted without using the EOG signals is −5.19 dB ± 5.54. Despite the ostensibly difference of the ASR values, there is not a statistically significant difference among them (F = 0.592; p-value < 0.44). So, it seems that when the EOG signals are included during the ICA decomposition, ICA separates more properly the ocular artifacts. But until now, there is not any clear evidence proposing for sure the use of EOG signals in ICA.