in my opinion the best way is to have enough trials in order to have the possibility to compute statistical analyses only on the remaining artefact-free trials....

Dear Alexandre,

Apart of having enough sample of ocular-artifacted trials, in my opinion the easiest way is:

1) to compute linear derivation VEOG+ minus VEOG-, and HEOG+ minus HEOG-.

2) to compute ICA-based ocular correction (EEGLAB)

Alternatively, regression-based (Gatton, Coles, Donchin, 1983) procedure is also useful.

Good luck and best regards,

Francisco Muñoz.

I don't have enough experience (and time) to correctly identify ocular artifacts with ICA but I think I will use it in future experiment! 

Francisco, you said to compute linear derivations but I don't understand how to do this? What are VEOG+ and VEOG- ?

I used the Gatton process in BrainVision but I did not find it in EEGlab.

Moreover, I used two frontal electrodes as EOG and I wonder how to deal with it in eeglab ?Indeed, when I check topographic view, one EOG seems to be correctly placed but the other one seems to be placed at the other eye...

Thank for your replies !

You will find answers on methodology for controlling EOG in ERPs in Picton et al, Audiometry, 1981

Dear Alexandre,

From my experience, the best way to deal with it, is to avoid it.

If your setup allows your subject to have the eyes closed, you just ask them hold their fingers on their eyelids. That works most of the time, except for the very nervous or anxious types.

Best regards,

Daniel

Dear Alexandre, as Salvatore, I would just do the analysis on artefact free trials, if you have enough statistical power. Personally, I use artefact correction procedures in EEProbe, but only with children who blink too much and for whom we would lose a lot of data. For adults, artefact rejection is not a problem. As Daniel, we try to control adult blinking. Since we work with language (often written) and images, we use an eyeblink prompting protocol (where participants are encouraged to blink between trials) to reduce artefacts during critical trials.