If you want to estimate/classify using implicit measures, whether a user is being deceptive while filling out an online questionnaire ... then methods such as EEG are not suitable.
Recording EEG data takes a lot of preparation and at least one experimenter to prepare the "user" (in your case the person filling out the questionnaire). And if you are talking about online questionnaires, I do not see a point in having an experimenter visit the user.
On the other hand, using (often toy-like) dry-electrodes that could be prepared by a single person, will not yield a sufficient data quality.
I think the easiest might be to consensually access the webcam of a user and evaluate the video stream for e.g., eye blinks, or other tentative variables.
Generally, this sounds like a very difficult task that will probably result in a high number of false positives and false negatives.