Hi Rasa Bhattarai

ERP is normally track specific event and you should design your experiment in a way that allow you to define and track such event, I assume you know how to use EEGLAB as you mentioned ERPLAB, I am not so familiar with ERPLAB but you can create an event list which contain all the events you are tracking, there is tutorial explain how you can create the event list and this is something you have to do and design yourself.

For the EEG out put should be the same for all EEG devices, as they follow the same channels location name, each channel has specific name and location, and the difference between one EEG device and another is only the reliability and the validity of the EEG device, some medical devices has better reading and more accuracy, but still following the same channels naming system such as 10-20 electrodes position system for example.

As you explained in your question, you are using moving objects as stimuli recorded from 2 channels EEG device, I am here just giving suggestion: how about you study the changes in the band powers of Theta, Alpha, Beta, Gamma. actually the power changes of these bands are related to some cognitive states such as attention or emotion and cognitive load or engagement, there is specific formula to calculate the changes of these states according to the change of these band powers.

i hope this help

regards Atef

Hello Rasa,

the basic idea of ERP in comparison to the continous recording of EEG is to sample epochs of EEG samples and average them with respect to experimental conditions to enhance the signal-to-noise-ratio. It is essential to determine a specific point in time when an experimentally defined epoch starts and how many sample points should be averaged. This point in time is usually named a trigger.

In many experimental setups there is a unit which controls the presentation of stimuli and the EEG recording unit. The presentation unit should be able to give an output (trigger) to the recording unit for instance marking the onset of a stimulus.

Finally you get a recording with an additional trace of triggers which can be used by an analytic appliance (workstation+matlab for instance) to average samples which are defined by the triggers found in the respective trace. For you, that would mean you should have at least three channels in your recording, two EEG and one trigger trace.

There would be a lot more to tell about ERP but I hope my short answer gives you a chance to start your continuing search for a solution at a most fruitful and practically oriented point to get in.

Regards,

Frank

Thank you all for your suggestions. Atef Eldenfria, your suggestion would be very helpful. As you mentioned " How about you study the changes in the band powers of Theta, Alpha, Beta, Gamma. actually the power changes of these bands are related to some cognitive states such as attention or emotion and cognitive load or engagement, there is specific formula to calculate the changes of these states according to the change of these band powers. "

Would you please elaborate more about the formulas for the calculation of these state of band powers ?

Hi Rasa Bhattarai

yes some cognitive state are related to the changes of band power, for example Engagement, Cognitive load, Emotion (Arousal and Valence):

Engagement = Beta power/Alpha + Theta

Cognitive load = ERD ∕ ERS % = (baseline interval band power - test interval band power)/(baseline interval band power) ∗100

Concentration = (SMR + Beta) / Theta

Emotion Arousal = Beta F3/ Alpha F3

Emotion Valence = (Alpha F4 / Beta F4) - (Alpha F3 / Beta F3)

do some research on how to measure cognitive load or engagement using brain power and you will find enough materials cover this.

i hope this help

Regards

Atef

I want to ask one question regarding erp analysis. I have collected 30 trials of data for one visual stimulus experiment.

I wanted to know that "Do we need to perform the pre-processing steps (Creating eventlist, filtering, epoching etc.) on all the trials of data at once ? I mean that do we need to create a single dataset of one person by combining all the trials and then apply pre-processing methods " OR "Do we need to perform the pre-processing steps on single trial of data and then combine it ?"

What is the method used ? How to combine all the trials in a single dataset ? Would someone explain it with an example ? Any suggestions Atef Eldenfria , Frank Schmitz