Please specify your question, what do you mean with " I couldn't find any sources describing how to plot d2 results"?? The the d2 gives you more than 1 parameter to quantify the performance of the participants, e.g. items finished, omission error, confusion error, etc. Do you need an univariate plot, like histogram or Box-plots, or bivariate plot? Are you interested in the mean values, the variances, the correlation with other variables? The is not THE method to plot d2 performace.

The figure 4 in the paper you mentioned does only show the performace values of each participant and the mean EEG values. To show the correlation a scatterplot would have been morte suiteable.

Dear Rainer Duesing , first thanks for yous answer.

What I am looking for is a method for validating the single-channel EEG signal. Reading a little about d2 test, it seemed an almost good way to examine that.

But I really don't know which parameter of it is better for this.

May I ask you if you have any other/better suggestions how can I provide a scientific reason to prove that my signal is EEG (or disprove it)?

What are you trying to validate? What ist your research question? Before looking at any plots or statistics, I highly recommend to get familiar with the things you are interested in, in the first place!

What do you mean with "provide a scientific reason to prove that my signal is EEG"?? This is a non-sensical question! If you recorded EEG, you have an EEG signal. A more valid question is, how the signal to noise ratio is in your EEG recording and which parameter you are interested in. Frequencies, components, Time-Frequency analyses, coherence/coherency.... again, I recommend to get familiar with your research field. Nobody can tell you what to look for, I YOU dont know what you are interested in.

"validating the single-channel EEG signal", what do you want to validate?!?! There is not THE EEG signal. If you know nothing about your EEG device, start with the simplest analyses, in my opinion showing that alpha frequency increases when your eyes are closed as compared to eyes open. If you look live/online at the EEG signal, you can see alpha already by visual inspection (around 10Hz). Next step, show simple forms briefly on a monitor. Typical EEG components of visual activity in the ocipital area should occur. ETC.

Dear Mr. Rainer Duesing ,

What I use is TGAM module, which is the main module used in all NeuroSky devices.

It is used for reporting Attention Level and Meditation Level, but I am trying to use it for recording EEG, as the company has claimed that one of the out puts of the module is raw EEG (and NO further information!)

The whole output is digital so I have translated them to analog to see a continuous signal. Now I have to examine whether this signal is really EEG or not.

I plotted the signal's spectra (amp-freq) , and yes, the alpha peak while eyes closed was higher that eyes open in all tests.

But what I am looking for is a more reliable analyze to prove the EEG-being of the signal.

1) If you record EEG with a computer, it is always digital. What you mean with "translate them to analog", I dont understand this phrase, the signal remains digital, but it may appear continous/analog, if you have a sufficiently large samplig rate (which should be at least twice as high as your highest frequency of interest [Nyquist theorem]).

2) If you have found increased alpha with eyes closed, you can be pretty sure that your device records an EEG signal and therefore cortical activity. Case closed? I am pretty sure that this is not what you want, but your questions are nor formulated very precisely. What do you mean with "reliable analyze to prove the EEG-being of the signal "? Of what signal? What are you trying to demonstrate in the first place? What have other researchers in your field already used? As already mentioned, you can be pretty sure that you recorded cortical activity. But now what? What are you really looking for: frequencies, event related potentials...? A correlation of one or more of the afore mentioned possible variabels with some external stimuli? If you just want to test the basic functionality of your device, do a simple attention/perception experiment, there are tons of them very well documented and validated. If you can't replicate them, your device may be not so good, if you can replicate them, fine.

3) As I understood it correctly, you have only one electrode? I think in this case your possibilities are very limited. As mentioned before, it is about the signal to noise ratio that is important. You can increase this ratio by increasing the trials/duration of your experiment AND by combining information from several electrodes to one signal (cluster formation), where the latter is not possible in your case.

4) It seems that you need some more basic information about EEG/time series analysis. I highly recommend the book Mike X. Cohen - Analyzing Neural Time Series Data, which is absolutely brilliant and easy to grasp (as compared to other textbooks). And another major advantage is that he provides MATLAB codes for all analyses so that you can replicate the findings and modify the code for your own needs (I also recommend MATLAB for EEG analysis, since it seems to be the most flexible programm, for which you can install tons of specialized EEG software (e.g. EEGLAB) and/or program your analyses by yourself, as Cohen did most of the time).

I wish you luck!

Dear Mr. Rainer Duesing ,

I really appreciate your time. Sure I am going to study the references you offered. Thank you a lot!