Since the market for EEG devices is constantly expanding, finding the right one can indeed be tough! You could ask colleagues or collaborators for recommendations, and most likely they will tell you their experience with a specific brand. Alternatively, you might want to look for EEG papers on Google Scholar (you can search e.g. for “EEG” and “arousal”) and check their “methods” sections where the experimental setup including all devices should be listed.
If you have to start from scratch, then you can get started based on the following four classification criteria:
1) How many electrodes?
One of the core decisions is the number of EEG electrodes in the cap/headset. Generally, EEG headsets pick up the brain’s electrical activity as it is propagated to the scalp surface.
EEG headsets with 10 – 20 electrodes are typically used in both academic and commercial research to get insights into cognitive measures such as workload, affective style or engagement/alertness in response to external stimuli. Here, the data analysis focuses on global effects measured from the surface.
If you are interested in conducting source-based analyses to determine where inside of the brain the registered activity originated from, more dense electrode arrays (>20 electrodes) are required (see Michels et al., 2004). These are typically used in clinical or neuroscientific studies (e.g., diagnosis and localization of epileptic seizures).
2) Should I use wet or dry electrodes?
Electrodes pick up signals from the skin. “Wet electrodes” are typically silver/silverchloride-based and require a conductive gel for improved signal transmission. They are primarily used in clinical and neuroscience research.
“Dry” electrodes do not require conductive pastes, which makes them ideal candidates for neuromarketing and research scenarios where wet electrodes are not accepted by respondents. It is important to note that data accuracy is better when using wet electrodes, as these allow for an improved connection between electrodes and scalp (see Searle & Kirkup, 2000).
Thus, several unintrusive headsets using wet electrodes have been developed to satisfy consumer and neuromarketing needs without sacrifices in data quality.
3) Which amplifier to go for?
The EEG signal is an analog signal and consists of very low electrical voltage fluctuations that need to be amplified. Comparable to playing music, signal quality is reduced if amplifiers of lesser quality are used.
Amplifiers are often the most expensive hardware element in an EEG system. Be sure to check the quality of the amplifier – its bandwidth, sampling rate, digitization and the supported number of channels – prior to the purchase. Some vendors offer EEG electrodes with pre-amplification of the electrode signal at the recording site.
Here, small amplifiers are built into each electrode head, so the signal is already amplified before it is sent to the main amplifier. In any case, you want to ensure that you have a stable, high-quality amplifier that will resolve/digitize the signal properly without data loss or drops (see Talgren et al., 2004).
4) Should I pick a wired or a wireless system?
It is important to note that the EEG signal has to cover two major distances: First, the EEG signal has to be sent from the electrode to the amplifier. Consecutively, the amplified signal has to be sent to the computer where the data is recorded.
Cables are generally very sensitive to mechanical movements and electrical interference from the surroundings, and in former years experiments could only be done in electromagnetically shielded lab chambers with respondents sitting statically in front of the screen, executing minimal behavioral responses such as button presses or mouse clicks.
Nowadays, signals can be pre-amplified (see 3) and cable length between electrodes and amplifiers can be reduced, therefore minimizing interference effects and noise. The amplified signal is then digitized and sent wirelessly to the recording computer. It might be worth checking how the wireless transmission from the amplifier to the recording computer is accomplished (Bluetooth, radio frequency etc.) as this determines the “action space” of your respondents.
Of course, if your respondents are carrying a backpack with the recording computer then their “action space” is solely driven by the computer’s battery life (see Gramann et al., 2006).
In the end, the choice for an EEG system should be driven by the specific study you want to run, the type of task the respondent will be completing as well as the type of analyses you would like to conduct.