Institute for Language Sciences Labs


Running an EEG experiment: FFR

Last updated on 20 April 2015 by Chris

If you have suggestions on how to improve this document, or find mistakes, please send them to

Wikipedia has the following to say about FFR:

Frequency following response (FFR), also referred to as Frequency Following Potential (FFP), is an evoked response generated by continuous presentation of low-frequency tone stimuli. Part of the auditory brainstem response (ABR), the FFR reflects sustained neural activity integrated over a population of neural elements.

Running an FFR experiment is very much akin to running a regular EEG experiment. The main difference is in how the equipment has to be set-up and how the stimuli are presented. The recording is done with the regular ActiveTwo Biosemi module. Not the ABR module! The module has to set to speedmode 7. In this speedmode the module samples 16,384 times per second and is limited to recording signals from only 32 pin-type electrodes (instead of 64) and 8 additional flat-type electrodes.

Electrode Placement

Electrode placement is as follows:

  • Right Mastoid : External 1, flat-type.
  • Left Mastoid : External 2, flat-type.
  • Driven Right Leg (DLR) : DLR pin-type from A set
  • Common Mode Sense (CMS) :CMS pin-type from A set
  • Cz : pin-type
  • … : pin-type
  • Facial electrodes (To better check for ocular artifacts. Omit these if you have your participant close their eyes.)

You will want to add some additional electrodes around Cz to act as backup. A cross pattern is recommended (i.e. C1, C2, FCz, and CPz). The 32-electrodes do not necessarily include electrodes that are correctly labeled with the positions you will intend to place them. This is not a problem since they are interchangeable. Simply use any of the regular pin-type electrodes and make a note of which one goes where. Try to use the same electrodes on the same locations in each session as to reduce confusion.

MASTOIDS ARE THE MOST IMPORTANT ELECTRODE LOCATIONS IN FFR EXPERIMENTS ! It can hardly be stressed enough! Make absolutely sure the left and right mastoid electrodes are placed correctly! A correct position is characterized by being on the bone extrusions and has little to no muscles between the skin and the electrode.

Participant Comfort

Frequency-following responses are very small. In order to reduce the muscle noise it is best to have the participant fully relaxed. Having them lay down is the optimal solution but a more practical solution is to have them sit in chair that has a neck-support feature. One such chair can be found in the eye-tracking lab but be aware that that lab might be in use. Furthermore, you can ask the participant to close their eyes during the experiment and dim the lights when you start running the experiment.

Audio Stimuli

You will want to play the audio via the in-ear eartips since they provide the least amount of noise. For across session consistency you’ll want to calibrate the audio volume prior to the start of every session. Eighty dB SPL is generally acceptable even though this is considered on the loud side. Because it can be unpleasant for long periods I recommend trying it out on yourself first!

Inserting the eartips can be a hassle. The eartips have to be inserted deeper than you will initial expect to ensure accurate audio playback (i.e. the top of the tips need to be nearly 2 mm into the ear canal). Ensure correct insertion or risk exposing the participant to deformed audio and introduce volume variance across sessions.

To further reduce the stimulus artifact introduced by the electronics that produce the sound you’ll want to use stimuli of alternating polarity. That is, play the sound followed by its inversion followed by the normal sound, followed by its inversion, et cetera.

Some questions answered:

Both the ABR module and the regular EEG modules allows to make FFR. The downside of using the ABR module is that you are limited to only one reference electrode and that this electrode is a flat type. This entails that it cannot be placed at Cz (flat-types and hair do not mix) and can only be placed on the forehead. This is a serious limitation and hence it is preferable to you use the ‘normal’ EEG module in speedmode 7 (16k Hz) for it allows the usage of pin-type electrodes with a cap to hold them in place.

Since the signal-to-noise ratio is pretty horrible you will want a large number of sweeps. Generally having 3,000 sweeps will give you a decent average but it would not be a bad idea to play it on the safeside and make 4,000 or even 6,000 sweeps. Note that this entails having your participant sit still for quite a long time!