EEG Quality (EQ) has been built to give you an easy way to assess whether you are recording high quality brain signals or not. Basically, it is a scoring system for the quality of the signal in each sensor. Each sensor is given a black, red, orange, light green or dark green colour to indicate the quality of the EEG signal for that specific sensor.
Overall EEG Quality as displayed in your EMOTIV application is presented as a score of 0-100. This number makes it easy to understand the quality of your signal at a glance. How do we calculate it? See below.
Contact Quality is a simple measurement of impedence of the channel. While this measurement is used widely in neuroscience and medical settings, it is really only a useful reading to a trained EEG professional and does not give the full picture on the quality of the signal. Signals that have low impedance are not necessarily brain signals - and you might have found that sitting your headset on the table can produce a false positive - leading the user to falsely believe they are recording high quality brain data.
EEG Quality is a machine-learning trained algorithm that automatically determines the quality of the signal based on multiple metrics. Each of these metrics are important in assessing whether the recording data accurately captures the underlying brain signal. These metrics include
Contact Quality (CQ) (as above) - an impedance measurement that indicates the quality of the electrical signal passing through the sensors and the reference. It can be one of 4 values: 0 - very bad (black), 1 - bad (red), 2 - ok (orange) and 4 - good (green).
ML Signal Quality (SQ): A machine learning algorithm trained on high quality EEG recordings that were assessed and collected by the EMOTIV Research team. The Signal Quality algorithm assess data over 2 seconds and is scored 0 - very bad (black), 1 - bad (red), 3 (light green) and 4 -good (dark green)
Signal Magnitude Quality (SMQ): A measure of the signal amplitude. Sometimes, the above metrics i.e. CQ, SQ are good but the signal amplitude is very small and therefore small power fluctuations in the FFT would be undetectable. Often, this is due to poor sensor hydration or poor scalp contact. SMQ gives scores for the average signal magnitude within a 2 second epoch. The scores can be 0 - very low (black), 1 - low (red), 4 - ok (green).
Each of these metrics are important in determining the quality of the signal, therefore the sensor colour will not be dark green unless all of these metrics have a “good” score of 4. The “EQ Map” UI in your EMOTIV software displays the minimum score of any of these metrics. The resulting EQ Score for each channel can be 0,1,2,3,4 corresponding to the colours black, red, orange, light green and dark screen, respectively.
Overall EQ is calculated and displayed as a number out of 100 so you can see your EEG Quality for all channels at a glance. Overall EQ is calculated as the sum of the worst 3 channels divided by the sum of the maximum score of these 3 channels, as a percentage. The resulting mathematical formula is (min1 + min2 + min3 / 12) * 100.