EEG Meta Data

This page specifies the meta-data for EEG recordings.

If a stream has the content-type EEG, we recommend that meta-data about the stream adheres to the structure and naming laid out in the following. While all meta-data is optional, we recommend that any stream should describe at least the channel labels.

<channels>            <!-- per-channel meta-data; might be repeated --> 
  <channel>           <!-- description of one channel, repeated (one for each channel in the time series) --> 
    <unit>            <!-- measurement unit (strongly preferred unit: microvolts) --> 
    <type>            <!-- channel content-type (EEG, EMG, EOG, ...) --> 
    <label>           <!-- channel label, according to labeling scheme; 
                           the preferred labeling scheme for EEG is 10-20 (or the finer-grained 10-5) --> 
    <location>        <!-- measured location (note: this may be arbitrary but should then include 
                           well-known fiducials (landmarks) for co-registration) --> 
      <X>             <!-- coordinate axis pointing from the center of the head to the right, in millimeters --> 
      <Y>             <!-- coordinate axis pointing from the center of the head to the front, in millimeters --> 
      <Z>             <!-- coordinate axis pointing from the center of the head to the top, in millimeters --> 
    </location> 
 
    <hardware>        <!-- information about hardware properties --> 
      <model>         <!-- model of the sensor --> 
      <manufacturer>  <!-- manufacturer of the sensor --> 
      <coupling>      <!-- type of coupling used (can be Gel, Saline, Dry, Capacitive) --> 
      <material>      <!-- conductive material of the sensor (e.g. Ag-AgCl, Rubber, Foam, Plastic) --> 
      <surface>       <!-- type of the contact surface (e.g., Plate, Pins, Bristle, Pad) --> 
    </hardware> 
 
    <impedance>       <!-- measured impedance value during setup, in kOhm --> 
 
    <filtering>       <!-- information about the hardware/software filters already applied to the data (e.g. notch) --> 
      <highpass>      <!-- highpass filter, if any --> 
        <type>        <!-- type of the filter (FIR, IIR, Analog) --> 
        <design>      <!-- design of the filter (e.g., Butterworth, Elliptic) --> 
        <lower>       <!-- lower edge of the transition frequency band (in Hz) --> 
        <upper>       <!-- upper edge of the transition frequency band (in Hz) --> 
        <order>       <!-- filter order, if any --> 
      </highpass> 
 
      <lowpass>       <!-- highpass filter, if any --> 
        <type>        <!-- type of the filter (FIR, IIR, Analog) --> 
        <design>      <!-- design of the filter (e.g., Butterworth, Elliptic) --> 
        <lower>       <!-- lower edge of the transition frequency band (in Hz) --> 
        <upper>       <!-- upper edge of the transition frequency band (in Hz) --> 
        <order>       <!-- filter order, if applicable --> 
      </lowpass> 
 
      <notch>         <!-- notch filter, if any --> 
        <type>        <!-- type of the filter (FIR, IIR, Analog) --> 
        <design>      <!-- design of the filter (e.g., Butterworth, Elliptic) --> 
        <center>      <!-- center frequency of the notch filter (in Hz) --> 
        <bandwidth>   <!-- width of the notch frequency band (in Hz), if known --> 
        <order>       <!-- filter order, if applicable --> 
      </lowpass> 
    </filtering> 
  </channel> 
</channels> 
 
<reference>           <!-- signal referencing scheme --> 
  <label>             <!-- name of the dedicated reference channel(s), if part of the 
                           measured channels (repeated if multiple) --> 
  <subtracted>        <!-- Yes if a reference signal has already been subtracted from the data, otherwise No --> 
  <common_average>    <!-- Yes if the subtracted reference signal was a common average, otherwise No --> 
</reference> 
 
<fiducials>           <!-- locations of fiducials (in the same space as the signal-carrying channels) --> 
  <fiducial>          <!-- information about a single fiducial (repeated for each one) --> 
    <label>           <!-- label of the location (e.g., Nasion, Inion, LPF, RPF); can also cover Ground and Reference --> 
    <location>        <!-- measured location (same coordinate system as channel locations) --> 
      <X>             <!-- coordinate axis pointing from the center of the head to the right, in millimeters --> 
      <Y>             <!-- coordinate axis pointing from the center of the head to the front, in millimeters --> 
      <Z>             <!-- coordinate axis pointing from the center of the head to the top, in millimeters --> 
    </location> 
  </fiducial> 
</fiducials> 
 
<cap>                 <!-- EEG cap description --> 
  <name>              <!-- name of the cap (e.g. EasyCap, ActiCap, CustomBiosemiCapA) --> 
  <size>              <!-- cap size, usually as head circumference in cm (typical values are 54, 56, or 58) --> 
  <manufacturer>      <!-- manufacturer of the cap (e.g. BrainProducts) --> 
  <labelscheme>       <!-- the labeling scheme for the cap (e.g. 10-20, BioSemi-128, OurCustomScheme) --> 
</cap> 
 
<amplifier>           <!-- information about the used amplification (e.g. Gain, OpAmps/InAmps...) --> 
  <settings>          <!-- settings of the amplifier --> 
  </settings> 
</amplifier> 
 
<location_measurement>  <!-- information about the sensor localization system/method --> 
  <model>               <!-- model name of the system, e.g. CMS10 --> 
  <manufacturer>        <!-- manufacturer of the measurement system, e.g. Polhemus, Zebris --> 
  <locationfile>        <!-- file system path of the (backup) location file, if any --> 
  <settings>            <!-- settings of the location measurement system --> 
  </settings> 
</location_measurement> 
 
<acquisition> 
  <manufacturer>      <!-- manufacturer of the amplifier (e.g. BioSemi) --> 
  <model>             <!-- model name of the amplifier (e.g. BrainAmp) --> 
  <precision>         <!-- the theoretical number of bits precision delivered by the amplifier 
                           (typical values are 8, 16, 24, 32) --> 
  <compensated_lag>   <!-- amount of hardware/system lag that has been implicitly 
                           compensated for in the stream's time stamps (in seconds) --> 
</acquisition>