You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: tutorials/ConceptsGuide/coordinateSystem.md
+22-26Lines changed: 22 additions & 26 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -6,24 +6,21 @@ parent: Concepts guide
6
6
grand_parent: Tutorials
7
7
nav_order: 5
8
8
---
9
-
EEGLAB electrode coordinate systems
10
-
=========
9
+
# EEGLAB electrode coordinate systems
11
10
12
-
EEGLAB supports electrode coordinate systems with the nose pointing towards the direction +X (the origin - loosely defined - is situated at the center of the head, and the top of the head points towards the direction +Z). Electrode coordinate formats where the nose points in another direction are automatically converted, so the nose points toward +X. In terms of distance units or fiducial definition or position, EEGLAB is quite flexible. Nevertheless, for BIDS compatibility, the EEGLAB coordinate system for scanned electrodes is equivalent to the CTF MEG coordinate system, with the center of the head being situated between the LPA and RPA fiducials (see below).
11
+
EEGLAB supports electrode coordinate systems with the nose pointing towards the direction +X (the origin - loosely defined - is situated at the center of the head, and the top of the head points towards the direction +Z). Electrode coordinate formats where the nose points in another direction are automatically converted, so the nose points toward +X. In terms of distance units or fiducial definition or position, EEGLAB is quite flexible. Nevertheless, for BIDS compatibility, the EEGLAB coordinate system for scanned electrodes is equivalent to the CTF MEG coordinate system, with the center of the head being situated between the left pre-auricular (LPA) and right pre-auricular (RPA) fiducials (see below).
13
12
14
-
Three fiducial or anatomical landmark points are typically used to define a system. See this Fieldtrip [FAQ page](https://www.fieldtriptoolbox.org/faq/how_are_the_different_head_and_mri_coordinate_systems_defined/#details-of-the-mni-coordinate-system) for details on how the origin and the axes are defined in different systems.
13
+
Three fiducial or anatomical landmark points (i.e., nasion, LPA, and RPA) are typically used to define a system. See this Fieldtrip [FAQ page](https://www.fieldtriptoolbox.org/faq/how_are_the_different_head_and_mri_coordinate_systems_defined/#details-of-the-mni-coordinate-system) for details on how the origin and the axes are defined in different systems.
15
14
16
-
- If digitized electrode locations are not recorded, we recommend importing the BEM template electrode file (default) to assign electrode locations based on 10-5 channel labels.
17
-
- If digitized electrode locations are imported in EEGLAB, make sure that the orientation of the coordinate system is correct, with 'LPA' on the left (+Y), 'RPA' on the right (-Y), and the nasion facing forward (+X; up). If necessary, electrodes may be rotated in the horizontal plane using the <i>Rotate Axis</i> push button of the EEGLAB channel editor (menu item <spanstyle="color: brown">Edit → Channel locations</span>).
18
-
- You may use [scanned electrode coordinates](../09_source/Channel_Locations.md) and [MRI](../09_source/Custom_head_model.md) for localizing EEG sources.
15
+
We recommend importing the BEM template electrode file (default) to assign electrode locations based on 10-5 channel labels, if digitized electrode locations are not recorded. If digitized electrode locations are imported into EEGLAB, make sure that the orientation of the coordinate system is correct, with 'LPA' on the left (+Y), 'RPA' on the right (-Y), and the nasion facing forward (+X; up). If necessary, electrodes may be rotated in the horizontal plane using the *Rotate Axis* push button of the EEGLAB channel editor (menu item <spanstyle="color: brown">Edit → Channel locations</span>).You may use scanned electrode coordinates and EEGLAB's [get_chanlocs](https://github.com/sccn/get_chanlocs/wiki) for electrode digitization and [MRI images](../09_source/Custom_head_model.md) for custom head models and localizing EEG sources.
19
16
20
17
## EEGLAB template files
21
18
22
-
EEGLAB uses two electrode montage for 2-D representation and source localization: the BESA and standard MNI BEM montages. These electrode systems are based on a sphere that best matches the geometry of the human head. This coordinate system is shifted upward compared to the coordinate system defined as nasion, left, and right fiducials (LPA and RPA). It is also tilted forward, so Cz is defined as vertical. The reason for shifting the coordinate system up is that a sphere would not match the head well -- it would not match the head well if we used a sphere centered at the origin of the nasion, LPA, and RPA coordinate frame.
19
+
EEGLAB uses the BESA electrode montage for 2-D representation and the standard MNI BEM source localization. These electrode systems are based on a sphere that best matches the geometry of the human head. The coordinate system is shifted upward compared to the coordinate system defined as nasion, left, and right fiducials (LPA and RPA). It is also tilted forward, so Cz is defined as vertical. The reason for shifting the coordinate system up is to best fit the sphere to the head; the sphere would not match the head well if we used a sphere centered at the origin of the nasion, LPA, and RPA coordinate frame.
23
20
24
21
25
22
26
-
This is the reference frame when you use spherical coordinates for your 10-20 channel montage. The 10-10 channel coordinates from the boundary element model template also use the same reference frame.
23
+
This is the reference frame when you use spherical coordinates for your 10-20 channel montage. The channel coordinates from the BEM template also use the same reference frame.
27
24
28
25
### 2-D representation biases
29
26
@@ -35,11 +32,11 @@ When plotting 2-D scalp maps, Fpz is situated at the outer limit of the head. Th
35
32
36
33
In general, if you want to perform source localization with 10-20 montage and do not have scanned electrode locations, we advise that you use the electrode of the BEM template EEGLAB location file (the default when you select *Look up locs* in the channel editing window. This file is well-validated.
37
34
38
-
It does not mean you are stuck with the associated 2-D electrode layout, though. To achieve the desired 2-D effect, you may apply any linear transformation of the 3-D electrode coordinates. These transformations may be compensated for when coregistration them with the head model. For example, if you want to place the fiducials close to the outer limit of the head plot, you can transform the BEM template coordinates by shifting the center of the sphere down by 40 millimeters (*Opt. head center* in the channel editing window). The result is shown below. The figure below shows 81 electrodes in the original BEM coordinate system, the change in the origin of the 3-D coordinate frame, and the resulting 2-D project.
35
+
It does not mean you are stuck with the associated 2-D electrode layout, though. To achieve the desired 2-D effect, you may apply any linear transformation of the 3-D electrode coordinates. These transformations may be compensated by the co-registration of the electrodes with the head model. For example, if you want to place the fiducials close to the outer limit of the head plot, you can transform the BEM template coordinates by shifting the center of the sphere down by 40 millimeters (*Opt. head center* in the channel editing window). The result is shown below. The figure below shows 81 electrodes in the original BEM coordinate system, the change in the origin of the 3-D coordinate frame, and the resulting 2-D project.
39
36
40
37
41
38
42
-
This is not an ideal layout, as electrodes near the outer head limit are more spaced than electrodes near the center. A better option is to ask EEGLAB to change the head limit (the default is 0.5). Once you make these changes, it will affect all the 2-D plots for this dataset (EEGLAB 2023.0 and later versions only). On the command line, type
39
+
This is not an ideal layout, as electrodes near the outer head limit are more spaced than electrodes near the center. A better option is to ask EEGLAB to change the head limit (the default is 0.5). Once you make these changes, it will affect all the 2-D plots for this dataset (EEGLAB 2023.0 and later versions only). To change the head limit, type the following on the command line:
43
40
44
41
```matlab
45
42
EEG.saved = 'no';
@@ -51,32 +48,32 @@ You may actually add any [topoplot.m](http://sccn.ucsd.edu/eeglab/locatefile.php
51
48
52
49
53
50
54
-
## Other EEGLAB template files
51
+
###Other EEGLAB template files
55
52
56
-
There is an extensive collection of EEGLAB template files when you look up electrode coordinates. For example, below, we import an EGI 128 channel file and show it the layout. It is fine to perform source localization with this file as it is impossible to use the template BEM location file (EGI channels are not defined in the 10-20 system). If Fieldtrip is installed, its available [electrode files](https://www.fieldtriptoolbox.org/template/electrode/) are also shown.
53
+
There is an extensive collection of EEGLAB template files when you look up electrode coordinates. For example, we import an EGI 128 channel file below and show the layout. It is fine to perform source localization with this file as using the template BEM location file is impossible for the EGI setup (EGI channels are not defined in the 10-20 system). If Fieldtrip is installed, the available [electrode files](https://www.fieldtriptoolbox.org/template/electrode/) within the Fieldtrip toolbox are also shown.
57
54
58
55
59
56
60
-
# Create and optimize your own montage
57
+
##Create and optimize your own montage
61
58
62
-
Sometimes you might want to create your own montage. For example, you might want to add some electrodes in specific locations, and usually cap manufacturer will accommodate you. You might also want to maximize head coverage, including adding electrodes as low as possible (see this [project](https://github.com/arnodelorme/optimize_montage)). We recommend maximum head coverage instead of dense coverage in the upper head region. Because of volume conduction, the bottom electrodes capture a lot of depth information and are useful for source localization.
59
+
Sometimes you might want to create your own montage. For example, you might want to add some electrodes in specific locations, and cap manufacturers will usually accommodate these changes. You might also want to maximize head coverage, including adding electrodes as low as possible (see this [project](https://github.com/arnodelorme/optimize_montage), as an example). We recommend maximum head coverage instead of dense coverage in the upper head region. Because of volume conduction, the bottom electrodes capture a lot of depth information and are useful for source localization.
63
60
64
-
# Other 2-D layout
61
+
##Other 2-D layout
65
62
66
-
If you are adamant about using a specific electrode layout for 2-D representation, you may use do so. However, you should not use the electrode locations for source localization or 3-D plotting as they are defined for 2-D plotting only (even though EEGLAB will automatically infer 3-D coordinates for them).
63
+
If you are persistent about using a specific electrode layout for 2-D representation, you may do so. However, you should not use the electrode locations for source localization or 3-D plotting as they are defined for 2-D plotting only (even though EEGLAB will automatically infer 3-D coordinates for them).
67
64
68
-
EEGLAB allows importing a variety of layouts. For example, below, after the tutorial dataset *eeglab_data.set*, we use the *eeglab_montage11_layout.loc* layout. To do so, call the channel editor using menu item <spanstyle="color:brown">*Edit > Channel locations*</span>, then click on the *Look up locs* button. Depending on the layout, you may have to adjust the plotting radius in the channel editor interface, so the entire head is visible. EEGLAB also allows importing Fieldtrip's [layouts](https://www.fieldtriptoolbox.org/template/layout/) if it is installed. The figure below shows some of types of 2-D electrode layouts available.
65
+
EEGLAB allows importing a variety of layouts. For example, after importing the tutorial dataset *eeglab_data.set*, we load the *eeglab_montage11_layout.loc* layout. To do so, call the channel editor using menu item <spanstyle="color:brown">*Edit > Channel locations*</span>, then click on the *Look up locs* button. Depending on the layout, you may have to adjust the plotting radius in the channel editor interface, so the entire head is visible. EEGLAB also allows importing Fieldtrip's [layouts](https://www.fieldtriptoolbox.org/template/layout/) if it is installed. The figure below shows some of types of 2-D electrode layouts available.
69
66
70
67
71
68
72
-
# Considerations about fiducials
69
+
##Considerations about fiducials
73
70
74
-
The left and right preauricular (LPA & RPA) points are commonly used as ear anatomical points (also known as fiducials). Unfortunately, they are poorly defined.
71
+
The left and right pre-auricular (LPA & RPA) points are commonly used as ear anatomical points (also known as fiducials). Unfortunately, they are poorly defined in different electrode montages, although errors in defining fiducial locations can systematically change source localization results [(reference)](https://doi.org/10.1109/NER.2019.8717065).
75
72
76
73
The default EEGLAB electrode coordinate system for datasets with anatomical landmarks labeled 'LPA' and 'RPA' as shown in the figure below. For backward compatibility purposes, this coordinate system is used regardless of whether the points labeled 'LPA' and 'RPA' are marked according to their formal definition (points on the posterior root of the zygomatic arch, see [here](https://www.fieldtriptoolbox.org/faq/how_are_the_lpa_and_rpa_points_defined/) for more details).
77
74
78
-
The exact ear anatomical landmark may be included with the EEG or MEG data (e.g., in the AnatomicalLandmarkCoordinateSystemDescription field of *_coordsystem.json for [BIDS specification](https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/03-electroencephalography.html#coordinate-system-json-_coordsystemjson)). It is identical to the [CTF coordinate system](https://www.fieldtriptoolbox.org/faq/how_are_the_different_head_and_mri_coordinate_systems_defined/#details-of-the-ctf-coordinate-system) for MEG.
79
-
- Units in millimeters
75
+
The exact ear anatomical landmark may be included with the EEG or MEG data (e.g., in the AnatomicalLandmarkCoordinateSystemDescription field of *_coordsystem.json for [BIDS specification](https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/03-electroencephalography.html#coordinate-system-json-_coordsystemjson)). It is identical to the [CTF coordinate system](https://www.fieldtriptoolbox.org/faq/how_are_the_different_head_and_mri_coordinate_systems_defined/#details-of-the-ctf-coordinate-system) for MEG. Please note the following consideration when including the fiducial locations within the datasets:
76
+
- Units are in millimeters
80
77
- The origin is exactly between the points labeled as 'LPA' and 'RPA'
81
78
- The X-axis points towards and goes through the nasion
82
79
- The Y-axis points approximately towards the 'LPA,' orthogonal to the X-axis
@@ -86,12 +83,11 @@ As shown below, even when properly defined, the 3 points LPA, RPA, and nasion ma
86
83
87
84
88
85
89
-
Eventually, the center of the reference frame defined by the LPA, RPA, and nasion fiducials is not critical. Fiducials must be defined accurately to align head montage with MRI scans, but their relative position matters little. For display purposes, it is important that electrodes be organized similarly to one of the standard 2-D layouts and that we may align them to a 3-D head model for source localization purposes. Both processes do not depend on using the reference frame defined by the LPA, RPA, and nasion fiducials.
86
+
Eventually, the center of the reference frame defined by the LPA, RPA, and nasion fiducials is not critical. Fiducials must be defined accurately to align head montage with MRI scans, but their relative position does not matter as much. For display purposes, it is important that electrodes be organized similarly to one of the standard 2-D layouts and that we may align them to a 3-D head model for source localization purposes. Both processes do not depend on using the reference frame defined by the LPA, RPA, and nasion fiducials.
90
87
91
88
## Use the helix-tragus junction for fiducials
92
-
PA points are palpable anatomical features but are challenging to locate in anatomical MR head images and 3-D EEG electrode scans. Furthermore, some experiment protocols call for using other anatomical points on the ears (e.g., ear canal, ear lobes, etc.) while retaining the inaccurate PA label. While consistency is critical when gathering data for a study, collaborative or data mining projects are undermined when accurate labels or descriptions are not used. With photogrammetry improving the availability of electrode localization, we strongly recommend using the helix-tragus junction (LHJ & RHJ) as the ear fiducials: these points are identifiable in 3-D head models as well as MR head images.
Pre-auricular points are palpable anatomical features but are challenging to locate in anatomical MR head images and 3D EEG electrode scans. Furthermore, some experiment protocols call for using other anatomical points on the ears (e.g., ear canal, ear lobes, etc.) while retaining the inaccurate pre-auricular labels. While consistency is critical when gathering data for a study, collaborative or data mining projects are undermined when accurate labels or descriptions are not used. With photogrammetry improving the availability of electrode localization, we strongly recommend using the helix-tragus junction (LHJ & RHJ) as the ear fiducials: these points are identifiable in 3-D head models as well as MR head images. This is the coordinate system we recommend for [<i>get_chanlocs</i>](https://github.com/sccn/get_chanlocs/wiki), an EEGLAB plug-in for photogrammetric electrode localization using 3-D head models.
95
91
96
-
This is the coordinate system we recommend for [<i>get_chanlocs</i>](https://github.com/sccn/get_chanlocs/wiki), an EEGLAB plug-in for photogrammetric electrode localization using 3-D head models. The default EEGLAB coordinate system, except the aforementioned helix-tragus junction (HJ) points, is used as the ear anatomical landmarks.
0 commit comments