Difference between revisions of "Filtering scope"
(7 intermediate revisions by 2 users not shown) | |||
Line 4: | Line 4: | ||
|version = 6.1 or higher | |version = 6.1 or higher | ||
}} | }} | ||
− | + | ==Overview== | |
− | In EEG/MEG data analysis frequency | + | In EEG/MEG data analysis frequency filtering is one of the most basic and at the same time important operations. In BESA Research filtering is implemented as a linear operation so it is interchangeable in the processing pipeline - it can be changed at every processing step in almost all situations. There are however some situations where filtering cannot be applied in a strictly linear manner - e.g. averaging in the presence of slow drifts. In general, due to the linear nature of filtering we recommend not to apply filtering in the pre-averaging processing stage (apart from high pass to reduce drifts) as it can still be applied afterwards. You will be prompted by the application if you encounter this type of situation in data analysis: |
[[File:Filter_notification.jpg]] | [[File:Filter_notification.jpg]] | ||
− | Also at some stages of the data processing, | + | Also at some stages of the data processing, filtering is of major importance, therefore we made it more accessible and you can easily redefine it. Please have a look at the table below: |
{| class="wikitable" style="text-align:center;" | {| class="wikitable" style="text-align:center;" | ||
! style="font-weight: bold;" | Filter type | ! style="font-weight: bold;" | Filter type | ||
Line 16: | Line 16: | ||
! style="font-weight: bold;" | Averaged buffer | ! style="font-weight: bold;" | Averaged buffer | ||
! style="font-weight: bold;" | FFT | ! style="font-weight: bold;" | FFT | ||
− | ! style="font-weight: bold;" | lienear | + | ! style="font-weight: bold;" | lienear correlation / regression |
! style="font-weight: bold;" | Combine Conditions | ! style="font-weight: bold;" | Combine Conditions | ||
! style="font-weight: bold;" | ERP | ! style="font-weight: bold;" | ERP | ||
! style="font-weight: bold;" | Coherence | ! style="font-weight: bold;" | Coherence | ||
− | ! style="font-weight: bold;" | Time Frequency plot (single click) | + | ! style="font-weight: bold;" | Time Frequency plot<br>(single click) |
! style="font-weight: bold;" | TopViewer | ! style="font-weight: bold;" | TopViewer | ||
! style="font-weight: bold;" | Source Analysis | ! style="font-weight: bold;" | Source Analysis | ||
Line 26: | Line 26: | ||
| style="font-weight: bold;" | Low/High pass | | style="font-weight: bold;" | Low/High pass | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
− | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be | + | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be redefined to default values(2-35Hz) |
| style="background: #BA9588" |as used during buffer creation | | style="background: #BA9588" |as used during buffer creation | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
Line 39: | Line 39: | ||
| style="font-weight: bold;" | Notch | | style="font-weight: bold;" | Notch | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
− | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be | + | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be redefined to default values(2-35Hz) |
| style="background: #BA9588" |as used during buffer creation | | style="background: #BA9588" |as used during buffer creation | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
Line 52: | Line 52: | ||
| style="font-weight: bold;" | Bandpass | | style="font-weight: bold;" | Bandpass | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
− | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be | + | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be redefined to default values(2-35Hz) |
| style="background: #BA9588" |as used during buffer creation | | style="background: #BA9588" |as used during buffer creation | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
Line 65: | Line 65: | ||
| style="font-weight: bold;" | Polygraphic | | style="font-weight: bold;" | Polygraphic | ||
| style="background: #F2D16B" |Main* | | style="background: #F2D16B" |Main* | ||
− | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main* / can be | + | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main* / can be redefined to default values(2-35Hz) |
| style="background: #BA9588" |as used during buffer creation | | style="background: #BA9588" |as used during buffer creation | ||
| style="background: #F2D16B" |Main* | | style="background: #F2D16B" |Main* | ||
Line 78: | Line 78: | ||
| style="font-weight: bold;" | artifact correction | | style="font-weight: bold;" | artifact correction | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
− | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be | + | | style="background: repeating-linear-gradient(45deg, #F2D16B,#F2D16B 10px, #BA9588 10px, #BA9588 20px);"| Main / can be redefined to default values(2-35Hz) |
| style="background: #BA9588" |as used during buffer creation | | style="background: #BA9588" |as used during buffer creation | ||
| style="background: #F2D16B" |Main | | style="background: #F2D16B" |Main | ||
Line 96: | Line 96: | ||
==Specific ERP settings== | ==Specific ERP settings== | ||
− | The ERP filter settings can be set during paradigm definition. Note that there are two different settings - for artifact scan and for averaging | + | The ERP filter settings can be set during paradigm definition. Note that there are two different settings - for artifact scan and for averaging. |
[[File:filter_ERP.jpg]] | [[File:filter_ERP.jpg]] | ||
==Specific Source Analysis settings== | ==Specific Source Analysis settings== | ||
− | For source analysis (SA), if you press | + | For source analysis (SA), if you press the '''ESI/MSI''' button on the toolbar ribbon ''Main'' filter settings will be used. Filters for SA can be set during sending a block of data to Source Analysis (right-click on the highlighted block and select '''Source Analysis'''). |
[[File:filter_SA.jpg]] | [[File:filter_SA.jpg]] | ||
+ | In case of enabled artifact correction, you will be prompted just before averaging if you want to average artifact corrected data or not. The detailed description of difference in both approaches can be found [[Using BESA to correct blink and EKG artifacts in MEG data#Averaging the raw data|here]]. | ||
==Notch filter== | ==Notch filter== | ||
− | Note that because of its generic nature - power line artifact removal - notch filter is treated differently. The notch filter value populates from ''Main'' setting to all subsequent applications. | + | Note that because of its generic nature - power line artifact removal - the notch filter is treated differently. The notch filter value populates from ''Main'' setting to all subsequent applications. |
==Bandpass filter== | ==Bandpass filter== | ||
− | + | The bandpass filter is complementary to the high/low pass filter combination. For convenience, it is treated differently (it also has different filter characteristics) and also populates to all subsequent applications from ''Main'' setting. | |
==Polygraphic filter== | ==Polygraphic filter== | ||
Line 116: | Line 117: | ||
[[File:filter_poly.jpg]] | [[File:filter_poly.jpg]] | ||
+ | |||
+ | ==See also== | ||
+ | *[[Review#Filtering|Filtering]] | ||
+ | *[[Using BESA to correct blink and EKG artifacts in MEG data]] | ||
+ | *[[Pipeline for simultaneous EEG-fMRI recording]] | ||
[[Category:Preprocessing]] | [[Category:Preprocessing]] |
Latest revision as of 14:23, 4 May 2021
Module information | |
Modules | BESA Research |
Version | 6.1 or higher |
Contents
Overview
In EEG/MEG data analysis frequency filtering is one of the most basic and at the same time important operations. In BESA Research filtering is implemented as a linear operation so it is interchangeable in the processing pipeline - it can be changed at every processing step in almost all situations. There are however some situations where filtering cannot be applied in a strictly linear manner - e.g. averaging in the presence of slow drifts. In general, due to the linear nature of filtering we recommend not to apply filtering in the pre-averaging processing stage (apart from high pass to reduce drifts) as it can still be applied afterwards. You will be prompted by the application if you encounter this type of situation in data analysis:
Also at some stages of the data processing, filtering is of major importance, therefore we made it more accessible and you can easily redefine it. Please have a look at the table below:
Filter type | Review window | Search Average View | Averaged buffer | FFT | lienear correlation / regression | Combine Conditions | ERP | Coherence | Time Frequency plot (single click) |
TopViewer | Source Analysis |
---|---|---|---|---|---|---|---|---|---|---|---|
Low/High pass | Main | Main / can be redefined to default values(2-35Hz) | as used during buffer creation | Main | Main | None | ERP | ERP | Main | Main | SA |
Notch | Main | Main / can be redefined to default values(2-35Hz) | as used during buffer creation | Main | Main | None | Main | Main | Main | Main | Main |
Bandpass | Main | Main / can be redefined to default values(2-35Hz) | as used during buffer creation | Main | Main | None | ERP | ERP | Main | Main | Main |
Polygraphic | Main* | Main* / can be redefined to default values(2-35Hz) | as used during buffer creation | Main* | Main* | None | Main* | Main* | Main* | Main* | n.a. |
artifact correction | Main | Main / can be redefined to default values(2-35Hz) | as used during buffer creation | Main | Main | Main | ERP | ERP | Main | Main | Main/SA |
Main settings
The main setting for artifact correction can be set using EdF or menu entry Filters/Edit Filter Settings...
Specific ERP settings
The ERP filter settings can be set during paradigm definition. Note that there are two different settings - for artifact scan and for averaging.
Specific Source Analysis settings
For source analysis (SA), if you press the ESI/MSI button on the toolbar ribbon Main filter settings will be used. Filters for SA can be set during sending a block of data to Source Analysis (right-click on the highlighted block and select Source Analysis).
In case of enabled artifact correction, you will be prompted just before averaging if you want to average artifact corrected data or not. The detailed description of difference in both approaches can be found here.
Notch filter
Note that because of its generic nature - power line artifact removal - the notch filter is treated differently. The notch filter value populates from Main setting to all subsequent applications.
Bandpass filter
The bandpass filter is complementary to the high/low pass filter combination. For convenience, it is treated differently (it also has different filter characteristics) and also populates to all subsequent applications from Main setting.
Polygraphic filter
Polygraphic channels are filtered using Main settings. Note that there are additional settings that can be applied to the polygraphic filter accessible via Filters/Polygraphic/Additional Chans...