[Eeglablist] Muscles Movement Intention Detection

[Yahya Shirazi]

Hi,

Building on Cedric's valuable insights, I just wanted to highlight that MRCP is usually considered a delta wave (hence probably the 5Hz upper range of the bandpass), but you would hopefully observe it without limiting your data to 5Hz.

Also, consider whether you want to analyze your data at the Channel level (similar to most of the MRCP research) or the source level. This choice helps you decide what to do with your ICA results. If you perform the channel-level analysis, ICA will help you identify and remove artifactual sources. If you want to work on the source level (given that you have enough number of electrodes), then you need only to pick the ICs that are dipolar and have  "brain"-like features (for example, ICLABEL brain classification).

You will find both methods described in detail on EEGLAB tutorials.

Best
Yahya
______________________
Seyed (Yahya) Shirazi, Ph.D., Asst. Project Scientist, Swartz Center for Computational Neuroscience, UCSD, neuromechanist.github.io <https://urldefense.com/v3/__http://neuromechanist.github.io/__;!!Mih3wA!FGJAwf70HHedowsdWxQLd2IjTBA5CiF5weP7W-3FAgQ1w4ztD3HsAXKcMdJMreEVETS6GYAaCciJztC3nriOYwEhGA$ >
> 
> On Mar 19, 2024, at 10:46 AM, Cedric Cannard <ccannard at protonmail.com> wrote:
> 
> Hi Faruk,
> 
> Plot after filtering, the plot is blank because you plotted before removing DC offset.
> 
> No need to to Cleanline since you lowpass at 5 Hz, which is too low by the way for ERP. Common cutoff frequencies are 15 or 20 hz (still no need to do cleanline since you are removing line noise already).
> 
> To remove bad data, you can also use the clean_rawdata plugin (in Tools), to use the artifact subspace reconstruction method (ASR) to remove bad segments (do not use reconstruct if using ICA).
> 
> Yes, if you want to remove the independent components from the data, you have to flag them (manually or with the ICLabel plugin that automatically classifies them into classes). In Tools > Flag components (to flag them), and then Tools > Remove components from data.
> 
> Default in EEGLAB is to flag them, and remove them at the STUDY level. But you can also remove them here individually to fine tune your preprocessing.
> 
> Hope this helps,
> 
> Cedric
> 
> 
> 
>> On Tuesday, March 19th, 2024 at 4:51 AM, Faruk Alioglu via eeglablist <eeglablist at sccn.ucsd.edu> wrote:
>> 
>> Hi everyone,
>> 
>> I have the following approach for Detection of MRCP in EEGLAB. Could you
>> tell me if there are serious mistakes with my processing?
>> 
>> 1. Import Raw-Data together with Event markers and channel locations
>> 2. Remove Baseline, because when I plot the raw data the screen is blank
>> 3. Basic FIR Filter (0.05-5Hz)
>> 4. Cleanline
>> 5. Manually Inspection of bad data
>> 6. Decompose by ICA
>> 
>> This is also something that I didn't fully understand. Is it enough when I
>> just press Decompose Data by ICA or must I use Remove components from data
>> to have an effect on my data.
>> 
>> 7. Extracting Epochs and plotting ERP images etc.
>> _______________________________________________
>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
>> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu
>