I am using an ActiveTwo BioSemi System for gathering EEG data, and have just read this useful information (http://www.biosemi.com/faq/adjust_filter.htm) about DC recording. If I understand this correctly, this system applies a low-pass filter as default, but I wondered whether it also applied some sort of high-pass filter?
My justification for asking this question is that whilst reading around the subject of filtering (specifically pp.185 in Luck, S. J. (2005). An Introduction to the event-related potential technique) I noted this sentence: "Some amplifiers allow you to do no high-pass filtering at all (these are called DC recordings)..." I also tried to apply a high-pass filter offline as part of my standard pre-processing, and was not able to filter using a cut-off of 0.01Hz which is optimal for investigating the effects of interest in my data (but could apply a less optimal 0.5 Hz filter).
Therefore I wonder if someone can tell me if a high-pass filter is applied as a default of the BioSemi ActiveTwo system?
IF YES - please can you tell me the default parameters of this filter?
IF NO - can anyone explain what is meant by Luck in the line I quoted above? And can anyone explain why I cannot use a high-pass filter of 0.01 on my data?
Thanks in advance for your help and guidance.
please help - Offline Filtering Optimisation Question
-
Coen
- Site Admin
- Posts: 1125
- Joined: Fri Mar 26, 2004 7:00 pm
- Location: Amsterdam, Netherlands
- Contact:
Re: please help - Offline Filtering Optimisation Question
The answer is: NO
Best regards, Coen (BioSemi)
Best regards, Coen (BioSemi)
Re: please help - Offline Filtering Optimisation Question
Hello.
"I also tried to apply a high-pass filter offline as part of my standard pre-processing, and was not able to filter using a cut-off of 0.01Hz which is optimal for investigating the effects of interest in my data (but could apply a less optimal 0.5 Hz filter)."
If you are doing some offline pre-processing, aiming for a cut-off of 0.01Hz,
you should try with a time-symmetric FIR with at least a 200 seconds impulse response length.
The filter's output will be 200s longer than the original signal.
Trimming 100s both at the start and the end, you will get a high passed version of the original signal.
(mind that the FIR must have a time-symmetric impulse response).
I hope this helps.
Best regards,
RR
"I also tried to apply a high-pass filter offline as part of my standard pre-processing, and was not able to filter using a cut-off of 0.01Hz which is optimal for investigating the effects of interest in my data (but could apply a less optimal 0.5 Hz filter)."
If you are doing some offline pre-processing, aiming for a cut-off of 0.01Hz,
you should try with a time-symmetric FIR with at least a 200 seconds impulse response length.
The filter's output will be 200s longer than the original signal.
Trimming 100s both at the start and the end, you will get a high passed version of the original signal.
(mind that the FIR must have a time-symmetric impulse response).
I hope this helps.
Best regards,
RR