Dear Brainstorm community,
I have a problem I am trying to solve. I have a dataset of EEG + ECG data recorded from 42-weeks old newborns, and I am preprocessing it to obtain the cleanest data possible. Since the cardiac is a major presence in these data, I wanted to get rid of it. To do so, I used the ECG channel to detect the heartbeats. To remove them, I have tried to calculate the SSP, but Brainstorm itself tells me that 9 channels are a very low number for the algorithm. If I try to use it anyway, the SSP components tend to remove data of an entire channel (like O2 or T8, which are in fact the one where the heartbeats are more evident) instead of the artifact associated with the heartbeat itself. Are there any other processes that can be done with few electrodes?
Thanks!
Did you run the SSP after having detected hearbeat events from those channels to did you run the SSP over the entire recording with no events marked?
Dear Sylvain, I did it the wrong way. After heartbeat detection, SSP work much better! Thank you for the indication.
Now, however, I am facing a similar problem: during an acquisition the EEG channels worked fine, but the cardiac channel was not working. Is it possible to remove the cardiac artifact (which is pretty evident in the EEG data) without the ECG? I have tried to put cardiac events manually, but the SSP remove not only cardiac activity but also something like 60% of the whole (EEG) signal.
You can indeed select an EEG channel where heartbeats are clearly visible as a template channel for the SSP calculations instead of the ECG channel.