How could we combine fnir and eeg signal for classification perpose? The sampling frequency of fnir and eeg are 2hz and 256 hz respectively.
Dear Muhammad Ahmad Kamran,
It is not so easy. Because the meaningful information of EEG will be removed in case of under sampling. It needs better idea than that of re-sampling.
Thank you.
Dear Md. Asadur Rahman ,
It is an awkward question! One hand, The length of processing window for NIRS must be extremely wide to illuminate its slow dynamic information. On the other hand, the long period of EEG processing window has some problems such as non-stationary. To tackle this problem you can use nonlinear processing methods which support a longer period of EEG processing window. Furthermore, you can use morphological feature extraction methods based on pick detection that is very common in this kind of EEG-NIRS parallel processing. However, if you find any other solution please inform me.
Regards,
You should indeed resample EEG, however definetly do not resample the original signal; you should resample some integral (over the resampling period) EEG metrics such as EEG band power etc. Hope this helps
Dear Kamran Antonio Maria Chiarelli , your suggestion makes me confident because I applied a procedure almost you suggested and get a beautiful result.
Thank you
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification