I am working on classification of schizophrenia and i see the channels placed at central sulcus give good accuracy then using all channels. I want to this automatically using some algorithm.
Here is a paper related to it,
Article A review of channel selection algorithms for EEG signal processing
but it seem most of them are computationally extensive or working on feature reduction leading to channel reduction.
I am looking for some direct channel selection algorithm. Any latest paper in your mind? or any idea please
Hello:
I recommend you to use the Fast Neighborhood Component Analysis (FNCA) for feature selection, which is provide on Matlab as fscnca.
Yang, Wei, Kuanquan Wang, and Wangmeng Zuo. "Fast neighborhood component analysis." Neurocomputing 83 (2012): 31-3
Best regards
Hello,
I would be happy to add my own opinion here, but would benefit from knowing a bit more about what you mean by "classification of schizophrenia".
More generally, I think it depends on what you want the classifier to achieve and whether you want the features to have neurobiological significance (informativeness) or instead want a measure that differentiates one population from another. Additionally, I think it will also depend on what the person with schizophrenia is doing during the EEG recordings and what type(s) of stimuli are being presented.
In short, some more details here would be essential to give you a good reply.
best,
Micah
Micah M Murray subject has to press a button after every one second which generate a tone after while eeg is being recorded
Conference Paper Motor Imagery Classification with Covariance Matrices and No...
Given the cognitive impairments in schizophrenia patients, I suggest you start with frontal channels. See this publication
Article Comparing EEG Nonlinearity in Deficit and Nondeficit Schizop...
See the subject-specific EEG channel selection @ Article Subject-specific EEG channel selection using non-negative ma...
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification