Abstract
Neuroimaging experiments in general, and EEG experiments in particular, must take care to avoid confounds. A recent TPAMI paper uses data that suffers from a serious previously reported confound. We demonstrate that their new model and analysis methods do not remedy this confound, and therefore that their claims of high accuracy and neuroscience relevance are invalid.
Highlights
1 INTRODUCTION A recent paper [8] presents a novel neural-network architecture, EEGChannelNet, for determining object class from EEG signals recorded from human subjects observing ImageNet [1] images as stimuli
The data used in [8] suffers from a confound and exhibits abnormally high classification accuracy with many different classifiers
Accuracy degrades to chance with new data collected to eliminate the confound: randomized trials and trials where the training and test data have different class presentation order
Summary
A recent paper [8] presents a novel neural-network architecture, EEGChannelNet, for determining object class from EEG signals recorded from human subjects observing ImageNet [1] images as stimuli. We document problems with the classifiers and training regimen: I Their new classifier EEGChannelNet exhibits the same flawed characteristics as the LSTM used in Spampinato et al [9], addressed in [6]. All remaining claims [8] are contingent on the confounded data, which results in refutation of the entire paper
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
More From: IEEE transactions on pattern analysis and machine intelligence
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.