Comparing EEG/ERP-Like and fMRI-Like Techniques for Reading Machine Thoughts

摘要

fMRI and ERP/EEG are two different sources for scanning the brain for building mind state decoders. fMRI produces accurate images but it is expensive and cumbersome. ERP/EEG is cheaper and potentially wearable but it gives more coarse-grain data. Recently the metaphor between machines and brains has been introduced in the context of mind state decoders: the "readers for machines' thoughts". This metaphor gives the possibility for comparing mind state decoder methods in a more controlled setting. In this paper, we compare the fMRI and ERP/EEG in the context of building "readers for machines' thoughts". We want assess if the cheaper ERP/EEG can be competitive with fMRI models for building decoders for mind states. Experiments show that accuracy of "readers" based on ERP/EEG-like data are considerably lower than the one of those based on fMRI-like images.