We read with interest the article by Silverstein and colleagues (Silverstein, Snodgrass, Shevrin, & Kushwaha, 2015) who questioned the putative specificity of the P3b event-related potentials (ERP) component as a neural signature of conscious access to a visual representation. Prior to this new study, numerous empirical reports revealed that a brain response peaking ∼300 msec after stimulus onset and maximally distributed over parietal electrodes – the so called P3b – is closely related to subjective visibility (Sergent et al., 2005 and Vogel et al., 1998). These experimental findings provided the bases to develop neuronal and computational theories of consciousness such as the global workspace model (Dehaene and Changeux, 2011, Dehaene et al., 2006 and Dehaene and Naccache, 2001). Silverstein and colleagues used a ‘passive attentive’ version of a masked visual odd-ball paradigm while recording scalp ERPs. In each trial, subjects were presented with either the masked word ‘LEFT’ (in 80% or 20% of trials) or the masked word ‘RIGHT’ (in 20% or 80% of trials). Word frequency was balanced across subjects, who were asked to carefully attend to the masked sequence. Not only were they instructed that this sequence contained a masked word, but also that: “however implausible it might seem, our prior data suggested that the stimuli would nonetheless be unconsciously perceived and produce brain wave effects – but only if they maintained their attention”. When contrasting ERPs elicited by rare and frequent masked words, Silverstein and colleagues identified a P3b ERP component followed by a late, and sustained, slow wave (LSW). Given that participants subjectively reported the absence of conscious perception of words, and that they performed at chance-level in a stimulus detection task performed after the main experiment, Silverstein and colleagues concluded that a P3b can be observed during unconscious perception. If valid, their interpretation would then simply invalidate the P3b as a possible candidate neural signature of conscious access. This original and provocative study, however, raises both methodological and conceptual concerns which need to be addressed before one can adopt Silverstein and colleagues' interpretation.