No electrophysiological evidence for semantic processing during inattentional blindness.

• We report a novel test of whether semantic processing can occur for unseen stimuli. • Word pairs are presented that, when seen and task relevant, elicit an N400 and robust decoding. • Inattentional blindness to the words abolishes N400 and renders semantic decoding at chance. • N400 and above-chance decoding are observed when words are seen yet task irrelevant. • Decoding may suggest global broadcasting when words are seen yet task irrelevant. • Visual Awareness Negativity is found when words are seen yet task irrelevant. • P3b is found only when words are seen and task relevant. A long-standing question concerns whether sensory input can reach semantic stages of processing in the absence of attention and awareness. Here, we examine whether the N400, an event related potential associated with semantic processing, can occur under conditions of inattentional blindness. By employing a novel three-phase inattentional blindness paradigm designed to maximise the opportunity for detecting an N400, we found no evidence for it when participants were inattentionally blind to the eliciting stimuli (related and unrelated word pairs). In contrast, participants noticed the same task-irrelevant word pairs when minimal attention was allocated to them, and a small N400 became evident. When the same stimuli were fully attended and relevant to the task, a robust N400 was observed. In addition to univariate ERP measures, multivariate decoding analyses were unable to classify related from unrelated word pairs when observers were inattentionally blind to the words, with decoding reaching above-chance levels only when the words were (at least minimally) attended. By comparison, decoding reached above-chance levels when contrasting word pairs with non-word stimuli, even when participants were inattentionally blind to these stimuli. Our results also replicated several previous studies by finding a "visual awareness negativity" (VAN) that distinguished task-irrelevant stimuli that participants noticed compared with those that were not perceived, and a P3b (or "late positivity") that was evident only when the stimuli were task relevant. Together, our findings suggest that semantic processing might require at least a minimal amount of attention. [ABSTRACT FROM AUTHOR]