When errors occur in speeded choice tasks, these errors can be reported with a high reliability and within the fraction of a second. This so-called error awareness presumably relies on an evidence accumulation process, in which evidence for an error is sampled from multiple sources. The error positivity, an event-related potential (ERP) peaking between 200 and 500 ms after errors, has been identified as a neural correlate of the accumulated evidence, as it predicts whether and how confidently participants become aware of an error. However, little is known which cognitive and physiological mechanisms provide the evidence for this process, and thus determine whether or not an error is explicitly detected. Early ideas assume that the error negativity, an early ERP correlate of error processing, reflects a brain signal that provides this evidence. However, correlative approaches did not reveal a robust relationship between the error negativity and error awareness, leaving its role for error awareness unclear. In the present project, we use a novel target-masking paradigm to uncover the contribution of the error negativity and other cognitive and physiological variables to error awareness. With this paradigm, we are able to elicit errors for which an error positivity but no error negativity can be observed. In a first part of the project, we utilize this paradigm to investigate whether an error negativity is a necessary precondition (1) for participants becoming aware of an error with high confidence, (2) for errors eliciting heart rate deceleration and pupil dilation, that is, autonomous reactions that have been discussed as being involved in error awareness, and (3) for the occurrence of a so-called early error sensation, which refers to the feeling of having detected an error already before its occurrence. In a second part, we use the target-masking paradigm to explore which metacognitive cues embedded in the task can influence error awareness. Here, we investigate the role of knowledge about error likelihood as well as stimulus contexts and action effects that are predictive of error occurrence. The results of this project will contribute to an understanding of the cognitive and neural architecture of error processing in the human brain.