This study was designed to test the theorized link between reinvestment, motor chunks, and conscious processing, to provide a thorough examination of reinvestment theory. We measured electroencephalographic power and connectivity alongside self-reported conscious processing and behavioral indices of chunking in a 2 (group) × 5 (block) mixed-model design. Fifty-five individuals acquired a motor sequence (blocks: A1, A2 A3, A4) via relatively explicit (errorful) or implicit (errorless) paradigms. Then they performed in a pressure condition (block: T). Results confirmed that chunking characterizes both modes of acquisition. However, explicit acquisition resulted in quicker chunking, reduced conscious processing, and increased cortical efficiency (left-temporal high-alpha power). In support of reinvestment theory, self-reported conscious processing tended to increase under pressure among explicit trainees only. In contrast to reinvestment theory, this had no adverse effect on performance. Our results endorse explicit acquisition as an effective mode of training and provide a new neurophysiological explanation why.