p a i f s H R m T w t f n O c p F C c s s a W orking memory has never been so well established as a construct for understanding psychopathology. In fact, it is one of a select few constructs around which new thinking about pathophysiology in mental illness is being organized. As the construct gains in prominence, so does interest in finding out what precisely about working memory goes wrong, leading to psychiatric symptoms. In refining our ideas about what goes wrong in working memory, we may overturn the whole apple cart of current thinking. The construct of working memory, attributed to the seminal 1974 work of Baddeley and Hitch (1), originally emerged as an effort to bring order to studies of short-term memory. In the 1980s and 1990s, working memory found its greatest champion in the work of Patricia Goldman-Rakic. The successes of Goldman-Rakic and her colleagues recording from electrodes in macaque prefrontal cortex made “reverberating circuit” a household word and served to establish the importance of working memory to the cognitive deficits in schizophrenia (2). Today, working memory has become one of the organizational constructs around which the new U.S. National Institute of Mental Health (NIMH) Research Domain Criteria (RDoC) approach to studying mental illnesses is developing. At the same time, two articles by independent groups published in this issue of Biological Psychiatry serve to raise further questions about what characterizes the working memory deficits found in patients with schizophrenia. In July of this year, two dozen experts in working memory, drawn from basic and clinical research, spent 2 days in Washington, DC, wording and rewording series of opinions about the neuroscience of working memory. The purpose of this exercise was to provide a test case for the NIMH broader RDoC initiative (for more information, see http:// www.nimh.nih.gov/research-funding/nimh-research-domain-criteriardoc.shtml). The RDoC initiative was initiated by NIMH director Tom Insel to ask a simple question: can a reconceptualization of mental illnesses in terms of cognitive and affective mechanisms and the brain systems that underlie those mechanisms lead to better research and better treatments? Because so much is known about working memory, it was reasoned, the process of expert consultation must succeed here if it were to succeed for any construct. In the context of this process, working memory was defined as the active maintenance and flexible updating of task relevant information (to include items, goals, strategies, etc.) that has limited capacity in a form that resists interference. The group also accepted as inevitable the existence of edge cases, such as tasks that involve both working memory and cognitive control (including control of attentional resources), and other tasks that evoke both working and long-term memory (Figure 1). Upon arriving at a consensus on what was meant by working memory, the directive shifted. The expert panel then turned to the task of building over the course of several hours what might keep a team of graduate students busy for a year: a translational account of working memory, from genes to functional