Working memory and attention are complex cognitive functions that are disrupted in several neuropsychiatric disorders. Mouse models of such human diseases are commonly subjected to maze-based tests that can neither distinguish between these cognitive functions nor isolate specific aspects of either function. Here, we have adapted a simple visual discrimination task, and by varying only the timing of events within the same task construct, we are able to measure independently the behavioral response to increasing attentional demand and increasing length of time that information must be maintained in working memory. We determined that mPFC lesions in mice impair attention but not working memory maintenance.
[Supplemental material is available for this article.]A deficit in working memory, the ability to actively maintain and manipulate information over short time intervals (Baddeley 1992), is a core cognitive symptom of schizophrenia (Park and Holzman 1992;Barch and Smith 2008) and depression (Harvey et al. 2004;Rose and Ebmeier 2006). Working memory is a complex construct that involves a number of discrete psychological processes, including the maintenance of information, flexible updating and manipulation, and sensitivity to interference, and it is strongly influenced by attentional processes (see the NIMH Research Domain Criteria working memory workshop proceedings [Research Domain Criteria Project 2010] and Awh and Jonides 2001). Because the neural substrates for each of these discrete processes are likely to be different, each component must be precisely isolated at both the behavioral and neural levels of analysis in order to relate behavioral and biological processes.Mouse molecular models involving genetic and virally mediated manipulations, including optogenetics, can provide powerful tools for identifying the cellular and molecular mechanisms that support cognitive functions such as working memory. However, the value of such cross-species translational research depends upon the existence of functional homology and the availability of accurate behavioral assays to isolate discrete cognitive processes in mice (see Ward et al. 2011 for discussion). Researchers have identified the medial prefrontal cortex (mPFC), consisting of the infralimbic, prelimbic, and anterior cingulate cortices, as the rodent brain structure homologous to the primate dorsal lateral PFC (DLPFC) due to the similarity of anatomical connections (Furster 1997;Ongur and Price 2000).Several rodent studies have shown that mPFC damage results in poor performance in working memory tasks (Brito and Brito 1990;Granon et al. 1994;Seamans et al. 1995;Ragozzino et al. 1998;Taylor et al. 2003). The majority of these studies used maze-based tasks, such as delayed alternation or delayed match or nonmatch to sample paradigms. Such maze-based tasks do not isolate the individual cognitive processes that are involved in solving working memory tasks, such as acquisition of the task rules, maintenance of information over the delay, flexible updating, or res...