This chapter discusses the need for ecologically valid neuropsychological assessments. First, there is a review of Neiser and Banaji's debate about whether the laboratory imposes an artificial situation that does not represent the everyday world. In the years that followed, neuropsychologists attempted to offer a definition of ecological validity that was specific to neuropsychology. This definition included ideas of verisimilitude and veridicality. Recently, Burgess et al. (2006) argued for a "function-led approach" to models and assessments that proceed backward from a directly observable everyday behavior to measure the ways in which a set of actions lead to a given behavior in normal and disrupted processing. The function-led approach is contrasted with traditional construct-driven assessments that were borrowed from nonclinical studies and measure abstract cognitive constructs. This chapter presents the reader with Goldberg's (2000) contention that existing neuropsychological procedures assess veridical, but not agent-centered, decision making, which limits the tests' ecological validity because most real-life decision making is agent-centered and adaptive, rather than veridical. Next, there is a discussion of potential enhancements to ecological validity via the inclusion of the interplay of "cold" cognitive processing of relatively abstract, context-free information, and "hot" cognitive processing involved when emotionally laden information.