Equisalience analysis (EA) is a new method for analyzing the functional architecture of human perception and cognition. EA allows one to compare the relative sensitivity of two or more perceptual/cognitive tasks to two or more dimensions of stimulus intensity. For a given task T, and two dimensions X and Y of stimulus intensity that can be used to control performance in T (e.g., X might be luminance and Y equiluminant green saturation) the X-to-Y equisalience function fT maps any intensity x of X onto the intensity y = fT(x) of Y that yields the same level of performance in task T. Along with fitting equisalience functions to data, we have developed procedures to test statistically whether two tasks (possibly with different dependent variables) share the same X-to-Y equisalience function. This question is of interest because, if two tasks with the same information requirements are found to have different equisalience functions, then the tasks probably reside in different processing streams. There are two primary ways of using EA to make scientific progress: First, by discovering tasks with different equisalience functions, we can begin to analyze cognitive processing into different functional streams; second, by enlarging families of tasks that all share the same equisalience function, we can delineate the functional boundaries between those streams.
Wright, Chubb, Winkler, & Stern (In press) provides an introduction to EA using, as an example, the question whether the ventral and dorsal streams of visual processing share the same access to information about color and luminance. Our most recent research using this tool asks whether judgments of numerosity operate consistently for objects discriminable by color, size, and orientation.