Mice are uniquely suited as experimental subjects for various approaches to the study of the molecular and genetic basis of behavior, and there has been a corresponding explosion in the use of mice in behavioral neuroscience. Rats and monkeys, however, remain the preferred species for high-order cognitive models largely due to the unavailability of valid, reliable and translatable endpoint measures of behavior in the mouse. Here we present further development and validation of a touchscreen-based operant method for measuring cognition that is comparable to methods used in other species and human patients. C57BL/6J mice were found to show good performance on visual discrimination and reversal learning using this method. Demonstrating the sensitivity of the paradigm to genetic factors, C57BL/6J and DBA/2J mice exhibited marked differences in discrimination and reversal learning. Systemic treatment with the selective D1-like agonist, SKF81297, produced an impairment in the early phase of reversal learning, but did not alter visual discrimination, in C57BL/6J mice. The same treatment impaired spatial working memory on the T-maze delayed alternation task, but did not alter control measures of behavior including motivation and locomotor activity. These data demonstrate the sensitivity of visual discrimination and reversal learning measured by this method to genetic factors and pharmacological challenge, and thereby provide an extension and further validation of the method for measuring cognition in mice. When combined with emerging molecular techniques uniquely suited to this species such as genetic engineering and RNA modification this paradigm could provide a powerful new tool for behavioral neuroscience.