Abstract
Research on cognitive disorders is challenging due to the complexity of functions studied and to the numerous variables involved. First, the concept of cognition as a mediated (semiotic) and systemic activity is reviewed. According to this concept, the result of a local lesion is not an isolated symptom but a syndrome, and the best neuropsychological approach is an analysis based on appropriate tests to disclose double dissociations and thereby provide clues to brain-behavior relationships. This approach takes into account the influence of task-relevant variables (confounders) related to the patient (e.g., age, education), to the lesion (size, etiology), and to the tests and testing conditions (ecological validity, examiner's experience), which need to be controlled and analyzed in multivariate statistical analyses, as illustrated in research on medial temporal lobe epilepsy. Other controversial issues such as single and double dissociations, single-case versus group studies, and the lesion method are also examined.