Neuropsychology is nothing without control: a potential fallacy hidden in clinical studies.

The need of appropriate methodological approaches in cognitive neuropsychology has been repeatedly addressed for over 30 years. Most of the debate has focused on whether single-case or group studies are more appropriate for drawing inferences with respect to an unimpaired cognitive architecture (e.g., Caramazza and McCloskey, 1988; Grodinsky et al., 1999). This controversy has not been resolved in either direction and, currently, both single-cases and group studies are commonly adopted in neuropsychological research. We will focus our attention upon the latter approach and will maintain that the presence of a control group does not per se guarantee an appropriate interpretation of results. According to the group study approach, an experimental groupof braindamagedpatients is typically selectedaccording to specific diagnostic tests or criteria; then an age-matched non-pathological control group is also selected on the same tests/criteria. These two groups are usually required to perform a task, which is the focus of the study. Then, in discussing the results, the immediate (and rather naı̈ve) conclusion is to ascribe the differences between the two groups to the presence/absence of the specific pathology. Here is where the fallacy can be incurred: the control group often differs from the experimental group, not only because of the absence of the pathology, but also because it lacks other characteristics, associated with the pathology itself. If we select a group of patients on the basis of their performance in a test for a specific neuropsychological deficit (e.g., anosognosia, apraxia, etc.) we are in fact selecting patients with that deficit and with a high probability of associated (often cognitive) impairments against patients (or healthy participants)