Null results in TMS: From absence of evidence to evidence of absence

Abstract

It is always difficult to interpret null results. But as a research method, transcranial magnetic stimulation (TMS) has so many degrees of freedom that null results are often dismissed as meaningless. We feel that this may be unnecessary, if not counterproductive. Null results seem to inherently fulfill an important role in brain mapping. In fact, without null results, neuroimaging as an enterprise would not make sense. We argue that null results are similarly important in TMS research. By itself, neuroimaging research leaves room for doubt concerning whether or not an activated region is actually necessary for intact task performance. Interference methods such as TMS can therefore complement brain research by testing the functional relevance of that region. However, if then only positive TMS results are taken seriously, the brain interference paradigm seems less informative than promised. But how can null results inform us if they only constitute absence of evidence? We suggest that three main arguments contravene interpretation of null results in TMS. These we call the localization argument, the neural efficacy argument, and the power argument. We proceed to discuss in turn how, and under which conditions, each of these arguments may be nuanced. These considerations lead us to value null results along a gradient of meaningfulness, rather than a dichotomy. This perspective may open up a new range of TMS applications, where research questions about the lack of functional relevance of a particular brain region become valid. In this context we make specific recommendations on experimentation and interpretation. We propose that it is often not only meaningful to interpret null results, but also useful to make such findings available to the community, especially now that improved methods and an expanded knowledge base make null results more interpretable than they have been in the past.