This is an impressive, well-written paper with interesting results. But a factor structure can be determined by the choice of items. In this case, the authors chose items that were similar to the obtained factor structure. I do concede that the obtained factor structure arose from BOLD activity, which gives this study its impressive validity.
Nevertheless, it would strengthen the authors' conclusions if this work were repeated with a set of words randomly sampled from the population of concrete nouns. That would allow confirmatory factor analysis to be applied.
