Semantic Verbal Fluency tests have been used in the detection of certain clinical conditions, like Dementia. In particular, given a sequence of semantically related words, a large num- ber of switches from one semantic class to an- other has been linked to clinical conditions. In this work, we investigate three similarity measures for automatically identify switches in semantic chains: semantic similarity from a manually constructed resource, and word as- sociation strength and semantic relatedness, both calculated from corpora. This informa- tion is used for building classifiers to distin- guish healthy controls from clinical cases with early stages of Alzheimer’s Disease and Mild Cognitive Deficits. The overall results indi- cate that for clinical conditions the classifiers that use these similarity measures outperform those that use a gold standard taxonomy