Localizing those anatomical structures of the brain 
that support language functions is one of the most 
exciting interdisciplinary goals of language inquiry, 
and one whose recent vigorous pursuit is bound to have 
far-reaching repercussions on the way we understand 
both human language behavior, and human cognition in 
general. 

Until the advent of neuroimaging techniques, post-mortem 
data, neuropsychological case studies, studies of 
childhood language acquisition, and experimental tests 
were the only (albeit, indirect) evidence available to 
researchers in this area. Now, neuroimaging techniques 
allow for the in vivo examination of the brain, identifying 
regions of brain activity by measuring signals associated 
with blood flow increases in metabolically active tissues. 

In this accessible paper, Sophie Scott and Richard Wise 
unveil the secrets of using neuroimaging to study patterns 
of brain activity related to a language task (namely, 
reading), and warn us away from the risks of wrong design 
choices and unwarranted results. Since the neuroimaging 
technique measures relative differences in blood flow, 
rather than absolute levels, the brain activity imaged 
during the task is contrasted to a baseline condition. 
Selection of an appropriate baseline condition is of 
crucial importance; if this condition reveals a distinct 
activity pattern, these activated regions will be subtracted 
from the task condition, and may wrongly be assumed to 
be irrelevant. 

A non-specialist willing to use this fascinating technique 
should be aware of the tricks of the trade, in order to 
steer clear of these and other similar traps. The authors 
knowledgeably address a number of them.