Background: fMRI provides a noninvasive means of identifying the location a
nd organization of neural networks that underlie cognitive functions. Objec
tive: To identify, using fMRI, brain regions involved in processing written
text in children. Methods: The authors studied nine normal right-handed na
tive English-speaking children, aged 10.2 years (range 7.9 to 13.3 years),
with two paradigms: reading Aesop's Fables and "Read Response Naming" tread
ing a description of an object that was then silently named). Data were acq
uired using blood oxygen level-dependent fMRI. Group data were analyzed wit
h statistical parametric mapping; individual data sets were analyzed with a
region-of-interest approach from individual study t maps. The number of ac
tivated pixels was determined in brain regions and an asymmetry index (AI =
[L - R]/[L + R]) calculated for each region. Results: The authors found st
rong activation in the left, middle temporal gyrus and left midfrontal gyru
s and variable activation in left inferior frontal gyrus for both reading t
asks in the group analysis (z > 5.5 to 9.1). All subjects had strong left-s
ided lateralization for both tasks in middle/superior temporal gyrus, infer
ior frontal gyrus, and middle frontal gyrus (AI = 0.76 to 1.0 for t = 4). R
eading Fables activated twice as many pixels in temporal cortex as the Read
Response Naming task; activation in dorsolateral prefrontal cortex was sim
ilar for both tasks. Small homologous right middle temporal region activati
on was seen with reading a fable. Conclusions: The neural networks that pro
cess reading appear to be lateralized and localized by middle to late child
hood. Reading text paradigms may prove useful for identifying frontal and t
emporal language-processing areas and for determining language dominance in
children experiencing epilepsy or undergoing tumor surgery.