Developmental Neuropsychology
ISSN: 8756-5641 (Print) 1532-6942 (Online) Journal homepage: https://www.tandfonline.com/loi/hdvn20
Age-Related Changes in Regional Brain Activation
During Phonological Decoding and Printed Word
Recognition
Panagiotis G. Simos, Joshua I. Breier, Jack M. Fletcher, Barbara R. Foorman,
Angeliki Mouzaki & Andrew C. Papanicolaou
To cite this article: Panagiotis G. Simos, Joshua I. Breier, Jack M. Fletcher, Barbara R. Foorman,
Angeliki Mouzaki & Andrew C. Papanicolaou (2001) Age-Related Changes in Regional Brain
Activation During Phonological Decoding and Printed Word Recognition, Developmental
Neuropsychology, 19:2, 191-210, DOI: 10.1207/S15326942DN1902_4
To link to this article: https://doi.org/10.1207/S15326942DN1902_4
Published online: 08 Jun 2010.
Submit your article to this journal
Article views: 149
View related articles
Citing articles: 5 View citing articles
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=hdvn20
DEVELOPMENTAL NEUROPSYCHOLOGY, 19(2), 191–210
Copyright © 2001, Lawrence Erlbaum Associates, Inc.
Age-Related Changes in Regional Brain
Activation During Phonological
Decoding and Printed Word Recognition
Panagiotis G. Simos and Joshua I. Breier
Vivian L. Smith Center for Neurologic Research
Department of Neurosurgery
The University of Texas–Houston Medical School
Jack M. Fletcher, Barbara R. Foorman, and Angeliki Mouzaki
Department of Pediatrics
The University of Texas–Houston Medical School
Andrew C. Papanicolaou
Vivian L. Smith Center for Neurologic Research
Department of Neurosurgery
The University of Texas–Houston Medical School
Using magnetic source imaging, age-related changes in spatiotemporal brain activa-
tion profiles associated with printed word recognition and phonological decoding
(pseudoword reading) were examined in 27 adults and 22 children without reading
problems. Adults showed a distinct spatiotemporal profile during reading of both
types of print consisting of bilateral activation of occipital cortices, followed by
strongly left-predominant activation of basal temporal regions, and, finally, left hemi-
sphere temporoparietal (including the angular gyrus) and inferior frontal activation.
Children lacked the clear temporal distinction in the engagement of basal and
temporoparietal areas and displayed significantly weaker activation of the left infe-
rior frontal gyrus. In addition, the consistent hemispheric asymmetries in the degree
of activation of basal temporal regions that were present in the adult readers were not
apparent in the children. In contrast, the strong left hemisphere preponderance in the
Requests for reprints should be sent to Panagiotis G. Simos, Department of Neurosurgery, The Uni-
versity of Texas–Houston Medical School, 6431 Fannin Street, Suite 7.154, Houston, TX 77030.
E-mail: [email protected]
192 SIMOS ET AL.
degree of activation of temporoparietal areas was present in children as well as adults,
regardless of the type of print they were asked to read. The data suggest that the degree
of specialization of cortical regions for reading, as well as the pattern of regional inter-
actions that supports this specialization, may change with age.
With the recent emergence of functional imaging methods that allow for the detec-
tion, localization, and quantification of brain activity associated with cognitive
functions, it has become possible to address systematically issues regarding the
neurophysiological basis of reading. Functional brain imaging is well suited for this
purpose because it is noninvasive and can be used repeatedly with both children and
adults. Three basic functional imaging methods have been applied for that purpose
thus far: positron emission tomography (PET), functional MRI (fMRI), and, most
recently, magnetic source imaging (MSI; otherwise known as magneto-
encephalography [MEG]). PET and fMRI capture blood flow and local metabolic
changes, averaged across seconds or even minutes, contingent on the differential
degree of activation of various brain structures during the performance of specific
tasks. In contrast, MSI, by directly measuring electrical currents in neuronal aggre-
gates during task performance, provides, in real time, a spatiotemporal map of brain
activity. Thus, MSI may provide a more direct index of cognitive, task-specific ac-
tivation than methods that rely on indexes of local blood flow and metabolism. In
addition, MSI is currently the only imaging technique supported by extensive
cross-validation data from clinical studies in which MSI results were compared di-
rectly with the results of invasive functional mapping techniques (Breier, Simos,
Papanicolaou, et al., 1999; Papanicolaou et al., 1999; Simos, Breier, et al., 1999;
Simos, Papanicolaou, et al., 1999). The excellent agreement found in these studies
between MSI and invasive mapping procedures clearly indicates that MSI accu-
rately identifies the brain regions activated during the performance of language
tasks. The language functions that have been examined thus far include phonologi-
cal analysis and comprehension of spoken utterances. It should also be emphasized
that the validity of MSI-derived maps has been established for individual partici-
pants and patients, rather than merely for average group results, as is typically the
case with other functional imaging techniques.
Having obtained sufficient evidence for the validity of the MSI technique, we
developed protocols that provided evidence for the existence of a distinct spatio-
temporal brain activation profile specific to reading and phonological decoding
(Breier , Simos, Zouridakis, & Papanicolaou, 1998; Breier, Simos, Zouridakis, &
Papanicolaou, 1999) in adults without reading problems. This profile was obtained
in the context of two reading tasks, one involving real words and another using
pseudowords. The latter, representing meaningless, yet pronounceable, combina-
tions of phonological elements (e.g., gnume, glisp) are often used to isolate phono-
logical aspects of letter-to-sound decoding (Rumsey et al., 1997). The main
features of the activation profile in the majority of participants include predomi-
PHONOLOGICAL DECODING 193
nant activation of the left basal temporal region (after the initial bilateral occipital
activation), followed by activation of the left temporoparietal region (TMP; i.e.,
posterior superior temporal, supramarginal and angular gyri) and of the inferior fron-
tal region. Temporoparietal areas appear to be intimately involved in the phonological
analysis of words (Eden & Zefiro, 1998; Simos, Breier, Wheless, et al., 2000),
whereas extrastriate and basal temporal areas are likely to be involved in vi-
sual/graphemic analysis (Breier et al., 1998, Breier, Simos, Zouridakis, et al., 1999;
Pugh et al., 1996). On the other hand, there have been suggestions for a role of inferior
frontal areas in phonological processing (Pugh et al., 1996; Shaywitz et al., 1998).
These data are consistent with findings from other imaging modalities regarding the
spatial extent of activation associated with reading (Shaywitz et al., 1998).
At present, very little is known regarding the development of neurophysio-
logical processes associated with the establishment of reading skills. Activation
profiles indicating the spatial extent of task-specific cortical engagement are ex-
pected to change with time during acquisition of any new skill (Castro-Caldas, Pe-
terson, Reis, Stone-Elander, & Ingvar, 1998; Poldarak, Desmond, Glover, &
Gabrieli, 1998), revealing underlying changes in functional brain organization to
meet the new demands. Reading, in particular, is a relatively late achievement in
the course of human development and depends on skills that do not appear to be di-
rectly related to those normally involved in understanding spoken language, such
as phonological awareness (Liberman, 1998). Moreover, according to some ac-
counts, word recognition may depend on different cognitive operations in begin-
ning readers than in experienced readers. For instance, it is likely that word
identification may rely more on sublexical, phonological decoding operations than
on direct (i.e., whole-word) lexical access (Share, 1995) during the early stages of
reading acquisition. Thus it is reasonable to expect changes in brain activation as
such skills are acquired and become increasingly more refined. Changes in brain
activation profiles may, in turn, reflect modifications in certain aspects of brain
function such as increased participation of particular areas (perhaps relative to
other areas) and also changes in the manner in which regional engagement is inte-
grated across different areas to support reading.
Previous attempts to clarify these issues have relied on event-related potentials
(ERPs). These methods possess adequate temporal resolution for capturing
changes in brain activation in real time (e.g., Ackerman, Dykman, & Oglesby,
1994; Licht, Bakker, Kok, & Bouma, 1992) but lack the spatial resolution required
to identify brain regions engaged in a particular task. In contrast, MSI relies on
scalp-recorded evoked fields (EFs) that afford greater spatial resolution than ERPs
in terms of the localization accuracy of the intracranial events that produce EFs.
MSI reliably probes both temporal and spatial aspects of brain function by provid-
ing accurate information regarding the degree of task-specific engagement of a
given area as well as the relative timing of this engagement. This technique can
now be used with confidence in the study of complex cognitive functions such as
194 SIMOS ET AL.
reading. In this context, we report on the use of whole-head MSI to describe a dis-
tinct spatiotemporal brain activation profile that is specific to reading and phono-
logical decoding in children, and we compare this profile with the one obtained
from adults using similar reading tasks.
METHOD
Participants
EFs were collected from 12 children (7 boys, age 8–15 years) and 16 adults (9 men,
age 23–38 years) during performance of a real-word recognition task. Two differ-
ent participant groups, one consisting of 10 children (7 boys, age 8–14 years) and
11 adults (6 men, age 24–41 years) were tested on a pseudoword rhyme-matching
task. All participants were skilled at decoding as indicated by their performance
above the 80th percentile on the Word Attack subtest of the Woodcock–Johnson
Psycho-Educational Battery (Woodcock & Johnson, 1989). All participants were
right-handed (with scores of +.40 or above on the Edinburgh Handedness Inven-
tory; Williams, 1991), with English as their primary language and no history of
hearing deficit, neurological or psychiatric disorder, or visual impairment.
Stimuli and Tasks
Evoked magnetic responses to printed word and pseudoword stimuli were obtained
during two tasks—one involving real words and the other pseudowords. In an at-
tempt to match the tasks given to children and adults on overall difficulty, we sys-
tematically manipulated word readability levels. In the real-word task given to the
adult readers, stimuli were 63 English nouns. Word frequency ranged from “very
frequent” to 9 occurrences per million for some words (Thorndike & Lorge, 1944).
A different list of 63 nouns was chosen for the children. The list consisted of words
with substantially higher readability levels as indicated by frequency of occurrence
greater than 20 per million in the corpus of second-grade-level reading material ac-
cording to the norms provided by Zeno, Ivens, Millard, and Duvvuri (1995).
Thirty-three words from each list were used as targets and the remaining 30 as
distractors. There were no significant differences (p > .05) between the two lists of
targets in concreteness ratings (Paivio, Yuille, & Madigan, 1968). The target stim-
uli were presented for study for a period of 1 min before each scan. The method of
presentation of the stimuli was the same in recording and study sessions. MEG re-
cordings were obtained during three blocks of trials. The target stimuli were re-
peated in each block in a different random order each time, mixed with 10 new
distractors. Words were presented for 1 sec to prevent potential contamination of
PHONOLOGICAL DECODING 195
the EF record by visual offset responses. The interstimulus interval varied ran-
domly between 3 and 4 sec across trials.
The pseudoword task involved sequential presentation of pronounceable letter
strings. Again, to match the tasks given to children and adults on overall difficulty
level, we varied the number of letter strings that participants in each group had to
decode on each trial. Adults were presented on each trial with two pseudowords,
one on top of the other, followed by a third word. They were asked to decide
whether either one of the two initial pseudowords rhymed with a third pseudoword
presented 2 sec later. Children were presented with a single pseudoword followed
by a second. Matching trials involved pseudowords that sounded alike but were or-
thographically dissimilar (e.g., gnume–noom) to discourage performing compari-
sons on the basis of orthographic information. Therefore, successful completion of
the task required phonological decoding of the unfamiliar letter strings (i.e., map-
ping the letter strings onto the appropriate phonological representations). Stimulus
duration (1 sec), interstimulus interval (2 sec onset to onset), and intertrial interval
(randomly varied across trials between 4 and 5 sec) were identical across tasks. All
stimuli were projected through a Sharp LCD projector (Model XG–E690U; Sharp
Electronics Corp., Mahwah, NJ) on a white screen located approximately 1.5 m in
front of the participant and subtended 1° to 4° and 0.5°of horizontal and vertical vi-
sual angle, respectively. In both tasks, participants were asked to respond to rhym-
ing pseudowords by raising their index finger after the second stimuusi in each
pair. EFs were recorded to the first stimulus of each pair to ensure that the brain ac-
tivity recorded corresponded to phonological decoding operations and did not re-
flect the additional cognitive operations that matching of the stimuli entailed.
Responding hand was counterbalanced across participants.
MEG Recording and Analyses
The principles underlying the capacity of MSI to identify brain areas that show
event-related increases in local neuronal activity can be described briefly as follows.
External stimuli are known to evoke neurophysiological activity as soon as they im-
pinge on sensory receptors. One basic aspect of such activity is the intracellular and
extracellular flow of ions that is associated with magnetic flux. The latter can be re-
corded from the head surface in the form of EFs. Application of a train of similar
stimuli results in the repeated evocation of such activity that is recorded and aver-
aged to improve signal quality. Several time series of average magnetic flux data
(epochs) are actually recorded, one per recording location. In this way, at any given
point in time one can reconstruct the distribution of magnetic fields on the head sur-
face. Given that scalp-recorded magnetic flux undergoes minimal distortion as it
propagates through the head to reach the surface, the location of underlying
neurophysiological events (activity sources) can be reliably determined.
196 SIMOS ET AL.
In our study, MSI data were recorded in a magnetically shielded room with a
whole-head neuromagnetometer (4D Neuroimaging, San Diego, CA) that con-
sisted of 148 magnetometer coils placed in a cryogenic container (dewar). The sig-
nal was filtered online with a bandpass filter between 0.1 and 50 Hz and digitized
with a sampling rate of 250 samples per second starting 150 msec before, and con-
tinuing for 1,000 msec after stimulus onset. To further reduce the amount of mag-
netic noise produced by remote sources, the MEG data were submitted to an
adaptive noise reduction procedure that is part of the 4D signal analysis software.
Throughout the recording session eye movements and blinks were monitored by
recording electrooculogram (EOG) from gold-disk electrodes affixed to the outer
canthus and the supraorbital ridge of the right eye (bipolar montage). Electrode re-
sistance was less than 5 KOhm. After signal processing, single-trial MEG epochs
were averaged together, separately for each task, after removing those during
which an eye movement or blink had occurred (as indicated by a peak-to-peak am-
plitude in the EOG in excess of 50 m V). For adults, average rejection rates were
10% (range 5%–19%) and 12% (range 4%–18%) in the word and pseudoword
tasks, respectively. Corresponding rates for children were 13% (range 7%–19%)
and 14% (range 6%–20%), ensuring that a minimum of 80 artifact-free EF epochs
would be available to calculate each averaged waveform.
The intracranial sources of the observed EFs at each time point throughout the
epoch were modeled as equivalent current dipoles, which were fitted independ-
ently of each other every 4 msec by using the nonlinear Levenberg–Marquardt al-
gorithm (Sarvas, 1986). Activity source solutions were considered satisfactory if
they were associated with a correlation coefficient of at least .90 between the ob-
served and the “best” predicted magnetic field distribution. The location of activity
sources was computed in reference to a Cartesian coordinate system defined by a
set of three anatomical landmarks (fiduciary points): the right and left external
meatus and the nasion. The line passing through the right and left external meati
served as the y-axis. The vertical line between the nasion and the midpoint of the
y-axis defined the x-axis, and the line perpendicular to the x–y plane, passing
through the intersection of the x-axis and y-axis, defined the z-axis. The position of
the magnetometers relative to the participant’s head was precisely determined us-
ing five coils, three of which were attached to the fiduciary points and two on the
forehead. The coils were turned on briefly at the beginning and again at the end of
the recording session and their precise location in three-dimensional space was de-
termined using a localization algorithm built into the system. During the recording
session a fiberoptic motion detector was used to ensure that the participant did not
change position relative to the sensor.
Precise coregistration of the MEG coordinate system onto MRI structural images
was achieved by aligning the MEG fiduciary points with high contrast cod-liver cap-
sules (3 mm in diameter) that were affixed to the participant’s nasion and inserted in
the external meatus prior to the MRI scan. Orthogonal coronal, sagittal, and axial
PHONOLOGICAL DECODING 197
T1-weighted images were obtained using a GE Signa 1.5 T system (TR = 13.6 msec,
TE = 4.8 msec, recording matrix 256 × 256 pixels, 1 excitation, 240 mm field of view,
and 1.4 mm slice thickness). Finally, the anatomical location of each source was de-
termined using a standard MRI atlas (Damasio, 1995).
We have established the validity of this procedure for (a) identifying cortical
regions specialized for basic receptive language function (Papanicolaou et al.,
1999; Simos, Breier, Zouridakis, & Papanicolaou, 1998b; Simos, Breier, et al.,
1999; Simos, Papanicolaou, et al., 1999) and (b) providing accurate information
regarding the degree of regional activation associated with engagement in lin-
guistic tasks (Breier, Simor, Papanicolaou, et al., 1999; Papanicolaou et al.,
1999; Simos, Breier, Zouridakis, & Papanicolaou, 1998a; Zouridakis, Simos,
Breier, & Papanicolaou, 1998). For the latter purpose we have shown in the se-
ries of aforementioned studies that the number of satisfactory activity sources
computed in each hemisphere during the late portion of the EF provides a reli-
able and valid measure of interhemispheric asymmetry in activity specific to
language function. In addition to the degree of regional activation, our study ex-
amined the temporal course of the engagement of cerebral regions as indicated
by the onset (i.e., earliest latency in milliseconds after stimulus onset) of reliably
localized neurophysiological activity in a particular area.
RESULTS
The selection of stimuli was successful in ensuring similar performance levels
across age groups for both tasks. Mean percentage correct identification of targets
was 89.4% (± 5) for adults and 87.2% (± 4) for children during the real-word task (p
> .05). Mean percentage correct detection of rhyming pairs on the pseudoword task
was 73.6% (± 6) and 69.8% (± 7), respectively (p > .05).
The number of activity sources found in a given area and the earliest latency at
which these sources were computed served as the dependent measures in all subse-
quent analyses. Each measure was analyzed separately using an analysis of variance
with two within-subjects variables (Area: temporoparietal, inferior frontal, basal
temporal; and Hemisphere: left, right). Age Group (children, adults) served as the
between-subjects variable. When applicable, the Greenhouse–Geisser method
(Geisser & Greenhouse, 1958) was used as a precaution against the effects of viola-
tions of the sphericity assumption, and the Bonferroni method was used for control-
ling family-wise Type I error when pair-wise means comparisons were performed.
Temporoparietal cortex (TMP) included the posterior third of the superior temporal
gyrus, the angular, and supramarginal gyri, and basal temporal cortex (BTC) in-
cluded the fusiform and lingual gyri. In the left hemisphere, region TMP largely
overlaps with Wernicke’s area, and the inferior frontal gyrus (IFG) region is essen-
tially coextensive with Broca’s area. On the basis of previous research, these are the
198 SIMOS ET AL.
three regions that consistently display reliable activity sources in the context of read-
ing tasks similar to those used in our study.
Degree of Activation
The number of activity sources found in each of the six prespecified areas is shown
in Figure 1. With respect to TMP activation, all groups showed strong left-hemi-
sphere preponderance. In addition, the mean number of sources in the left TMP re-
gion was greater in children than in adults in response to both types of stimuli.
In contrast, children displayed less pronounced asymmetries than adults with
respect to the degree of basal temporal activation, especially in response to
pseudowords. Further, activity in IFG in children was very sparse during the
pseudoword task and altogether undetectable during the word task. These trends
were reflected in significant Area × Hemisphere × Age Group interactions found
for both types of print (see Table 1). Further analyses indicated that whereas chil-
FIGURE 1 Degree of regional activation indicated by the mean number of reliable activity
sources found in each area during the course of the evoked magnetic fields in the four groups of
participants (e.g., adults and children during the real-word recognition task and the pseudoword
rhyme-matching task). TMP = temporoparietal regions; IFG = inferior frontal gyrus.
TABLE 1
Analysis of Variance Results for Degree of Activation
Words Pseudowords
Interaction F p t p F p t p
Area × Hem × Age F(2, 52) = 3.34 < .047 F(2, 38) = 4.84 < .014
Left IFG: t(21.10) = 3.73 < .001a,b t(11.35) = 3.64 < .004b
Children < Adults
Left TMP: t(13.6) = 3.6 < .003b t(17.53) = 5.55 < .0001b
Children > Adults
Children Adults Children Adults
Interaction F p t p F p F p t p F p
Area × Hem F(2, 22) = 5.29 < .014 F(2, 18) = 15.55 < .0001
Left–TMP > t(11) = 3.58 < .004 t(9) = 5.82 < .0001
Right-TMP
Area
TMP > Basal, IFG F(2, 30) = 9.11 < .001 F(2, 20) = 4.10 < .037
Hem: Left > F(1, 15) = 15.46 < .001 F(1, 10) = 20.36 < .001
Right
Note. Hem = hemisphere; IFG = inferior frontal gyrus; TMP = temporoparietal region.
aAll t tests were two-tailed. bt test for unequal sample sizes.
200 SIMOS ET AL.
dren displayed significant Area × Hemisphere interactions, no such interactions
were found in the adult data, which in contrast displayed significant Area and
Hemisphere main effects. Inspection of Table 1 shows that although the hemi-
sphere effect (left > right) was significant across all three areas in the adult data,
the effect only reached significance for area TMP in the children’s data. In addi-
tion, as is evident in Figure 1, children as a group displayed stronger left TMP acti-
vation than adults for both types of stimuli. In contrast, adults as a group displayed
stronger activation in the left IFG compared to children during both tasks. These
trends were evident in the majority of individual activation profiles such as those
shown in Figures 2 and 3.
Temporal Course of Activation
In contrast to the degree of regional activation, the temporal course of activity in
different areas varied substantially as a function of stimulus type (see Figures 4
and 5). In response to word stimuli, the onset of regional activation showed a
similar course in both groups, featuring bilateral occipital activation up to ap-
proximately 140 msec (not shown in Figures 4 and 5), followed by left
(Children: 229 msec, standard error [SE] = 28; Adults: 194 msec, SE = 24) and
then right (Children: 300 msec, SE = 26; Adults: 340 msec, SE = 20) basal tem-
poral activation. Left TMP activation started next (Children: 326 msec, SE = 62;
Adults: 303 msec, SE = 28), followed by activation in right TMP regions
(Children: 432 msec, SE = 49; Adults: 368 msec, SE = 39). In addition, adults
reliably showed activity in the left IFG within the same time range (on average)
as TMP activity (mean onset latency was 282 msec, SE = 16). These observa-
tions were reflected in a significant Area main effect (see Table 2). Further, a
Hemisphere main effect indicated that, across age groups and stimulus types, the
onset of activity in the left basal and TMP regions preceded the onset of activity
in homologous right hemisphere regions.
The temporal course of activation associated with pseudoword reading dif-
fered across age groups. Adults showed a pattern similar to what was found dur-
ing word reading. It consisted of bilateral occipital (within approximately the
first 150 msec after stimulus onset), followed by left basal temporal (222 msec,
SE = 36), and subsequently by left TMP (579 msec, SE = 62) and left IFG acti-
vation (640 msec, SE = 92), with no significant time lag in the onset of activity
between the two latter areas. In contrast, children’s activation maps did not dis-
play a clear temporal distinction between left basal (260 msec, SE = 21) and left
TMP regions (297 msec, SE = 41), although there was a significant time lag be-
tween the onset of activity in both these regions and the onset of activity in the
left IFG area (886 msec, SE = 135).
FIGURE 2 Representative individual activation profiles obtained during performance of the
real-word recognition task. The anatomical location of activity sources that became active be-
tween 100 and 250 msec after stimulus onset is indicated by filled circles, whereas the location of
sources that became active between 250 and 450 msec is indicated by filled triangles. Finally, ac-
tivity sources that became active next, between 450 and 650 msec, are displayed as filled
squares. Early activity sources (< 100 msec after stimulus onset), which are typically localized in
posterior occipital cortices bilaterally, are not shown. LH = left hemisphere; RH = right hemi-
sphere.
201
FIGURE 3 Representative individual activation profiles obtained during performance of the
pseudoword rhyme-matching task. The anatomical location of activity sources that became ac-
tive between 100 and 250 msec after stimulus onset is indicated by filled circles, whereas the lo-
cation of sources that became active between 250 and 450 msec is indicated by filled triangles.
Finally, activity sources that became active next, between 450 and 650 msec, are displayed as
filled squares. Early activity sources (< 100 msec after stimulus onset), which are typically local-
ized in posterior occipital cortices bilaterally, are not shown. LH = left hemisphere; RH = right
hemisphere.
202
FIGURE 4 Time course of regional activation indicated by the mean latency of onset of reli-
able activity sources found in each area during the course of the evoked magnetic fields recorded
over the left hemisphere for the four groups of participants (e.g., adults and children during the
real-word recognition task and the pseudoword rhyme-matching task). TMP = temporoparietal
regions; IFG = inferior frontal gyrus.
FIGURE 5 Time course of regional activation indicated by the mean latency of onset of reli-
able activity sources found in each area during the course of the evoked magnetic fields recorded
over the right hemisphere for the four groups of participants (e.g., adults and children during the
real-word recognition task and the pseudoword rhyme-matching task). TMP = temporoparietal
regions; IFG = inferior frontal gyrus.
203
204
TABLE 2
Analysis of Variance Results for Onset of Activation
Words
Main Effect F p t p
Area F(1, 26) = 18.11 < .0001a
BTC® TMP
Left-BTC® Left-IFGb (adults only) t(15) = 2.98 < .009
Hem F(1, 26) = 14.43 < .001
Left® Right
Pseudowords
Children Adults
Interaction F p F p t p F p t p
Area × Hem × Age F(1, 19) = 17.54 < .0001
Area × Hem F(1, 9) = 8.63 < .017 F(1, 10) = 10.83 < .008
Left-BTC® Left-IFG t(9) = 6.63 < .0001 Left-BTC® Left-TMP t(10) = 3.38 < .007
Left-TMP® Left-IFG t(9) = 5.63 < .0001 Left-BTC® Left-IFG t(10) = 5.60 < .0001
Right-BTC® Right-TMP t(9) = 3.59 < .006 Left-BTC® Right-BTC t(10) = 3.37 < .007
Note. BTC = basal temporal cortex; TMP = temporoparietal region; IFG = inferiro frontal gyrus; Hem = hemisphere.
aOnly two areas (BTC, TMP) were included in the analysis of variance because only 2 children and 3 adults had activity sources in the right IFG, and only 3
children had activity in the left IFG. bPlanned comparison.
PHONOLOGICAL DECODING 205
DISCUSSION
Distinct spatiotemporal profiles of cerebral activation were obtained from children
and adult participants as shown in Figures 6 through 9. Although children’s activa-
tion profiles were dominated by activity in the left TMP, adult profiles indicate ac-
tivity that is distributed over more regions, including the right TMP and the left
IFG. In addition, visual association areas (in basal temporal regions) were activated
essentially in parallel with the left TMP in children, whereas adults showed a clear
precedence in activation of the former over the latter areas. It is noteworthy that al-
though TMP areas show dramatically reduced activation in children with severe
reading disability, little or no difference is typically found between children who
are reading impaired and children who are nonimpaired with respect to activation
of BTC regions (Simos, Breier, Fletcher, Bergman, & Papanicolaou, 2000; Simos,
Breier, Fletcher, Foorman, et al., 2000).
Each profile can be further described in terms of (a) the amount of regional acti-
vation and (b) the relative timing of this activation following the presentation of
printed stimuli. With respect to the amount of regional cerebral activation, several
age-related differences were noted. Specifically, children showed increased left
TMP engagement, reduced left IFG activation, and reduced hemispheric asymme-
tries for basal temporal areas compared to adults during reading of both real words
and pseudowords. It is important to note that both groups consisted of skilled read-
FIGURE 6 Group spatiotemporal activation profile displaying the mean latency of onset of
activity sources in each area. The degree of regional activation is indicated by the size of the
filled circles. Brain regions in both hemispheres have been sorted on the horizontal axis based on
onset latency. TMP = temporoparietal regions; IFG = inferior frontal gyrus.
FIGURE 7 Group spatiotemporal activation profile displaying the mean latency of onset of
activity sources in each area. The degree of regional activation is indicated by the size of the
filled circles. Brain regions in both hemispheres have been sorted on the horizontal axis based on
onset latency. TMP = temporoparietal regions; IFG = inferior frontal gyrus.
FIGURE 8 Group spatiotemporal activation profile displaying the mean latency of onset of
activity sources in each area. The degree of regional activation is indicated by the size of the
filled circles. Brain regions in both hemispheres have been sorted on the horizontal axis based on
onset latency. TMP = temporoparietal regions; IFG = inferior frontal gyrus.
206
PHONOLOGICAL DECODING 207
FIGURE 9 Group spatiotemporal activation profile displaying the mean latency of onset of
activity sources in each area. The degree of regional activation is indicated by the size of the
filled circles. Brain regions in both hemispheres have been sorted on the horizontal axis based on
onset latency. TMP = temporoparietal regions; IFG = inferior frontal gyrus.
ers for their age. In addition, similar performance levels (indicated by comparable
accuracy rates) is evidence against the possibility that differences between age
groups in spatiotemporal activation profiles were due to increased effort commit-
ted by the less fluent younger readers. It is therefore very likely that these findings
reflect real differences in at least one aspect of brain function that supports read-
ing, namely age-related changes in the specialization of brain areas actively in-
volved in component processes of reading. The second important aspect of brain
function, cortico-cortical integration, is discussed later.
It is noteworthy that, for each group, the degree and anatomical distribution of
activation was similar across tasks. This finding suggests that the activation pro-
files associated with word and pseudoword reading reflect engagement of
neurophysiological processes that are common to the brain mechanisms that sup-
port reading of both types of print. Such common processes include graphemic/or-
thographic and phonological analysis. Data from our group and others suggest that
basal temporal areas are involved in graphemic/orthographic analysis (Breier,
Simos, Zouridakis, et al., 1999; Pugh et al., 1996), and that TMP areas (particu-
larly the posterior part of the superior temporal gyrus) play a crucial role in phono-
logical decoding (Eden & Zefiro, 1998; Pugh et al., 1996; Simos, Breier, Wheless,
et al., 2000). Reduced hemispheric asymmetries in basal temporal activation may
indicate incomplete specialization of the left basal temporal areas for processing
print, even in older children who are skilled readers for their age. Increased ac-
tivation in left TMP areas displayed by children during reading of both real
208 SIMOS ET AL.
words and pseudowords may reflect increased reliance on phonological decoding,
not only when this process is explicitly required by the experimental task (per-
forming phonological similarity judgments) but also in the context of word recog-
nition (Simos et al., in press).
With respect to the temporal course of regional activation, both groups showed
similar patterns during word reading, consisting initially of bilateral occipital ac-
tivity sources followed by left basal temporal activation, which was in turn fol-
lowed by left TMP activation. In addition, during word reading, adults showed
considerable activity in the left IFG that was first noted, at approximately the same
time (on the average) as the onset of activity in TMP areas.
In response to pseudowords, however, the two groups showed distinct temporal
activation profiles. In children, the typical course of activation consisted of left and
right basal temporal and left TMP activation (with very little temporal lag between
basal and TMP sources) followed several hundred milliseconds later by activity in
the left IFG. Adults, on the other hand, first showed left basal temporal followed
by right basal and then by left TMP and IFG activation, with a negligible time lag
between the latter two areas.
The finding of a negligible time lag between the onset of left basal temporal and
left TMP activation in children may further suggest that operations typically per-
formed in the left BTC in adult readers are perhaps not relied on to the same extent
in children. Rather, reading in children may rely more on processes that depend
predominantly on TMP structures. Finally, IFG activation appears to characterize
the adult readers in this study. This conclusion is supported by the following find-
ings: (a) Activity in the left IFG is very sparse in children compared to adults dur-
ing reading of pseudowords (in fact, activity in this area is essentially undetectable
during reading of real words in children), and (b) even when present, activity in
this area starts very late after stimulus presentation (and significantly later than in
adults).
It should be noted that age-related changes found in our study might not reflect
those occurring during the early stages of reading skill acquisition. Rather, it is
possible that early changes feature a progressive increase in the left hemisphere
preponderance in the degree of regional activation (Licht, Bakker, Kok, & Bouma,
1992), especially in TMP areas. This change may reflect a gradual increase in the
degree of specialization of these areas for reading-related processes and may occur
in contrast with oral language, the hemispheric representation of which appears to
be established very early in development (Entus, 1977; Fennel, Satz, & Morris,
1983; Hahn, 1987; Hiscock, 1988; Molfese & Betz, 1988). We are now collecting
longitudinal data from sufficiently large groups of young children during the early
years of reading acquisition. Longitudinal designs permit the analysis of develop-
mental trends in brain activation profiles that accompany reading acquisition, with
greater fidelity than cross sectional designs, by controlling for individual variabil-
ity in brain function.
PHONOLOGICAL DECODING 209
ACKNOWLEDGMENT
The research presented in this article was partly supported by National Science
Foundation Grant REC–9979968 to Andrew C. Papanicolaou.
REFERENCES
Ackerman, P. T., Dykman, R. A., & Oglesby, D. M. (1994). Visual event-related potentials of dyslexic
children to rhyming and nonrhyming stimuli. Journal of Clinical and Experimental
Neuropsychology, 16, 138–154.
Breier, J. I., Simos, P. G., Papanicolaou, A. C., Zouridakis, G., Wilmore, L. J., Wheless, J. W., Constantinou,
J. C., & Maggio, W. W. (1999). Cerebral laterality assessment using magnetoencephalography: A com-
parison with the intracarotid amobarbital procedure. Neurology, 22, 938–945.
Breier, J. I., Simos, P. G., Zouridakis, G., & Papanicolaou, A. C. (1998). Relative timing of cortical acti-
vation during a word recognition task. Journal of Clinical and Experimental Neuropsychology, 20,
782–790.
Breier, J. I., Simos, P. G., Zouridakis, G., & Papanicolaou, A. C. (1999). Temporal course of regional ac-
tivation associated with phonological decoding. Journal of Clinical and Experimental
Neuropsychology, 22, 465–476.
Castro-Caldas, A., Peterson, K. M., Reis, A., Stone-Elander, S., & Ingvar, M. (1998). The illiterate
brain: Learning to read and write during childhood influences the functional organization of the
adult brain. Brain, 121, 1053–1063.
Damasio, H. (1995). Human brain anatomy in computerized images. New York: Oxford University
Press.
Eden, G. F., & Zefiro, T. A. (1998). Neural systems affected in developmental dyslexia revealed by
functional neuroimaging. Neuron, 21, 279–282.
Entus, A. K. (1977). Hemispheric asymmetry in processing of dichotically presented speech and
nonspeech stimuli by infants. In S. J. Segalowitz & F. A. Gruber (Eds.), Language development and
neurological theory (pp. 63–73). New York: Academic.
Fennel, E. B., Satz, P., & Morris, R. (1983). The development of handedness and ear listening asymme-
tries in relation to school achievement: A longitudinal study. Journal of Experimental Child Psy-
chology, 35, 248–262.
Geisser, S., & Greenhouse, S. W. (1958). An extension of Box’s results on the use of the F distribution in
multivariate analysis. Annals of Mathematical Statistics, 1, 113–125.
Hahn, W. K. (1987). Cerebral lateralization of function: From infancy through childhood. Psychologi-
cal Bulletin, 101, 376–392.
Hiscock, M. (1988). Behavioral asymmetries in normal children. In D. L. Molfese & S. J. Segalowitz (Eds.),
Brain lateralization in children: Developmental implications (pp. 85–170). New York: Guilford.
Liberman, A. M. (1998). Why is speech so much better than reading? In C. Hulme & R. M. Malatesha (Eds.),
Reading and spelling: Development and disorders. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Licht, R., Bakker, D. J., Kok, B., & Bouma, A. (1992). Grade-related changes in event-related potentials
(ERPs) in primary school children: Differences between two reading tasks. Journal of Clinical and
Experimental Neuropsychology, 14, 193–210.
Molfese, D. L., & Betz, J. C. (1988). Electrophysiological indices of the early development of
lateralization for language and cognition, and their implications for predicting later development. In
D. L. Molfese & S. J. Segalowitz (Eds.), Brain lateralization in children: Developmental implica-
tions (pp. 171–190). New York: Guilford.
210 SIMOS ET AL.
Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values
for 925 nouns. Journal of Experimental Psychology, 76, 1–25.
Papanicolaou, A. C., Simos, P. G., Breier, J. I., Zouridakis, G., Wilmore, L. J., Wheless, J. W.,
Constantinou, J. C., Gormley, W., & Maggio, W. W. (1999). Magnetoencephalographic mapping of
the language-specific cortex. Journal of Neurosurgery, 90, 85–93.
Poldarak, R. A., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. (1998). The neural basis of visual skill
learning: An fMRI study of mirror reading. Cerebral Cortex, 8, 1–10.
Pugh, K. R., Shaywitz, B. A., Constable, R. T., Shaywitz, S. E., Skudlarsky, P., & Fulbright, R. K.
(1996). Cerebral organization of component processes in reading. Brain, 119, 1221–1238.
Rumsey, J. M., Horwitz, B., Donohue, B. C., Nace, K., Maisog, J. M., & Andreason, P. (1997). Phonologi-
cal and orthographic components of word recognition. A PET–rCBF study. Brain, 120, 739–759.
Sarvas, J. (1986). Basic mathematical and electromagnetic concepts of the biomagnetic problem. Phys-
ics in Medicine and Biology, 32, 11–22.
Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cog-
nition, 55, 151–218.
Shaywitz, S. E., Shaywitz, B. A., Pugh, K. R., Fulbright, R. K., & Constable, R. T. (1998). Functional
disruption in the organization of the brain for reading in dyslexia. Proceedings of the National Acad-
emy of Sciences, 95, 2636–2641.
Simos, P. G., Breier, J. I., Fletcher, J. M., Bergman, E., & Papanicolaou, A. C. (2000). Cerebral mecha-
nisms involved in word reading in dyslexic children: A magnetic source imaging approach. Cere-
bral Cortex, 10, 809–816.
Simos, P. G., Breier, J. I., Fletcher, J. M., Foorman, B. R., Bergman, E., Fishbeck, K., & Papanicolaou,
A. C. (2000). Brain activation profiles in dyslexic children during non-word reading: A magnetic
source imaging approach. Neuroscience Letters, 290, 61–65.
Simos, P. G., Breier, J. I., Fletcher, J. M., Foorman, B. R., Castillo, E. M., & Papanicolaou, A. C. (in press).
Brain mechanisms for reading words and pseudowords: An integrated approach. Cerebral Cortex.
Simos, P. G., Breier, J. I., Maggio, W. W., Gormley, W., Zouridakis, G., Wilmore, L. J., Wheeles, J. W.,
& Papanicolaou, A. C. (1999). Atypical temporal lobe language representation revealed by MEG
and intraoperative stimulation mapping. Neuroreport, 10, 139–142.
Simos, P. G., Breier, J. I., Wheless, J. W., Maggio, W. W., Fletcher, J. M., Castillo, E. M., &
Papanicolaou, A. C. (2000). Brain mechanisms for reading: The role of the superior temporal gyrus
in word and pseudoword naming. Neuroreport, 11, 2443.
Simos, P. G., Breier, J. I., Zouridakis, G., & Papanicolaou, A. C. (1998a). Assessment of cerebral domi-
nance for language using magnetoencephalography. Journal of Clinical Neurophysiology, 15,
364–372.
Simos, P. G., Breier, J. I., Zouridakis, G., & Papanicolaou, A. C. (1998b). Identification of language-re-
lated brain activity using magnetoencephalography. Journal of Clinical and Experimental
Neuropsychology, 20, 706–720.
Simos, P. G., Papanicolaou, A. C., Breier, J. I., Wilmore, L. J., Wheeles, J. W., Constantinou, J. C.,
Gormley, W., & Maggio, W. W. (1999). Localization of language-specific cortex using MEG and
intraoperative stimulation mapping. Journal of Neurosurgery, 91, 787–796.
Thorndike, E. L., & Lorge, I. (1944). The teacher’s wordbook of 30,000 words. New York: Teacher’s
College, Bureau of Publications.
Williams, S. M. (1991). Handedness inventories: Edinburgh versus Annet. Neuropsychology, 5, 43–48.
Woodcock, R. W., & Johnson, M. B. (1989). Woodcock–Johnson Psycho-Educational Battery–Re-
vised. Allen, TX: DLM,.
Zeno, S. M., Ivens, S. H., Millard, R. T., & Duvvuri, R. (1995). The educator’s word frequency guide.
Touchstone Applied Science Associates.
Zouridakis, G., Simos, P. G., Breier, J. I., & Papanicolaou, A. C. (1998). Functional hemispheric asym-
metry assessment in a visual language task using MEG. Brain Topography, 11, 57–65.
