www.elsevier.com/locate/neulet
Received 26 May 1999; received in revised form 19 October 1999; accepted 21 October 1999
Abstract
The purpose of the study was to identify spatiotemporal brain activation proles associated with phonological decoding in dyslexic children using magnetic source imaging. For this purpose maps of regional cerebral activation were
obtained from eleven children diagnosed with dyslexia and ten children without reading problems during engagement
in a pseudoword rhyme-matching task. All dyslexic children showed aberrant activation maps consisting of reduced
activity in temporoparietal areas in the left hemisphere (including the posterior part of the superior temporal, angular
and supramarginal gyri) and increased activity in the right homotopic region. In contrast, the two groups of children did
not differ in the degree of activity in basal temporal areas that typically precedes temporoparietal activation. This is the
rst study to demonstrate the existence of distinct activation proles associated with phonological decoding in individual dyslexic children. q 2000 Published by Elsevier Science Ireland Ltd.
Keywords: Functional imaging; Magnetic source imaging; Magnetoencephalography; Dyslexia; Reading; Phonological decoding;
Temporal lobe
0304-3940/00/$ - see front matter q 2000 Published by Elsevier Science Ireland Ltd.
PII: S03 04 - 394 0( 0 0) 01 32 2- 7
62
discourage performing comparisons on the basis of orthographic information. Event-related magnetic elds (ERFs)
were recorded to the rst stimulus of each pair, to ensure
that the brain activity recorded corresponded to phonological decoding operations and did not reect the additional
cognitive operations that matching of the stimuli entails.
The responding hand was counterbalanced across children.
MSI data collection and analysis methods developed and
used in our laboratory have been described in detail elsewhere [11]. Briey, recordings were made in a magnetically
shielded room with a whole-head neuromagnetometer
(Biomagnetic Technologies). The intracranial generators
of the observed ERFs were modeled as single equivalent
current dipoles (ECDs) that were tted at successive 4 ms
intervals using the non-linear LevenbergMarquardt algorithm [16]. Although a variety of source modeling
approaches have been proposed by different investigators
[9,17] we decided to use a single-ECD source model that
is part of the BTi software. Alternative algorithms hold
many promises as tools for magnetic source localization.
However, they have not yet been validated against invasive
localization procedures. In contrast, there is currently a
wealth of data testifying to the validity of the single-ECD
model for reliably localizing and lateralizing neurophysiological activity associated with language functions
[2,11,20,22]. On the basis of this evidence, the singe-ECD
source model is part of the standard analysis protocol in
essentially all clinical applications of magnetoencephalography. For a given point in time, the ECD dipole tting
algorithm was applied to the magnetic ux measurements
obtained from a group of 3438 magnetometers, always
including both magnetic ux extrema. ECD computation
was restricted to those latency periods during which a single
pair of magnetic ux extrema dominated the left and/or the
right half of the head surface. ECD solutions were considered as satisfactory upon meeting two criteria: (1) a correlation coefcient of at least 0.90 between the observed and the
`best' predicted magnetic eld distribution, and (2) a 95%
condence volume of 3 cm 3 or smaller. Only ECDs
computed during the `late' portion of ERFs to the printed
pseudowords were considered in the analyses. Previous
studies have shown that these activity sources originate in
association cortices specialized for the analysis of linguistic
stimuli [1,11,22]. Early ERF components (i.e. those
recorded up to approximately 150 ms after stimulus onset)
originate in modality-specic (in this case visual) cortex and
probably reect early sensory analysis of the stimuli. The
procedure for precise coregistration of ECD coordinates on
structural, T1-weighted, MRI scans (TR 13.6 ms,
TE 4.8 ms, recording matrix 256 256 pixels, 1 excitation, 240 mm eld of view and 1.4 mm slice thickness) has
been described in detail elsewhere [19,27].
As expected, performance on the task was signicantly
higher for the non-dyslexic group (79.2 vs. 67.4% correct,
P , 0:05). The number of reliably localized activity
sources in temporoparietal (TMP) and basal temporal
cortices (BTC) in each hemisphere served as the dependent measure in the statistical analyses. These regions
were selected because they consistently displayed activity
sources in every child, in at least one hemisphere. TMP
areas included the posterior third of the superior (STGp)
and middle temporal gyrus (MTGp), the angular (ANG)
and supramarginal gyri (SMG). Selection of areas was
based on the results of our previous study [27], which
did not reveal any interactions involving these four
temporal and inferior parietal areas. Basal temporal
cortices included the fusiform and lingual gyri. As
previously mentioned, sources were computed during
the later portion of the evoked magnetic response thereby
excluding those which typically originate within modality-specic cortices. This measure has been found to be
the most reliable and valid index of the degree of regional, language-specic cerebral activation in several
studies involving neurologically intact volunteers and
patients [2,11,17,22,27].
Visual inspection of individual brain activation proles
indicated that all dyslexic children showed strong activation
in TMP regions in the right hemisphere with visibly weaker
activity in homotopic areas in the left hemisphere. The
opposite pattern was apparent for all non-dyslexic children
(with the exception of S20 who displayed slightly stronger
right than left temporoparietal activation). This highly
63
64
Dyslexics
Non-Dyslexics
a
BTC
Left
Right
Left
Right
474 (51)
297 (39)
351 (75)
415 (66)
268 (25)
260 (11)
436 (34)
248 (44)
by left TMP cortices in children with dyslexia. These operations may either be directly involved in the conversion of
print to sound (phonological decoding), specialized for
phonological analysis of speech sounds, or be common to
both processes [3,6,8,2325]. We are planning studies to
examine the developmental course that leads to the establishment of the activation prole seen in non-impaired readers.
These studies will address directly the proposition that the
distinct activation prole seen in older children with dyslexia
reects a functional decit involving left temporoparietal
areas or, alternatively, represents a more immature brain
mechanism supporting reading (we thank one of the
reviewers of an earlier version of this paper for suggesting
this possibility).
Regardless of the precise nature of this decit, the interindividual consistency of this phenomenon as revealed by
MSI underlines a potential use of this technique as a tool for
evaluating the effectiveness of educational intervention strategies that specically target the `phonological core' of
developmental reading disability.
[1] Breier, J.I., Simos, P.G., Zouridakis, G. and Papanicolaou,
A.C., Relative timing of cortical activation during a word
recognition task, J. Clin. Exper. Neuropsychol., 20 (1998)
782790.
[2] Breier, J.I., Simos, P.G., Papanicolaou, A.C., Zouridakis, G.,
Wilmore, L.J., Wheless, J.W., Constantinou, J.C., Maggio,
W.W., Cerebral laterality assessment using magnetoencephalography: a comparison with the Intracarotid Amobarbital Procedure, Neurology, 22 (1999) 938945.
[3] Breier, J.I., Simos, P.G., Zouridakis, G. and Papanicolaou,
A.C., Temporal course of regional activation associated
with phonological decoding: a MEG study, J. Clin. Exp.
Neuropsychol., 21 (1999) 465476.
[4] Eden, G.F. and Zero, T.A., Neural systems affected in
developmental dyslexia revealed by functional neuroimaging, Neuron, 21 (1997) 279282.
[5] Fletcher, J.M., Shaywitz, S.E., Shankweiler, D.P., Katz, L.,
Liberman, I.Y., Stuebing, C.K., Francis, D.J., Fowler, A.,
Shaywitz, B.A., Cognitive proles of reading disability:
comparisons of discrepancy and low achievement denitions, J. Edu. Psychol., 86 (1994) 623.
[6] Foorman, B.R., Francis, D.J., Fletcher, J.M. and Lynn, A.,
Relation of phonological and orthographic processing to
early reading: comparing two approaches to regressionbased, reading-level matched designs, J. Edu. Psychol., 4
(1996) 639652.
[7] Greenblatt, S.H., Subangular alexia without agraphia or
hemianopsia, Brain Lang., 3 (1976) 229245.
[8] Lyon, G.R., Toward a denition of dyslexia, Ann. Dysl., 45
(1995) 327.
[9] Mosher, J.C., Lewis, P.S. and Leahy, R., Multiple dipole
modeling and localization from spatio-temporal MEG
data, IEEE Trans. Biomed. Eng., 39 (1992) 541557.
[10] Olson, R.K., Kliegl, R., Davidson, B.J. and Foltz, G., Individual and developmental differences in reading disability, In
G.E. MacKinnon and T.G. Waller (Eds.), Reading Research:
Advances in Theory and Practice, Vol. 4, Academic Press,
Orlando, 1985, pp. 164.
[11] Papanicolaou, A.C., Simos, P.G., Breier, J.I., Zouridakis, G.,
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
65