IMAGING BRAIN STRUCTURE IN CHILDREN: DIFFERENTIATING LANGUAGE DISABILITY AND READING DISABILITY.
The purpose of this article is threefold. First, I will provide a tutorial on brain neuroanatomy for readers who are not familiar with the history and terminology of the organ of learning and behavior. Second, I will provide a brief overview of current knowledge on how brain structure differs in normal and poor readers. This knowledge has been greatly enhanced by structural magnetic resonance imaging of living children and adults. Finally, I discuss recent research in our laboratory and other laboratories indicating that reading disability is a heterogeneous disorder. Neuroanatomical differences between normally developing readers and those with reading disabilities may depend on whether the poor reader has a specific reading disability characterized by poor phonological awareness or an oral language disability characterized by difficulties with listening comprehension.
Specific reading disability in this context refers to a specific difficulty in understanding the sound structure of language, that is, phonological awareness -- representing or imaging phonemes and associating these images with letters (graphemes) (Shankweiler et al., 1995; Shaywitz et al., 1996). These poor readers are worse at reading nonsense words (where graphemephoneme associations are the only clue) than connected text (where oral vocabulary can also provide clues). Even with training, the ability to read nonsense words will never surpass the comprehension of connected text. Nonspecific reading disabilities associated with poor listening comprehension, by contrast, are characterized by difficulty associating meanings with graphemes and phonemes. This type of poor reader also has poor phonological awareness, but poor phonological awareness is not the only cause of his inability to comprehend text. The poor reader with poor listening comprehension has a small oral vocabulary, which places an upper bound on his ability to read. This type of poor reader is equally bad at reading nonsense words and connected text. With phonological training, the ability to read connected text will never surpass and may actually trail the ability to read nonsense words (Bishop & Adams, 1990).
Finally, this review will present evidence from imaging studies that different neurobiological risk factors characterize readers with listening comprehension and phonological disorders (Eckert & Leonard, 2000; Leonard et al., 2001a). Children with reading deficits that accompany general oral language impairment frequently have small, symmetrical auditory cortical regions whereas children with specific reading (not oral language) deficits tend to have extra gyri in auditory cortex and exaggerated cerebral and cerebellar asymmetries (Leonard et al., 2001a).
The view that comprehension and phonological ability are separate dimensions in reading ability is supported by the results of several recent studies (Catts, Fey, Zhang, & Toblin, 1999; Lombardino, Riccio, Hynd, & Pinheiro, 1997; Stringer & Stanovich, 2000). These studies show that oral language skills and general intelligence contribute variance to reading scores that is separate from that contributed by phonological skill. Computational models of reading and dyslexia also support the view that comprehension (semantics) and phonology are separable skills (Harm & Seidenberg, 1999; Manis, Seidenberg, Doi, McBride-Change, & Petersen, 1996).
In this article, I provide a rationale for looking at brain structure, describe the major brain regions of interest in the study of reading and language, summarize the technical and scanning issues in neuroimaging, present some results of structural imaging research, and discuss how developmental deviance in different neurobiological systems may be associated with oral and written language deficits.
Why Look at Brain Structure?
Language is a product of our brains. It is a biological species-specific behavior that has evolved to help humans adapt to their environment. No other animal can use arbitrary symbols to communicate about the past and the future. Because differences between the behavior of different species are due, in part, to differences in their brain structure, it is reasonable to suppose that differences between individuals could also be due, in part, to differences in their brain structure. This view is not biological determinism but rather assumes that neurobiological variation confers vulnerability to the effects of variation in the linguistic environment, as illustrated by an example outside the domain of reading.
Imagine that a child has a relatively large hand. A large hand has a large number of sensory receptors and a large cortical representation for the hand. The location of this cortical representation (called the hand knob) is actually visible with magnetic resonance imaging (Yousry et al., 1997). Reaching for octaves on the piano or the frets on a guitar is easier with a large hand. The easier it is to perform these musical activities, the more likely that practice will seem like fun. If practice is fun, it happens more often and the already large hand knob expands even more, because the size of a cortical representation increases with use. Many scientists have concluded that cortical size doesn't matter because functional imaging studies report that the extent of activated tissue decreases with practice (Raichle et al., 1994). What such studies actually show is that the overlap between fields decreases with practice because of more differentiated, higher resolution representations. It seems reasonable to expect that the number of functions that can be represented with high resolution depends on the amount of cortical tissue available.
Skilled musicians who started musical training early have larger auditory cortex and motor cortex (Amunts et al., 1996; Schlaug, Jancke, Huang, & Steinmetz, 1995). We do not yet know if the larger cortical areas predisposed the children to seek training or resulted from the training. In any case, this interaction between "innate talent" and experience is an example of a neurobiological Matthew effect -- the rich get richer and the poor get poorer (Stanovich, 1986). If, in the absence of an innate talent, participation in sports or music is encouraged and modeled by the parents, even a clumsy child will improve her musicianship and athleticism, and enlarge her cortical representations. It should be remembered, however, that a child with a neurobiologically driven aptitude for music or sports is more likely to seek out opportunities to exercise that aptitude, whether or not the opportunities are present at home.
The analogy to reading should be clear. Children with large cortical areas for processing auditory and visual word forms will, all other things being equal, have an advantage in processing language and text. If the environment is filled with language opportunities, reading is facilitated. Even if a child has small or mismatched auditory and visual areas, but parents and teachers work hard on reading skills and provide high-quality literacy experiences, special education may not be necessary. Such a child may still be at a disadvantage and when college comes may fail the foreign language requirement, or stumble on timed multiple-choice exams. Nevertheless, the child who has only neurobiological disadvantages is not as vulnerable as the child with both neurobiological and environmental disadvantages. All children should be identified early and given concentrated remediation, tailored to the individual pattern of deficits. In the sections below, we will describe where in the brain these neurobiological disadvantages might lie.
A Brief Introduction to Human Brain Anatomy
The brains of all vertebrates are organized according to a set of common principles. In the second week of fetal development, a longitudinal groove appears on the top of the embryo. This groove rounds into a tube that subsequently divides into a paired set of cerebral hemispheres and an unpaired brainstem. The cerebellum, consisting of a set of paired hemispheres attached to a core formed by the anterior lobe and vermis, forms on the back of the brainstem just anterior to the spinal cord. In humans, the cerebral hemispheres are so large that they cover both the brainstem and the cerebellum. The thalamus, a small oval structure at the front of the brainstem, serves as a waystation for channeling of incoming information from the internal and external environment.
The major difference between the brains of different mammals is in the size and folding pattern in their cerebral cortex. The cortex expands greatly during the weeks before birth, and because the head does not expand nearly as fast, the cortex ends up folding, like a crumpled piece of paper. The folds are called sulci and the exposed surface between sulci are called gyri. Two thirds of the cortical surface is found in the banks of the cortical sulci. The largest sulci are sometimes called fissures.
The MRI image of a brain section shown in Figure 1 comes from just lateral to the midline (a sagittal image) and shows the brain stem, the thalamus and one cerebellar and cerebral hemisphere. The gray areas in the figures are cortex while the white areas are the fibers that connect cortical areas, the brain stem and spinal cord. Each hemisphere has four lobes, named after the bones of the skull. The frontal, parietal, and occipital lobes can be seen in Figure 1 and the temporal lobe can be seen in Figure 2, which shows a sagittal image of the left hemisphere taken about 5 cm away from the midline (close to the edge of the hemisphere). The fissures that define the frontal and temporal lobes are indicated by arrowheads in Figure 2.
The cortex was called gray matter by early anatomists, and the name has stuck. Gray matter appears gray because it is composed mainly of cells and thickly branching terminal processes of axons and dendrites. The axon and dendrite terminations carry information between cells. The space between cells that is occupied by axonal and dendrite endings is called neuropile. Between the cortex and the brainstem are the fiber tracts (called white matter because of fatty sheathes that improve communication efficiency) that carry information in and out of the cortex, and small collections of cells called the basal ganglia -- the caudate, putamen and globus pallidus. In the middle of the white matter are large fluid-filled spaces: the lateral ventricles, one in each hemisphere. The ventricles are filled with cerebrospinal fluid (CSF), a watery liquid that also covers the brain. The bony skull and the CSF look black on the type of MRI scan used for visualizing brain structure.
The cortex that can be seen on the surface is called cerebral cortex or neocortex (for new cortex, to distinguish it from the "old" simpler cortex seen in reptiles and birds). Folded in on the medial surface is some old cortex -- archicortex -- called the hippocampus. It is very small -- about the size of the basal ganglia -- but it is essential for processing conscious experiences so that they can be retrieved from long-term memory. A neocortical region adjacent to the hippocampus, the parahippocampal gyrus, connects the hippocampus to the sensory processing regions. All input to and output from the hippocampus funnels through the parahippocampal gyrus.
The cerebellum and the basal ganglia work together with the cerebral hemispheres to translate intentions into actions. The cerebral cortex processes information from the environment and selects possible responses. These proposed responses are sent to the basal ganglia and the cerebellum. The basal ganglia and the cerebellum refine the instructions, like computer programmers who interpret recommendations of a management team. These instructions are then sent to the motor areas of the cerebral cortex, which regulate the final output. Neither the cerebellum nor the basal ganglia has much direct access to lower motor centers. They influence behavior through their effect on the cerebral cortex. The connections between the cerebral cortex, basal ganglia, and cerebellum are very plastic, that is, they are responsive to the effects of experience.
Visual, somatosensory and auditory systems relay in the thalamus before projecting to primary sensory cortex in the occipital, parietal and temporal lobes. Areas that receive projections from primary cortex are called unimodal sensory association cortex. Unimodal cortex projects to polymodal or heteromodal association cortex, so called because it processes more than one type of sensory input. Both the visual and the auditory thalamic systems have magnocellular and parvocellular components. In the visual system, the magnocellular component responds quickly, is sensitive to stimulus movement and location, and controls orienting movements and visual fixation. This system is sometimes called "where" (Ungerleider & Mishkin, 1982) or "how" system (Goodale & Humphrey, 1998). The parvocellular component responds more slowly and is sensitive to fine detail, and has been called the "what" system (Figure 3). One neurobiological risk factor for reading disability may be a defect in a magnocellular visual fixation system (Stein et al., 2000). It is also proposed that there is a magnocellular auditory analogue to the magnocellular visual system that is disrupted in dyslexia (Galaburda, Menard, & Rosen, 1994).
Cortical sulci and gyri. The pattern of sulci and gyri is very consistent in different species and reflects the amount of processing space needed for species-specific behaviors such as hoarding, echo detection, song learning, and speech. Visual animals have deep sulci in visual areas while olfactory animals have a deep sulcus that separates the large olfactory area from smaller nonolfactory areas. During development the cortex tends to fold at weak points, where the structure and layering patterns (and by implication, the information processed) suddenly change. The pattern of gyri may reflect the patterns of local and distant neural connections (Van Essen, 1997).
In humans, the folding patterns differ tremendously between individuals. Some of this variation is genetic. Studies in twins have found that the deeper parts of the sulci share more genetic variance than the superficial parts visible on the surface (Lohmann, Von Cramon, & Steinmetz, 1999). This finding suggests that the superficial arrangement of the sulci may depend more on extrinsic, mechanical constraints than on intrinsic functional constraints. The superficial appearance may also depend on connections formed during postnatal experience.
Human cerebral cortex is divided up into 52 regions, called Brodmann's areas, in honor of the German neuroanatomist who described them in the 19th century. The boundaries between regions are defined on the basis of subtle differences in the shapes, sizes, and layering patterns of neurons. Subsequent experimental neuroanatomical and neurophysiological research has demonstrated that these boundaries define functionally distinct areas. Anatomy actually matters! The area called BA 17, which has a thick layer of small granular cells but no large pyramidal cells is also called Visual 1 (V1) because it receives the primary visual projection from the eyes. This area is in the calcarine sulcus, which is labeled on Figure 1. The area in the anterior bank of the central sulcus, called BA 4, has giant pyramidal cells, and is called Motor 1 (M1) because stimulation causes quick muscle jerks produced by long axons that descend all the way from the cortex to the spinal cord. The posterior bank of the central sulcus has three different Brodmann areas (3, 2, 1). It is the site of primary somatosensory (touch, pain and joint position) processing. Neurologists and cognitive scientists who do functional imaging experiments have found that Brodmann's nomenclature is useful for describing the effects of brain damage and location of brain activity during cognitive tasks. A few of the Brodmann areas are labeled in Figure 3.
The boundaries between many Brodmann's areas occur in the depths of sulci. As a result, the location of sulci can be used to roughly define the limits of functional areas. Since sulci are visible with modern imaging techniques, it is possible to measure the size of different functional areas and compare them between people. Imaging experiments have shown correlations between the presence of sulci and functional activity (Crosson et al., 1999; Grosbras, Lobel, Van de Moortele, LeBihan, & Berthoz, 1999).
The largest fissure in the cortex is the sylvian fissure, which separates the temporal lobe from the frontal and parietal lobe. The lower bank of the sylvian fissure contains the primary auditory cortex on Heschl's gyms and auditory association cortex on the planum (Figure 4). The superior temporal sulcus, a speech-processing area, runs parallel and just below the sylvian fissure in the temporal lobe. In the left hemisphere, Broca's area is found anterior to (in front of) the precentral sulcus and Wernicke's area is found posterior to (behind) the postcentral sulcus (see Figure 4). These areas were named in the 19th century because clinicians observed a specific aphasia symptom after frontal damage (Broca's aphasia or nonfluent speech). When Wernicke's area is damaged, by contrast, speech is fluent but filled with errors (Wernicke's aphasia). There are no real boundaries for these areas and different textbooks put them in different places (Bogen & Bogen, 1976). A visible region in Broca's area is pars triangularis (see Figure 4). Wernicke's area includes parts of the supramarginal and angular gyri in the parietal lobe and the superior temporal gyrus in the temporal lobe.
Why measure size? Why would the size of functional areas matter and why might size be related to individual differences in ability? To explain why size might matter, it is necessary to explain some principles of cortical organization. The cortex is organized as a series of maps of the environment and the body. Different species have different types of maps. Species with large cortical surface areas have more cells and more maps than species with small extents of cerebral cortex (Jones, 1990). The more maps, the more stimulus and response features can be processed and organized. In addition, bigger maps have more detail. Larger, more numerous, more detailed maps enable more accurate information processing.
What is a cortical map? In the cortex, neurons that are close to one another have similar properties, that is, they are activated by similar stimuli and produce similar responses. This feature is referred to as topographic organization. The properties of neurons in a cortical map change in an orderly way. A cortical map is an actual physical representation of a stimulus or response dimension. The simplest example of a cortical map is V1, the area receiving the primary visual projection from the eyes. Due to the physics of the eye, each cell in the retina registers brightness contrast in a particular point in the environment. The nerve fibers leaving the eye stay close to their nearest neighbors and end up in the same topographic relationship in the cortex that they had in the eye. If you show the eye an arrow, for example, there will be a physical representation of that arrow lying along the calcarine fissure (see Figure 5). The longer the calcarine fissure, the greater the number of cells firing for each mm length of the arrow, and the greater the total activation in V1. Because different sensory inputs compete to control responses, a large cortical representation will influence behavior more than a small cortical representation. In A1 (primary auditory cortex), the cells map the frequency of the sound, not the location, because different parts of the cochlea in the ear are sensitive to different frequencies rather than locations. The maps in the cochlea and the auditory cortex are called tonotopic maps while the map in V1 is called a retinotopic map. S1 (primary somatosensory cortex) contains a the map of the body surface, while M1 contains a map of the muscles and joints.
One of the most important features of cortical maps is that they are distorted, so that information that is important activates a large number of neurons that occupy a large amount of space. In humans and other visual animals, the information falling on the center of the retina, called the fovea, activates many more cells and is mapped to a much larger region than information falling on the retinal periphery (see Figure 5). Humans have a visual mechanism, called foveation, that moves the head and eyes so that information of potential interest will fall on the cell-dense region of the fovea, and thus activate more cells in the cortex. More is generally better. When the point of the arrow is foveated, many more cortical cells will process information about the point than will fire to the rest of the arrow, even if the tail is many times larger than the point. It is easy to imagine how the size and distortion of cortical maps might promote or interfere with learning to read. In children with large foveal maps, for example, letters would loom large and associations with other stimuli, such as sounds, would tend to form faster than in children with small foveal maps. We will return to this point later.
The somatosensory and motor maps are equally distorted. The areas for the tongue, the lips and the fingers are huge, compared to the area for the trunk and legs. In the hamster and the mouse, by contrast, the area for the whiskers, the sensors used for exploring the environment, is huge, compared to the area for the forepaw. These species differences have arisen as a result of evolution. Different animals specialize for different ecological niches. If whiskers are useful, animals with large cortical representations for whiskers will survive and reproduce more, all other things being equal. If large fovea are useful, animals with appropriately distorted V1's will have an advantage.
There are two important principle to remember Evolutionary selection can only act on traits that vary and the trait that is selected for will vary with the environment (Dobzhansky, 1951). Humans function well in a tremendous range of environments and the traits rewarded in these different environments also cover a very broad range. Fine-motor skills aid a watchmaker but not a philosopher. Perfect pitch aids a musician but not a visual artist. It is reasonable to expect that variation in these skills is due in some part to variation and distortion of the cortical areas mapping information-processing abilities necessary for those skills. In fact, a recent study has found that musicians with perfect pitch have longer plana than nonmusicians or musicians without perfect pitch (Schlaug et al., 1995).
Technical Issues in Neuroimaging
What does this discussion have to do with imaging? Recent technical developments have made it possible to image the cortical sulci in individual healthy children in perfect safety, because magnetic resonance imaging (MRI) does not depend on radiation or X-rays. The child lies in a large, very powerful magnet for about 15 minutes, while a complicated computer program (called a pulse sequence) controls electric gradients that manipulate the spinning of water molecules in the brain. Because these spin properties depend on the local cellular environment, the signals from gray matter, white matter and cerebrospinal fluid are different.
In order to reconstruct an image, another computer program takes the signals and turns them into a two-dimensional array of numbers, which is called a digital image. Each location in the image is called a pixel (picture element) and is an average of the signals from one volume element (voxel) in the brain. The number representing each pixel is proportional to the strength of the signal at that location in the brain. The reason why a picture appears on the monitor is that a display program translates each number into a gray level. Low numbers are displayed as black and high numbers look white.
MRI is the most flexible imaging technique available. Because the signal depends to a large extent on a computer program (software), rather than on the design of the machine (hardware), many different types of brain images can be otained. As with all scientific techniques, it is essential to formulate the research question carefully and imagine what the data might look like before starting the study, because different types of image acquisitions are sensitive to different properties of neural tissue. No one technique is perfect for all questions. Once started, one is locked in to a particular pulse sequence. In a longitudinal study that is investigating changes over time in the same individual or collecting data over a long period, the protocol is fixed once data collection has started.
There have been steady improvements in MRI technology over the last 10 years. The main improvement that is relevant to structural imaging of the brain is the capacity for volumetric imaging -- acquiring a gapless series of thin sections of the whole brain in a few minutes. When the images are between 1 and 2 mm thick, it is possible to reconstruct the brain in any plane of section. In contrast, older techniques acquired 5 mm thick images separated by 2 mm gaps that could not be reconstructed into useable data.
The flexibility of MRI technology allows the acquisition of images that emphasize the contrast between CSF and white matter (T2), or images that emphasize the contrast between white matter and gray matter (T1). The volumetric methods produce T1 images. A negative characteristic of T1 images is that they are noisy, that is, the signals from gray and white matter are not uniform. Although the boundary between gray and white matter looks clear to our eyes (because the neurons in our cortex emphasize boundaries or edges), there are many pixels with the same gray level in both gray and white matter. There are also many pixels with gray levels intermediate between gray and white. Some of these pixels are in true transition zones between gray and white, but some have intermediate values because of an artifact called partial voluming or volume averaging. Volume averaging results when the signal is collected from a voxel of mixed tissue types. Volume averaging was a more serious problem in the old techniques that used thick sections with low resolution than in current volumetric techniques, but it still contributes to inaccuracy in the final image.
Turning anatomy into numbers: The segmentation problem. The major problem in trying to compare anatomy among individuals stems from difficulties in converting visual patterns into numbers. Scientific comparisons require numbers but translating patterns into numbers is never straightforward. There are two sets of problems -- segmenting the image into gray and white matter and characterizing individual differences in sulci in a meaningful way.
This uncertainty about the "real" tissue identity of individual pixels is called a segmentation problem. Many groups of computer scientists are trying to calculate the amount of gray and white matter automatically (Tang et al., 2000; Warfield, Kaus, Jolesz, & Kikinis, 2000). Computer scientists know that the problem is unsolved, but the pressure for results is so great that data from many semi-automated methods have been published. In a semi-automated classification the computer program offers the opportunity to change the pixel identity "by hand" (according to subjective perception of classification errors).
Although it is possible to teach scientists and technicians to do the hand adjustments reliably, it is difficult to judge the validity of the classification that results because the actual amount of gray and white matter in an individual brain is unknown. Nevertheless, several publications have reported developmental changes in gray and white matter volumes. Very few investigators report the actual values for gray and white matter volumes, so it is difficult to compare results between investigators. Another problem with the studies is that they are of "super controls," high-functioning children who have passed stringent inclusion criteria. For example, in a longitudinal NIH study, five children were rejected for every one scanned (Giedd et al., 1999), leaving a control group that is anything but "average." Although there are not yet enough reports on children for the replicability of gray matter segmentations to be established, there appears to be a gradual, apparently linear increase in white matter of between .5 and 1% a year (Giedd et al., 1999; Paus et al., 1999).
Measurement issues in quantifying brain anatomy. The shape, continuity, and branching patterns of most sulci vary among individuals. Each research group has taken a different approach to handling these individual differences. Computer scientists at the Montreal Neurological Institute (MNI) decided to ignore individual differences in sulcal patterns and average 300 brains. After this kind of image processing, only a few major sulci can be discerned because the position of most sulci is so variable that they disappear in the group average (Evans et al., 1992). The other extreme is to catalogue variations in sulci (Ono, Jubik, & Abernathy et al., 1990; Paus et al., 1996; Steinmetz, Ebeling, Huang, & Kahn, 1990). A middle position is to adopt a set of standard rules for defining comparable regions across brains (Caviness, Meyer, Makris, & Kennedy, 1996; Kennedy et al., 1998; Rademacher, Galaburda, Kennedy, Filipek, & Caviness, 1992). Such rules sometimes use easy-to-identify subcortical landmarks such as the corpus callosum or the anterior commissure as boundaries between anatomical subdivisions. Applying standard rules improves measurement reliability. Ignoring individual variability in sulcal position may, however, eliminate the consideration of anatomical information that is relevant to cognitive differences between individuals. Scientists at the MNI are now creating probabilistic atlases of anatomical variation in order to determine where each individual falls on a continuum of anatomical variation (Paus et al., 1996; Penhune, Zatorre, MacDonald, & Evans, 1996). Such atlases will be important for interpreting the results of functional imaging experiments. They may even prove useful for diagnostic purposes.
Validity. How can one assess the validity of the different approaches to quantifying brain anatomy? Another way of asking this question is to examine which dimensions of variation seem to correlate with behavioral variables such as handedness, cognitive function and diagnosis, or biological variables such as genes, sex, and family relationships. Ideally, different techniques would be compared on the same data set in order to determine the method that best elicits meaningful differences between groups. In the absence of such a principled approach, the field has been subjected to categorical statements about the best method (in general, that of the writer).
Researchers differ on whether volume or surface measurements are better indicators of cortical differences. In the early MRI studies, linear outlines were drawn on one thick section (Hynd, Semrud-Clikeman, Lorys, Novey, & Eliopulos, 1990). This kind of linear measurement fell out of favor when volumetric techniques became available. It is important, however, not to confuse these "linear" techniques with surface measurements that are made on a series of volumetrically acquired sections. This type of surface measurement has an anatomical rationale and is not simply dictated by technical limitations. The rationale underlying surface measurements is that the processing unit in the cerebral cortex is the cortical column, which lies perpendicular to the surface. The cortex is only 3 mm thick and this thickness is reasonably uniform across the cortex (with the exception of a few highly specialized areas). The surface of a sulcus has a standard relationship with the volume. Measurements of sulcal surfaces can be done quickly, avoiding the tissue segmentation problem, and have been shown to be sensitive to a number of biological and behavioral differences (Amunts et al., 1996; Eckert, Lombardino, & Leonard, 2001; Foundas, Leonard, & Heilman, 1995; Leonard et al., 1996; Steinmetz, Rademacher, & Fluang, 1989).
Research Findings in Language Disability and Dyslexia
The sylvian fissure. The structures that have been measured most frequently in studies of language and reading impairment are the auditory structures on the superior surface of the temporal lobe -- Heschl's gyrus and the planum temporale. Heschl's gyrus is one of the few structures that is present in every brain and can be recognized with little or no training (see Figure 3). It provides a convenient anterior boundary to the planum temporale (PT), a roughly triangular region that has the most robust known lateral asymmetry in the human brain. The earliest quantitative study was by Geschwind and Levitsky (1968), who measured planar length with calipers in 100 autopsy brains. The planum temporale was longer on the left in 65, equal on the two sides in 23 and longer on the right in 12. The authors proposed that this leftward asymmetry was the neurobiological substrate of left-hemisphere language dominance.
A subsequent study reported that the shape and slope of the sylvian fissure were different in the right and left hemispheres (Rubens, Mahwold, & Hutton, 1976). In the left hemisphere, the fissure slopes gently and is frequently capped by short ascending and descending branches, forming a sideways "T" (see Figure 6). The supramarginal gyrus (SM) is therefore, longer than it is tall and the angular gyrus (ANG) is more posterior and inferior. In the right hemisphere, the parietal planum (PP) tends to be long, the supramarginal gyrus is taller, and the angular gyrus is more anterior and superior (Figure 6). Several investigators have reported on the frequencies of these right/left differences using a qualitative category system (see Table 1) and proposed that these hemispheric asymmetries are related to functional hemispheric asymmetries in language and visuospatial function (Ide, Rodriguez, Zaidel, & Aboitiz, 1996; Steinmetz et al., 1990; Witelson & Kigar, 1992).
Table 1 Classification of Relationship Between Sylvian Fissure and Parietal Operculum Type Frequency Steinmetz Witelson Left Right Hem. Hem. Moderate length, terminal branches I HV 80% 65% Long, extra parietal sulci or no branches II-III H 15% <5% Very short IV V <5% 30%
Differences in planar asymmetry are visible in the last trimester of fetal life, although there are some indications that leftward asymmetry may increase with age, perhaps as a function of linguistic experience (Chi, Dooling, & Gilles, 1977; Sowell, Thompson, Jernigan, Homes, & Toga, 1999; Witelson & Paillie, 1973). Another possibility, suggested by Binder, is that the hemispheric differences in shape result from simple mechanical forces. During development, neuroplasticity driven by visual-spatial experiences coded in the right inferior parietal lobe might cause expansion and push the sylvian fissure forward and up (Binder et al., 1996). If individual differences in visual-spatial experience and ability are associated with differences in parietal expansion, one would predict that people with short planum temporale and long plana parietale might have an advantage at spatial and mathematical tasks that are performed in the posterior parietal regions (Dehaene, Spelke, Pinel, Stanescu, & Tsivkin, 2000).
There is mounting evidence for such a relationship. In our initial MRI study, the only two subjects with small PT and large PP, bilaterally, were dyslexic physicians in specialities requiring highly developed visuospatial ability (Leonard et al., 1993). Subsequently, Witelson and her colleagues have discovered that Albert Einstein had similar brain anatomy (Witelson, Kigar, & Harvey, 1999). In our longitudinal study of normal children we are finding an association with a long PP on the left and the broad mathematical achievement score on the Woodcock-Johnson tests of cognitive ability (Leonard & Lombardino, unpublished data).
Although the term planum temporale was originally coined to describe the posterior surface of the sylvian fissure in the temporal lobe, some authors have used the term to include both the PT, in the temporal lobe, and the PP, in the parietal lobe. This terminology is counterintuitive, because it puts a structure with the name temporale in the parietal lobe. This is not a trivial point because the PT and the PP have different population asymmetries (Steinmetz et al., 1990; Witelson & Kigar, 1992), particularly in right-handers (Steinmetz, Volkmann, Jancke, & Freund, 1991). When the planum temporale and planum parietale are summed together (P+), the opposing asymmetries cancel each other out and leftward asymmetry is greatly reduced (Leonard et al., 1993; Steinmetz et al., 1990).
The pronounced leftward asymmetry of the planum temporale has fostered the belief that the planum temporale and not the planum parietale is the site of phonological processing. The logic of this argument is that a structure that tends to be larger on the left should process information that is essential for language functions that are located on the left. Actually, we know very little about how linguistic and nonlinguistic auditory processing is distributed among the auditory cortical areas (Binder et al., 1996). The rationale for studying the planum therefore rests more on the ease with which it can be measured than on its known physiological correlates.
The planum in dyslexia research. The discovery that an anatomical correlate of functional asymmetry for language could be visualized with the naked eye had a tremendous impact on research in learning disabilities. Because dyslexia is considered to be a language disorder stemming from anomalous brain development, The Orton Society (now the International Dyslexia Society) set up a brain bank at Geschwind's suggestion, so that when individuals with dyslexia died, their plana could be examined. Eventually, eight brains were acquired and one of Geschwind's residents, Albert Galaburda, examined them. Galaburda found that all eight individuals had symmetrical plana due to expansion of the right planum, a finding consistent with anomalous development of hemispheric dominance (Geschwind & Galaburda 1987).
As is common in post-mortem studies, the subjects had multiple comorbid diagnoses, such as attention problems, seizures, language delay and mental retardation. Still, the idea that dyslexia had an anatomical basis in anomalous hemispheric asymmetry spread quickly, and is still one of the best known behavior-anatomy correlations in the human. Unfortunately, later work suggests that an enlarged right planum is not common in dyslexia (Beaton, 1997; Eckert & Leonard, 2000).
To understand the imaging literature in dyslexia, it is necessary to clarify the meaning of the word dyslexia. If this term is applied to all children with poor reading skills, then a larger-than-expected number of these children do indeed have symmetrical or right greater than left plana. If, on the other hand, it is reserved for children with a sizeable aptitude-achievement discrepancy, then most of those children do not have anomalous planar asymmetry (Filipek et al., 1995; Schultz et al., 1994).
In the paragraphs that follow, an attempt will be made to fit all the recent imaging results into a single framework. The results of a series of studies in ours and other laboratories converge on the following conclusions: (a) leftward planar asymmetry predicts good reading and language skills (Eckert et al., 2001; Gauger, Lombardino, & Leonard, 1997; Leonard et al., 1996; Rumsey et al., 1997); (b) children and adults with specific phonological deficits do not have symmetrical plana (Leonard et al., 1993; Leonard et al., 2001a); (c) phonological deficits are associated with extra gyri and sulci (see Figures 6 and 7) (Habib & Robichon, 1996; Hiemenz & Hynd, 2000); (d) different structural variables predict comprehension and phonological decoding deficits (Leonard et al., 2001); (e) variables such as socioeconomic status, age, handedness, and the presence of other anomalous brain structures influence the relationship between planar asymmetry and reading (Eckert et al., in press; Leonard et al., 2001).
The studies supporting these conclusions will be presented in historical sequence. After a series of imaging studies that, in general, supported the Galaburda results, the first study that did not find symmetry in dyslexia was published in 1993 (Leonard et al., 1993). There were many problems with this study -- the sample was small, it included both children and adults, the only measure of nonverbal ability was the Benton line orientation test, handedness was not measured quantitatively (all except one subject responded no to the question "Did they prefer to use their left hand for any skilled activity"), some of the subjects were related, and, most serious of all, history, not a set of defined research criteria, was used to diagnose the subjects. All the "dyslexic" subjects impressed clinicians with their significant reading deficits, given their cognitive capacities or achievements. This study might not be considered publishable today, but it was acceptable then, perhaps because it was the first dyslexia study to use the new volumetric techniques and visualize the planum in thin sections using anatomical guidelines. Another strength was that there were two control groups: one of nondyslexic family members and one of nonrelated children and adults from comparable backgrounds and cognitive ability.
The results were clear. Not a single dyslexic or control subject had symmetrical plana. In fact, asymmetries were more markedly leftward in the dyslexic than in either of the two control groups. The dyslexic participants did, however, have type II-III opercula and extra Heschl's gyri (Figure 7). There seemed to be a cumulative effect: The more anomalies in either hemisphere, the worse the performance on phonological decoding tests. Interestingly, the nondyslexic relatives also had increased incidence of type II-III opercula. Clark and Plante (1998) have also seen increased incidence of anomalous sulci in relatives of language impaired individuals.
Subsequently, several other studies of children and adults have confirmed that planar asymmetry is not anomalous in children and adults with specific reading disability (Best & Demb, 1999; Filipek, 1995; Habib & Robichon, 1996; Rumsey et al., 1997; Schultz et al., 1994). Meanwhile, however, studies of normal children and children with oral language impairment (Gauger et al., 1997; Leonard et al., 1996) that used exactly the same planar measurement techniques have reported planar symmetry in children with poor oral language. At first, it was hard to reconcile these findings with the dyslexia studies.
A review of all published studies on the planum (Eckert & Leonard, 2000; Eckert & Leonard, in press) suggests the following resolution of these discrepancies: Planar symmetry characterizes children whose poor reading skills stem from poor verbal ability or delayed language development. Marked leftward planar asymmetry, by contrast, characterizes children whose phonological decoding skills are unexpectedly poor, given their level of verbal ability and oral comprehension skills.
In a new study, Eckert has confirmed the relationship between planar asymmetry and reading skill in a sample of 39 sixth graders followed prospectively from kindergarten (Eckert et al., 2001). The sample was chosen to reflect the ethnic and income distribution of Alachua County, Florida. In this study most children had comparable phonological decoding and reading comprehension. (Oral language skills were not measured because the data were collected before we realized the theoretical importance of such measures.) There was a clear predictive relation between leftward planar asymmetry and all reading skills in right-handed children. As predicted by the epigenetic model, children who had both biological and sociological risk factors, that is, came from poor homes and had rightward planar asymmetry, performed in the severely impaired range, while their peers with leftward asymmetry had average reading scores. Children from middle-class homes, who had only the biological risk factor of rightward planar asymmetry, achieved average scores while their middle-class peers with leftward asymmetry scored in the superior range. This is an empirical demonstration that the Matthew effect applies to the interaction of brain structure and environment.
Oddly, planar asymmetry did not predict reading performance in the 12 nonright-handers. By chance, all the nonright-handers were average to good readers from middle-class homes. They also lacked other neural risk factors, which subsequent work suggests increase the probability of reading and language deficits (Leonard, Eckert, Lombardino, Given, & Eden, 2001b). So further work is necessary to verify the interaction of handedness, poverty and anatomical risk factors. A particularly intriguing aspect of this study is that in the eight right-handed children receiving special education services, there was a perfect relation between planar asymmetry and reading improvement between kindergarten and sixth grade. Those with leftward asymmetry were able to profit from remediation while those with rightward asymmetry were not. This preliminary result suggests that MRI screening has a potential role in predicting response to remediation.
Most children, including those with oral language impairments, have comparable phonological and comprehension skills (Bishop & Adams, 1990; Manis et al., 1996). We think that the relative levels of phonology and oral comprehension are the critical variables. Studies of specific reading disability where oral comprehension is better than phonology have not found the elevated incidence of symmetry or reversed planar asymmetry that has been reported in studies of oral language disorders where both phonological ability and oral comprehension are compromised (reviewed in Eckert & Leonard, 2000).
The inferior parietal lobule. One brain region that has been implicated repeatedly in dyslexia is the inferior parietal lobule, a complex region that includes the parietal operculum (supramarginal gyrus, angular gyrus, and planum parietale). Although measurements of the length and asymmetry of the planum parietale do not distinguish dyslexics and controls, other measures of this region do. Two published studies have used the Steinmetz technique to categorize-the sylvian fissure. Leonard et al. (1993) found an elevated incidence of type II and III patterns in a population of clinically diagnosed, well-educated dyslexics referred to above. In a subsequent study of carefully matched dyslexic children and controls, Hynd et al. (1998) found that both diagnosed and normal children with type III patterns had worse phonological processing skills than those with type I (normal) patterns. Although the frequency of the type III pattern was not elevated in children with reading disability, the four children who had type II patterns were all dyslexic boys.
Two other groups have found poor language skills associated with anomalous parietal opercula. In a new sample of dyslexic college students, those with type III patterns had worse scores on a number of auditory and phonological tests, in particular one that required them to blend two phonemes together (Eckert & Leonard, 2001). Habib and Robichon (1996) found that dyslexic students from an engineering school had symmetrical opercula and that asymmetry was inversely proportional to performance on a phonological categorization task. Each of these studies has found a different language skill associated with anomalous parietal morphology. Future work in this area should use comprehensive oral and written linguistic batteries.
Cumulative anatomical risk factors. Reading disability is a result of anomalous brain development that is not limited to one specific system. Two recent studies from our laboratory have investigated the ability of multiple anatomical variables to predict diagnosis. The anatomical measures used are described in Table 2. In a group of college students with RD, four variables -- the size of Heschl's second gyrus, the combined asymmetry of the planum temporale and planum parietale, cerebral asymmetry, and anterior cerebellar asymmetry -- distinguished students with a severe phonological decoding problem from RD students with poor verbal ability and average phonological decoding (Leonard et al., 2001). Another anatomical variable -- brain volume -- predicted scores on tests of verbal ability such as oral and written comprehension and verbal analogies.
Table 2 Anatomical Risk Factors for Language Impaired (LI) and Phonological Deficit (PD) Phenotypes Phenotype Structure and Rationale Measurement LI PD for Measurement Cerebral hemispheres: size Volume Reduced Increased of cortical maps Heschl's primary gyrus: Left Reduced Increased primary auditory cortex; surface functional dominance for area left-hemisphere language- processing areas Planum temporale: auditory Leftward Reduced Increased association cortex; functional asymmetry dominance for left-hemisphere language-processing areas Planum temporale & parietale; Leftward Reduced Increased anomalous intra- and asymmetry interhemispheric asymmetry Heschl's second gyrus; Left surface Reduced Increased anomalous gyral development area Cerebral hemispheres; Leftward Increased Reduced anomalous inter- asymmetry hemispheric asymmetry Anterior lobe, cerebellum; Leftward Reduced Increased control of eye movements; asymmetry sensorimotor control Note. One deviant measure does not noticeably increase the risk of a language or reading impairment. As the number of deviant measures increases, the risk of impairment increases. LI: Oral language impairment associated with poor listening comprehension; PD: Phonological deficit not associated with oral language deficits.
A more recent analysis of 142 children, which includes children studied by Eckert and Lombardino (Leonard, in preparation), identified additional anatomical risk factors for comprehension -- the size of Heschl's first gyrus, the site of primary auditory cortex, and reduced leftward asymmetry of the planum. Children with reduced brain volume, rightward asymmetry of the planum and small first Heschl's gyri were at high risk for specific language impairment, particularly if they had low levels of the phonological risk factors defined in the dyslexia study described in the preceding paragraph. All seven variables were weighted and combined to create a normalized anatomical risk factor score that varied along a continuum from -3 to 3. Children with scores lower than -.5 on this variable were defined as having a language impairment (LI) anatomical phenotype while children with scores higher than .5 were defined as having a phonological deficit (PD) anatomical phenotype.
The ability of the cumulative anatomical risk factor score to predict oral language impairments in children with reading disabilities has recently been confirmed in children undergoing reading remediation in a study directed by Eden and Given (Leonard et al., 2001b). Of the 22 children who received scans, all seven whose receptive and expressive language scores were below 80 had a LI anatomical phenotype.
The identification of multiple risk factors is consistent with modern concepts of distributed processing (Clark, 1997; Elman et al., 1996), which suggest that reading and language ability depends on the integrated activity of many diverse systems. One anomalous structure is very unlikely to cause a reading or language impairment, but as the number of anomalies grows, the risk increases because information processing in more and more structures is affected. There also appears to be some specificity. Marked asymmetries and duplicated Heschl's gyri affect phonological decoding ability more than comprehension, while reduced volume of the cerebral hemispheres and auditory structures in the left hemisphere (the first Heschl's gyrus and the planum temporale) affect comprehension more than phonological decoding. Table 2 summarizes the anatomical measures and their effect.
The effect of the anatomical risk factors on reading appears to be modulated by the environment. Figure 8 shows the reading scores on two subtests of the Woodcock-Johnson reading battery (Woodcock & Johnson, 1989) in children with different anatomical phenotypes. For each phenotype, the mean scores for poor children (receiving a lunch subsidy) are shown separately from those for middle-class children (not receiving a subsidy). As illustrated, there is a dramatic interaction between poverty and the effect of the anatomical phenotype. For middle-class children, anatomical phenotype has little effect on phonological decoding. In poor children, only those with the low-risk middle phenotype have above average scores. For passage comprehension, there is a nonsignificant trend for middle-class children with a PD phenotype to show better comprehension than phonological decoding, but all group means are above average. For poor children, once again, there is a dramatic effect of anatomy, with only children with the middle, low-risk phenotype scoring above average.
These preliminary data suggest the hypothesis that the low-quality linguistic input many poor children receive at home (Hart & Risley, 1995) may have a particularly devastating effect on children with anatomical risk factors for reading and language delay. Given the dramatic effects of environmental enrichment (Greenough, Black, & Wallace, 1987) on brain growth and learning ability, it seems possible that early linguistic input may prevent the effect of these anatomical risk factors. The fact that these risk factors did not appear to affect the middle-class children supports this view.
It is of particular interest that a number of children with the PD anatomical phenotype had normal reading scores but very slow visual inspection times (Grudnik et al., 2000). The fact that these children were from middle-class homes might suggest that a literate environment can influence phonological and comprehension skills but cannot increase the speed of basic sensory-processing mechanisms. It would be interesting to follow these children longitudinally to determine if they develop reading problems when challenged by more difficult material in high school and college.
How do we interpret these anatomical findings? We speculate that synchronous local activation of representations in cortical auditory, visual and motor maps is required for foveation, attention to visual detail in words, and automatizing phoneme-grapheme associations. We hypothesize that specific reading deficits stem from asynchrony between auditory, visual and motor activations. The lack of temporal correspondence results from a lack of spatial correspondence (duplications, exaggerated asymmetries), that is, "mismatched maps." Oral comprehension deficits, by contrast, are associated with low-resolution (small) maps of auditory and linguistic features, which may contribute to difficulty in segmenting meaningful units from fluent speech (Tallal, Miller, & Fitch, 1993). It is chastening to realize that this distinction between reading and listening deficits is not new. In 1990, Bishop and Adams summarized the findings of a follow-up study of language-delayed children in the following words:
Phonological proficiency is not the main determinant of reading acquisition: syntactic and semantic ability are responsible for the major part of variation in reading ability. We did not find any children who had significant reading or spelling difficulties in the context of otherwise normal verbal functions. The commonest type of reading problem was where the child had normal reading accuracy but poor comprehension for what had been read. These children typically also had poor understanding for spoken material. If specific reading retardation [is defined] as a difficulty in reading that is discrepant with both non-verbal and verbal abilities, then not a single child in this sample fitted this picture ... we question the view that a specific language impairment in the preschool period manifests itself as a specific reading disorder later on. Contrary to what we expected, children who grow out of their early language difficulties ... were not at risk for literacy problems. Children who still have evident language impairment at the age of 5.5 years are likely to have reading and spelling difficulties, but these will not be isolated problems, but will occur in the context of persisting deficits in comprehension and expression of spoken language. Our study indicates that expressive phonological competence does play a part in learning to read, but it suggests that the importance of phonological processing may have been overstated. Phonological factors are of particular theoretical interest because they seem able to explain variation in reading acquisition that is not accounted for in terms of other, more general, verbal abilities. However, it should be emphasized that other language skills exert the major influence on reading progress. (p. 1046)
The findings described in the preceding paragraphs suggest that there may be biological differences between children with poor language skills and those whose reading is unexpectedly poor given their language ability. Large groups of subjects with a range of abilities will be needed to verify that reading and listening disabilities differ in neurobiology and prognosis. Most anatomical studies of dyslexia have included small subject groups and have limited their analysis to only one variable. Other methodological weaknesses exist as well. The small number of subjects in most studies prevented the control of variables known to influence reading level or brain anatomy such as handedness, sex, general intelligence and home environment (Eckert & Leonard, 2000). Imaging technology has changed rapidly in the past 20 years, resulting in many improvements in our ability to visualize the brain. Until recently, different investigators have not had a collaborative approach to sharing data and techniques. Every group has developed its own techniques and rationale. The funding of a group of NIH pediatric study centers explicitly devoted to collaboration is a promising development and the next decade may finally see a consensus emerge concerning the structural underpinnings of reading and language development.
This chapter was written with the support of National Institute of Deafness and Communication Disorders grant RO1 DC002922. The work described could not have been accomplished without the participation of a large number of students, children and their families and my collaborators, Linda Lombardino and Mark Eckert.
Amunts, K., Schlaug, G., Schleicher, A., Steinmetz, H., Dabringhaus, A., Roland P. E., & Zilles, K. (1996). Asymmetry in the motor cortex and handedness. Neuroimage, 4, 216-222.
Beaton, A. (1997). The relation of planum temporale asymmetry and morphology of the corpus callosum to handedness, gender, and dyslexia: A review of the evidence. Brain and Language, 60, 255-322.
Best, M., & Demb, J. (1999). Normal planum temporale asymmetry in dyslexics with a magnocellular deficit. Neuroreport, 10, 607-612.
Binder, J., Swanson, S., Hammeke, T., Morris, G., Mueller, W., Fischer, M., Benbadis, S., Frost, J., Rao, S., & Haughton, V. (1996). Determination of language dominance using functional MRI. Neurology, 46, 978-984.
Bishop, D., & Adams, C. (1990). A prospective study of the relationship between specific language impairment, phonological disorders and reading retardation. Journal of Child Psychology and Psychiatry, 31, 1027-1050.
Bogen, J., & Bogen, G. (1976). Wernicke's region, where is it? Annals of the New York Academy of Sciences, 280, 834-843.
Catts, H., Fey, M., Zhang, X., & Toblin, J. (1999). Language basis of reading and reading disabilities; Evidence from a longitudinal investigation. Scientific Studies of Reading, 3, 331-362.
Caviness, V., Jr., Meyer, J., Makris, N., & Kennedy, D. (1996). MRI based topographic parcellation of human neocortex: An anatomically specified method with estimate of reliability. Journal of Cognitive Neuroscience, 8, 566-587.
Chi, J., Dooling, E., & Gilles, F. (1977). Left, right asymmetries of the temporal speech areas of the human fetus. Archives of Neurology, 34, 346-348.
Clark, A. (1997). Being there: Putting brain, body, and world together again. Cambridge, MA: MIT Press.
Clark, M., & Plante, E. (1998). Morphology of the inferior frontal gyrus in developmentally language disordered adults. Brain and Language, 61, 288-303.
Crosson, B., Sadek, J., Bobholz, J., Gokcay, D., Mohr, C., Leonard, C., Maron, L., Auerbach, E., Browd, S., Freeman, A., & Briggs, R. (1999). Medial frontal activity during word generation is centered within the paracingulate sulcus; An fMRI study of functional anatomy in 28 individuals. Cerebral Cortex, 9, 307-316.
Dehaene, S., Spelke, E., Pinel, P., Stanescu, R., & Tsivkin, S. (1999). Sources of mathematical thinking: Behavioral and brain-imaging evidence. Science, 284, 970-974.
Dobzhansky, T. (1951). Genetics and the origin of species, third ed. New York: Columbia University Press.
Eckert, M., & Leonard, C. (2000). Structural imaging in dyslexia; the planum temporale. Mental Retardation and Developmental Disabilities Research Reviews, 6, 198-206.
Eckert, M., & Leonard, C. (in press). Developmental disorders: Dyslexia. In K. Hugdahl & R. Davidson (Eds.), Brain asymmetry. Cambridge, MA: MIT Press.
Eckert, M., Lombardino, L., & Leonard, C. (2001). Tipping the environmental playground: Who is at risk for reading failure. Child Development, 72, 988-1001.
Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking innateness: A connectionist perspective on development. Cambridge, MA: MIT Press.
Evans, A., Marrett, S., Neelin, P., Collins, L., Worsley, K., Dai, W., Milot, S., Meyer, E., & Bub, D. (1992). Anatomical mapping of functional activation in stereotactic coordinate space. Neuroimage, 1, 43-53.
Filipek, P. (1995). Neurobiological correlates of developmental dyslexia: How do dyslexic brains differ from those of normal readers? Journal of Child Neurology, 10, 62-69.
Foundas, A., Leonard, C., & Heilman, K. (1995). Morphological cerebral asymmetries and handedness: The pars triangularis and planum temporale. Archives of Neurology, 52, 501-508.
Galaburda, A., Menard, M. T., & Rosen, G. D. (1994). Aberrant auditory anatomy in developmental dyslexia. PNAS, 91, 8010-8013.
Gauger, L., Lombardino, L., & Leonard, C. (1997). Brain morphology in children with specific language impairment. Journal of Speech, Hearing & Language Research, 40, 1272-1284.
Geschwind, N., & Galaburda, A. (1987). Cerebral lateralization: Biological mechanisms, association and pathology. Cambridge, MA: MIT Press.
Geschwind, N., & Levitsky, W. (1968). Human brain: Left-right asymmetries in temporal speech region. Science, 161, 186-187.
Giedd, J., Blumenthal, J., Jeffries, N., Castellanos, F., Liu, H., Zijdenbos, A., Paus, T., Evans, A., & Rapoport, J. (1999). Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience, 2, 861-863.
Goodale, M., & Humphrey, G. (1998). The objects of action and perception. Cognition, 67, 181-207.
Greenough, W., Black J. E., & Wallace, C. (1987). Experience and brain development. Child Development, 58, 539-559.
Grosbras, M., Lobel, E., Van de Moortele, D., LeBihan, D., & Berthoz, A. (1999). An anatomical landmark for the supplementary eye fields in human revealed with functional magnetic resonance imaging. Cerebral Cortex, 9, 705-711.
Grudnik, J., Leonard, C., Kranzler, J., Oakland, T., Eckert, M., & Lombardino, L. (2000). Auditory duplications predict slow inspection time in children and adults. Society for Neuroscience Abstracts, 26.
Habib, M., & Robichon, F. (1996). Parietal lobe morphology predicts phonological skills in developmental dyslexia. Brain and Cognition, 32, 139-142.
Harm, M., & Seidenberg, M. (1999). Phonology, reading acquisition and dyslexia: Insights from connectionist models. Psychological Review, 106, 491-528.
Hart, B., & Risley, T. (1995). Meaningful differences. Baltimore, MD: Paul Brookes.
Hiemenz, J., & Hynd, G. (2000). Sulcal/gyral pattern morphology of the perisylvian language region in developmental dyslexia. Brain and Language, 74, 113-133.
Hynd, G., Semrud-Clikeman, M., Lorys, A., Novey, E., & Eliopulos, D. (1990). Brain morphology in developmental dyslexia and attention deficit disorder/hyperactivity. Archives of Neurology, 47, 919-926.
Ide, A., Rodriguez, E., Zaidel, E., & Aboitiz, F. (1996). Bifurcation patterns in the human sylvian fissure: Hemispheric and sex differences. Cerebral Cortex, 6, 717-725.
Jones, E. (1990). Modulatory events in the development and evolution of primate neocortex. In E. Jones (Ed.), The cerebral cortex (pp. 311-362). New York: Plenum Press.
Kennedy, D., Lange, N., Makris, N., Bates, J., Meyer, J., & Caviness, V. (1998). Gyri of the human neocortex: An MRI-based analysis of volume and variance. Cerebral Cortex, 8, 372-384.
Leonard, C., Eckert, M., Lombardino, L., Oakland, T., Kranzler, J., Mohr, C., King, W., & Freeman, A. (2001). Anatomical risk factors for phonological dyslexia. Cerebral Cortex, 11, 148-157.
Leonard, C., Eckert, M., Lombardino, L., Given, B., & Eden, G. (2001b). Two anatomical phenotypes for reading impairment. Journal of Cognitive Neuroscience, 12(abs).
Leonard, C., Lombardino, L., Mercado, L., Browd, S., Breier, J., & Agee, O. (1996). Cerebral asymmetry and cognitive development in children: A magnetic resonance imaging study. Psychological Science, 7, 79-85.
Leonard, C., Voeller, K. S., Lombardino, L., Morris, M. K., Alexander, A., Andersen, H., Garofalakis, M., Hynd, G., Honeyman, J., Mao, J., Agee, O., & Staab, E. (1993). Anomalous cerebral structure in dyslexia revealed with magnetic resonance imaging. Archives of Neurology, 50, 461-469.
Lohmann, G., von Cramon, D., & Steinmetz, H. (1999). Sulcal variability of twins. Cerebral Cortex, 9, 754-763.
Lombardino, L., Riccio, C., Hynd, G., & Pinheiro, S. (1997). Linguistic deficits in children with reading disabilities. American Journal of Speech-Language Pathology, 6, 71-78.
Manis, F., Seidenberg, M., Doi, L., McBride-Chang, C., & Petersen, A. (1996). On the bases of two subtypes of developmental dyslexia. Cognition, 58, 157-195.
Ono, M., Jubik, S., & Abernathy, C. (1990). Atlas of the cerebral sulci. New York: Thieme.
Paus, T., Tomaiuolo, F., Otaky, N., MacDonald, D., Petrides, M., Atlas, J., Morris, R., & Evans, A. (1996). Human cingulate and paracingulate sulci; Pattern, variability, asymmetry, and probabilistic map. Cerebral Cortex, 6, 207-214.
Paus, T., Zijdenbos, A., Worsley, K., Collins, D., Blumenthal, J., Giedd, J., Rapoport, J., & Evans, A. (1999). Structural maturation of neural pathways in children and adolescents: In vivo study. Science, 283, 1908-1911.
Penhune, V. B., Zatorre, R. J., MacDonald, J. D., & Evans, A. C. (1996). Interhemispheric anatomical differences in human primary auditory cortex; Probabilistic mapping and volume measurement from magnetic resonance scans. Cerebral Cortex, 6, 661-672.
Rademacher, J., Galaburda, A., Kennedy, D., Filipek, P., & Caviness, V. (1992). Human cerebral cortex: Localization, parcellation, and morphometry with magnetic resonance imaging. Journal of Cognitive Neuroscience, 4, 352-374.
Rubens, A., Mahwold, M., & Hutton, J. (1976). Asymmetry of the lateral sylvian fissures in man. Neurology, 26, 620-624.
Rumsey, J., Donahue, B., Brady, D., Nace, K., Giedd, J., & Andreason, P. (1997). A magnetic resonance imaging study of planum temporale asymmetry in men with developmental dyslexia. Archives of Neurology, 54, 1481-1489.
Schlaug, G., Jancke, L., Huang, Y., & Steinmetz, H. (1995). In vivo evidence of structural brain asymmetry in musicians. Science, 267, 699-701.
Schultz, R., Cho, N., Staib, L., Kier, L., Fletcher, J., Shaywitz, S., Shankweiler, D., Katz, L., Gore, J., Duncan, J., & Shaywitz, B. (1994). Brain morphology in normal and dyslexic children: The influence of sex and age. Annals of Neurology, 35, 732-742.
Shankweiler, D., Crain, S., Katz, L., Fowler, A., Liberman, A., Brady, S., Thornton, R., Lundquist, E., Dreyer, L., Fletcher, J., Stuebing, K., Shaywitz, S., & Shaywitz, B. (1995). Cognitive profiles of reading-disabled children; Comparison of language skills in phonology, morphology, and syntax. Psychological Science, 6, 149-156.
Shaywitz, S., Shaywitz, B., Pugh, K., Skudlarski, P., Fulbright, R., Constable, R., Bronen, R., Fletcher, J., Liberman, A., Shankweiler, D., Katz, L., Lacadie, C., Marchione, K., & Gore, J. (1996). The neurobiology of developmental dyslexia as viewed through the lens of functional magnetic resonance imaging technology. In G. Lyon & J. Rumsey (Eds.), Neuroimaging; A window to the neurological foundations of learning and behavior in children (pp. 79-94). Baltimore, MD: Paul Brookes.
Sowell, E., Thompson, P., Jernigan, T., Holmes, C., & Toga, A. (1999). Progression of structural changes in the human brain during the first three decades of life: In vivo evidence for post-adolescent frontal and striatal maturation. Society for Neuroscience Abstracts, 25.
Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-407.
Stein, J., Richardson, A., & Fowler, M. (2000). Monocular occlusion can improve binocular control and reading in dyslexics. Brain, 123, 164-170.
Steinmetz, H., Ebeling, U., Huang, Y., & Kahn, T. (1990). Sulcus topography of the parietal opercular region: An anatomic and MR study. Brain and Language, 38, 515-533.
Steinmetz, H., Rademacher, J., & Huang, Y. (1989). Cerebral asymmetry: MR planimetry of the human planum temporale. Journal of Computer Assisted Tomography, 13, 996-1005.
Steinmetz, H., Volkmann, J., Jancke, L., & Freund, H.-J. (1991). Anatomical left-right asymmetry of language related temporal cortex is different in left- and right-handers. Annals of Neurology, 29, 315-319.
Steinmetz, H., Rademacher, J., Jancke, L., Huang, Y., Thron, A., & Zilles, K. (1990). Total surface of temporoparietal intrasylvian cortex; Diverging left-right asymmetries. Brain and Language, 39, 357-372.
Stringer, R., & Stanovich, K. (2000). The connection between reaction time and variation in reading ability; Unraveling covariance relationships with cognitive ability and phonological sensitivity. Scientific Studies of Reading, 4, 41-54.
Tallal, P., Miller, S., & Fitch, R. H. (1993). Temporal processing in the nervous system; Implications for the development of phonological systems. Annals of the New York Academy of Sciences, 682, 27-47.
Ungerleider, L., & Mishkin, M. (1982). Two cortical visual systems. In D. Ingle, M. Goodale & R. Mansfield (Eds.), Analysis of visual behavior (pp. 549-586). Cambridge, MA: MIT Press.
Van Essen, D. (1997). A tension-based theory of morphogenesis and compact wiring in the central nervous system. Nature, 385, 313-318.
Witelson, S., & Kigar, D. (1992). Sylvian fissure morphology and asymmetry in men and women: Bilateral differences in relation to handedness in men. Journal of Comparative Neurology, 323, 326-340.
Witelson, S., & Paillie, W. (1973). Left hemisphere specialization for language in the newborn: Neuroanatomical evidence of asymmetry. Brain, 96, 641-646.
Woodcock, R., & Johnson, M. (1989). Woodcock-Johnson Psycho-Education Battery-Revised. Chicago: Riverside Publishing Co.
Yousry, T., Schmid, U., Alkadhi, H., Schmidt, D., Peraud, A., Buettner, A., & Winkler, P. (1997). Localization of the motor hand area to a knob on the precentral gyrus: A new landmark. Brain, 120, 141-157.
Requests for reprints should be addressed to: Christiana M. Leonard, P.O. Box 100244, McKnight Brain Institute, University of Florida, Gainesville, FL 32610. firstname.lastname@example.org
CHRISTIANA M. LEONARD, Ph.D., is professor of neuroscience, McKnight Brain Institute, University of Florida.
|Printer friendly Cite/link Email Feedback|
|Author:||Leonard, Christiana M.|
|Publication:||Learning Disability Quarterly|
|Date:||Jun 22, 2001|
|Previous Article:||CURRENT UNDERSTANDING OF THE GENETIC BASIS OF READING AND SPELLING DISABILITY.|
|Next Article:||THE USE OF BRAIN ELECTROPHYSIOLOGY TECHNIQUES TO STUDY LANGUAGE: A BASIC GUIDE FOR THE BEGINNING CONSUMER OF ELECTROPHYSIOLOGY INFORMATION.|