|Home | About | Journals | Submit | Contact Us | Français|
One of the core defining components of autism is impairment in communication, typically manifested as a delay in speech development. To date, neuroimaging studies have shed limited light on the mechanisms behind delay in speech development in autism. We performed magnetoencephalographic based auditory language mapping in two cases of high-functioning autism. Overall, two distinct characteristics were found: the use of atypical language pathways and cortical hyperexcitability. These neurophysiological findings parallel those reported in two other developmental disorders, developmental dyslexia and Rett syndrome. We discuss common mechanisms that may account for cognitive delays across these developmental disorders.
The neurobiological mechanisms underlying autism remain largely unknown. Although preventing or curing autism is the ultimate goal, such a possibility does not appear to be within our grasp at this time. Thus, optimizing cognitive functioning may be a more pragmatic, but important, goal. In order to design optimal cognitive rehabilitation treatments, the organization of neural networks as well as the dynamic changes in the structure of these neural networks with cognitive development and remediation needs to be better understood. Because language is such an important aspect of human development, much interest is focused on remediating the language deficits found in autism.
Neuroimaging has provided some insight into the organization of the cognitive language processing network in individuals with autism. Functional magnetic resonance imaging and magnetoencephalography studies suggest that the brain may not process language in the same way in typically developing individuals and individuals with autism. During language tasks, functional magnetic resonance imaging activation is abnormally decreased or increased in typical language areas and increased in atypical cortical areas, in individuals with autism as compared to typically developing individuals.1–5
Magnetoencephalography has advantages over functional magnetic resonance imaging. Because it directly measures the magnetic fields emitted by active synapses, instead of the indirect blood-oxygen signal measured with functional magnetic resonance imaging, it allows the millisecond-by-millisecond temporal evolution of cortical activation pattern to be studied. This may be important since the timing of cortical activation is different in individuals with developmental cognitive disorders as compared to typically developing individuals. Second, the morphology and neurophysiological characteristics of the evoked magnetic field can be examined.
Previous magnetoencephalography studies have shown that the auditory evoked field follows an opposite pattern of hemispheric dominance maturation in individuals with autism as compared to typically developing individuals. While both typically developing individuals and individuals with autism demonstrated reduced lateralization to syllable stimuli in childhood, as age increases cortical activation has been shown to become left lateralized in typically developing individuals and right lateralized in individuals with autism.6 This suggests that individuals with autism eventually overuse the right hemisphere to process language, but that this language dominance does not develop until adolescents. Interestingly, evoked potential studies demonstrate similar findings for older children with Asperger Syndrome.7
None of the magnetoencephalography studies that have examined individuals with autism have used functional mapping to examine the spatiotemporal dynamics of cortical activation during language tasks. We provide two examples of the spatiotemporal dynamics of cortical activation during simple auditory receptive language tasks in an older child and an adolescent with high-functioning autism. Our results suggest that children with autism may demonstrate patterns of cortical reorganization similar to those indentified in developmental dyslexia,8,9 and cortical hyperexcitability, not unlike the neurophysiological characteristics described in Rett syndrome.10
Participant A was a 16 year old male with an unremarkable medical history and family history significant for a brother of learning disabilities. He experienced language and social regression with development of stereotypic behavior at 18 months of age. Elements of his childhood developmental history were diagnostic of Autistic Disorder.11 Intensive speech therapy resulted in slowly, but gradually, improvement in speech. He continues to have poor social skills, obsessive and repetitive behaviors, sensory adversities and cognitive perseveration. He was diagnosed with attention deficit disorder at 8 years of age but stimulants were ineffective. Electroencephalogram and sleep study were normal. At 14 years of age he was reevaluated and received the diagnoses of Learning Disorder, Not Otherwise Specified and Pervasive Developmental Disorder, Not Otherwise Specified. His full scale intelligence quotient, as determined by the Wechsler intelligence Scale for Children 4th Edition, was 84 and considered to be in the low average range. On examination he demonstrated reduced initiation and maintenance of eye contact and social interactions, mild to moderate psychomotor delay, flat monotone dysfluent speech. At times he smiled, laughed and giggled spontaneously but inappropriately.
Participant B was an 8 year old boy with marked speech and social delay. Elements of his childhood developmental history were diagnostic of Autistic Disorder.11 A generalized seizure disorder was diagnosed at approximately 1 year of age and was controlled by valproic acid. He had particular difficulties with reading, following auditory directions and is described as a visual learner. He had perseverative interests and did not understand peer-to-peer relationships. Electroencephalogram demonstrated intermittent focal slowing and intermittent sharp waves over the right posterior quadrant. On examination he demonstrated speech that was poorly modulated in rate and volume and he used repetitive stereotypic phrases. He had good eye contact, but decreased reciprocity. Prior to the magnetoencephalography scan, the participant completed the Comprehensive Test of Nonverbal Intelligence which revealed a score of 85 which is one standard deviation below average and is considered to be in the low average range.
After description of the study to the participant and parent, written informed consent was obtained in accordance with our institutional review board regulations for the protection of human subjects. A whole-head 248-channel axial gradiometer system (4D, 3600, San Diego, Ca) was used for magnetoencephalography recording in both cases. Standard procedures for participant preparation and stimulus delivery were used.12 Participant A performed the continuous recognition memory task twice.12 This task required the participant to indicate whether any of 135 sequentially presented aural words match words on a pretest list. Participant B performed two auditory word rhyme tasks. In each task the participant heard 64 sequentially presented word pairs and indicated if the pairs rhymed. Equivalent results were found for both runs for both participants. Auditory evoked field waveforms13 and the significant functional mapped current estimates14 are presented. An example of an auditory evoked field from a typically developing adolescent is provided for comparison to our participants.
Although the early N1m component is within reasonable limits (Table 1), the late components are abnormal. Instead of a normal N400m which is typically characterized by a slowly increasing field with a peak between 400–500ms, a very sharp peak occurred at 350ms and 358ms in the left and right hemispheres, respectively. In addition, a third prominent peak occurred at 544 ms in the right hemisphere (Figure 1A). Functional localization demonstrated bilateral activation of the superior temporal gyrus near the auditory cortices and the right, but not the left, inferior frontal gyrus. The current estimate waveforms from the right superior temporal gyrus clearly demonstrated peculiar recurrent activation (Figure 2A). This pattern of activation suggests overuse of right hemisphere language pathways.
The auditory evoked field demonstrated a very high amplitude N1m peak in the left and right hemispheres (Table 1; Figure 1B) with no discernable late component. Functional localization demonstrated significant activation of the left superior temporal gyrus and inferior frontal gyrus. Current estimates from these areas demonstrated earlier activation in the left inferior frontal gyrus (~155ms) with later activation of the left (~200ms) and right (~220ms) superior temporal gyrus – a reverse of the typical pattern of activation (Figure 2B).
This report outlines two cases of autism with two neurophysiological abnormalities in common. First, the auditory evoked field waveforms demonstrated clearly abnormal morphology. Second, functional localization demonstrated cortical reorganization of the language system in both cases. Despite these commonalities, these cases also demonstrate the heterogeneity in cortical responses to auditory language stimuli in autism.
The abnormal functional localization demonstrated in these two cases of autism may be significant in two ways. First, the functional patterns described above are not unlike patterns seen in developmental dyslexia, in two ways. Neuroimaging studies have suggested that young adults with compensated developmental dyslexia overuse right temporoparietal and inferior frontal gyrus areas during reading tasks in order to compensate for weak left hemisphere language areas.8 Likewise, individuals with autism could also be overusing right hemisphere areas. This would be consistent with the pattern of activation demonstrated in Participant A and previous neurophysiological studies.6, 7
Children with developmental dyslexia demonstrate a different abnormal pattern of cortical activation as compared to young adults with compensated developmental dyslexia. Magnetoencephalography studies have shown that children with developmental dyslexia demonstrate a reverse sequence of left hemisphere activation with the left inferior frontal gyrus activating prior to left temporoparietal.9 This is similar to Participant B who demonstrated a reverse in the typical sequence of superior temporal gyrus and inferior frontal gyrus activation. Thus, these two developmental disorders, developmental dyslexia and autism, may share common patterns of cortical reorganization. Studying this overlap in reorganization may help us better understand neurodevelopmental disorders and how to apply research knowledge and treatments from other neurodevelopmental disorders to autism.
Second, abnormal cortical localization of language has implications for the developmental of multisensory integration and speech prosody – areas in which individuals with autism have difficulties. The posterior superior temporal gyrus, a brain area involved in verbal language processing, appears to be critical for multisensory integration.15 Thus, posterior superior temporal gyrus abnormalities may affect the development of language and multisensory integration. Second, speech prosody is typically processed by the right superior temporal gyrus and inferior frontal gyrus.16 If these areas are compensating for dysfunctional left hemisphere language areas, they may be unable to process speech prosody simultaneously, leading to speech that is poorly modulated in volume, inflection and emotional content.
The auditory evoked fields in the cases above demonstrate indications of cortical hyperexcitability. For example, Participant A demonstrated multiple late components and Participant B demonstrates high amplitude early waveforms. Several lines of evidence support the idea of cortical hyperexcitability in autism, including an abnormally high number of excitatory pyramidal cells and underconnectivity in long-distance frontoposterior reciprocal pathways.17 Rett syndrome is characterized by autistic behavior and cortical hyperexcitability.10 Animal models of Rett syndrome demonstrate decreased cortical inhibitory18 and reduced long-term potentiation and depression.19 The auditory evoked fields in this report support the notion that cortical hyperexcitability could be a factor that links the deficits in cognitive development in the various forms of autism.
This report confirms and clarifies previous neurophysiological findings in autism and outlines parallels between autism and other developmental disorders. Uncovering these parallels may help us understand the neurological basis of autism. Such information will no doubt be of great benefit in furthering diagnosis and treatment of this devastating disease.
This project was supported by NS046565 to Dr. Richard Frye. The authors would like to thank Mr. David Strickland and Benjamin Malmberg for their technical assistance.
Statement of conflict of interest: Dr. Frye is the director of the medically-based Autism clinic at University of Texas. Dr. Frye provides expert opinion for child neurology cases for both the plaintiff and defendant, including those in which autism is suspected and vaccine injury cases. All proceeds from such work are provided to the Department of Pediatrics at University of Texas with some of these funds supporting Dr. Frye’s research.