NLMFF interviewed Timothy Roberts, Ph.D.,
Oberkircher Family Chair in Pediatric Radiology
and Vice Chair of Research for the Department
of Radiology at Children's Hospital of Philadelphia
and Professor of Radiology at the University
of Pennsylvania School of Medicine, about what
MEG imaging studies can reveal about autism
spectrum disorders and his recent grant co-funded
by the NLM Family Foundation and Lurie Family
Foundation entitled, “MEG of Language
Impairment in Autism.”
NLMFF: Can you tell us about your background and about how you became interested in studying autism spectrum disorders in the first place?
Roberts: I trained as a medical physicist at Cambridge University (BA – 1988; PhD 1992) and immediately undertook a post-doctoral fellowship at UCSF in Neuroradiology. At this point, MEG as a clinical technology was in its true infancy with only a handful of centers using the technology, but I was immediately struck by its capabilities in mapping brain function in space and time. I was particularly fortunate to be part of the team setting up the first MEG system at UCSF and developing its applications (many thanks to Prof Howard Rowley for this collaboration). Of particular note, I was struck by the fact that a large number of patients, especially those with autism spectrum disorders, were not adequately served by conventional (albeit powerful) radiology techniques such as CT and MRI. It became a mission to apply MEG as a technology for studying the temporal dynamics of neuronal activity on the brain to patients with ASD.
NLMFF: You recently received a grant co-funded by the NLM Family Foundation and the Lurie Family Foundation for a project entitled, “MEG of Language Impairment in Autism.” What do you hope to accomplish through this grant?
Roberts: The goals of this grant are to elucidate the neural mechanisms of language impairment in autism. We believe that at some stage in the progression from sensory stimulus detection, through feature extraction and subsequent processing to higher cognitive computation, deviations in temporal patterns of brain activity will be identified in the children with autism and that these will serve as “temporal signatures”, providing a specific means of classifying the abnormalities of brain function. It seems promising that such “signatures” or brain-level phenotypes will have great utility in unraveling the clinical heterogeneity observed in the ASD populations we serve, as well as providing concrete measures for experimental model validation and genetic/genomic correlation.
NLMFF: What are some of the most important precedents in the field of neuroimaging for the work that you propose to do? For instance, what have we learned about dyslexia and how the brain processes language?
Roberts: There are clearly some commonalities in disorders affecting language function. In dyslexia, for example, we and others have reported deficient rapid temporal processing abilities (based on the pioneering work of Tallal) and neural correlates thereof, measured by MEG. Others (e.g. Papanicolaou) have shown abnormal hemispheric patterns of activity (crossed dominance between expressive and receptive language areas). Our goal is to identify both (i) systems of neuronal dysfunction that are common across clinical diagnoses, but are associated with behavioral performance impairment, and (ii) unique signatures of neuronal dysfunction that are specific to language impairment in ASD as distinct from other developmental and psychiatric disorders.
NLMFF: What are the unique advantages of MEG compared with other technologies available to image the brain?
Roberts: Well, as the realtors say “location, location, location”, we have to argue for “timing, timing, timing”. While spatial organization of brain function is important, it is also the case that brain activity (especially of complex functions) involves a network of rapidly communicating brain centers. MEG is uniquely able to provide not only spatial but also temporal (sub-millisecond) characterization of brain function as well as focusing directly on neuronal electrical activity, rather than indirect hemodynamic correlates of such activity, which may be significantly delayed and displaced.
NLMFF: We know that autism is highly heterogeneous. How might your research help identify subtypes of autism? What might the implications of finding subtypes be for diagnosis and treatment?
Roberts: By identifying brain-level phenotypes, or “signatures”, it seems a more specific characterization of children and adults with ASD will be achieved, allowing for a quantitative account of the clinically and behaviorally-observed heterogeneity. The degree to which such signatures overlap with other (co-morbid) conditions might point the direction for therapeutic interventions in various subsets of patients – that is identifying some patients with ASD as likely to respond to a treatment that is available for another clinical diagnosis because of similarities in their brain-level dysfunction.
NLMFF: Can MEG imaging be used to identify specific “interferences” in the mental process of transforming intention to speak into production of speech? Might these specific points represent targets for potential therapeutics?
Roberts: This is a really worthwhile goal. As of now, we are addressing the ability of MEG to detect receptive language activity that may be present but not be manifest behaviorally. We are terming this “passive probes of language competence”. It may be beyond the scope of the present proposal to extend this towards expressive language, although we will certainly be investigating activity in inferior frontal lobe structures that are conventionally considered to subserve speech production.
NLMFF: Do you think the term “language impairment” captures the essence of the communication problems faced by people with autism? Might the inability to speak, for example, be the result of a neuromuscular or fine motor planning impairment or, at the other extreme, reflect a timing or sequencing issue instead? In your view, might MEG imaging studies reveal for at least some cases of autism an intact Chomskian “language facility”?
Roberts: Again, the use development of “passive probes of language competence” really gets at this issue, allowing us to resolve “competence” from “performance”. It should be borne in mind that spoken language is only one aspect of communication and that there are various manifestations of communication impairment across the spectrum. We hope that our findings spur other research groups into pursuing similar approaches addressing the broader picture of communication impairment.
NLMFF: What do you think the future holds for autism research? What do you consider the single largest obstacle to making progress in the field?
Roberts: It is an exciting time for autism research. Increasing awareness of the disorder combined with tremendous advances in functional imaging methodologies, such as MEG, offer encouraging perspectives. Rigorous scientific method and multi-disciplinary studies (combining e.g. imaging, genetics, neurobiology and clinical and behavioral assessments) are critical. We are tremendously grateful to foundations such as the Nancy Lurie Marks Family Foundation for their support of research endeavors such as ours. By encouraging new researchers to enter this field, we are recruiting fresh bright ideas from a range of scientific and sociological backgrounds. This has to be our recipe for success.