Radiology: fMRI shows promise for autism diagnosis
 | |
| MRI brain scan highlights the amygdala, one of the areas of the brain that is reduced in volume in subjects with Conduct Disorder. Credit: G.Fairchild et al, 2011. |
“Children with autism were distinctly different [from normal children] in their responses to spoken narrative. This study is very preliminary. However, it is the first proof of principle that you can put a child in a scanner, sedate him or her and use fMRI to get an indication of autism,” Joy Hirsch, PhD, of the functional MRI research center at Columbia University Medical Center Neurological Institute in New York City, told Health Imaging News.
According to the Centers for Disease Control and Prevention, it is estimated that as many as one in every 110 children is affected by autism. “However, the diagnosis of autism currently remains limited to parent and clinician observation of missed developmental milestones,” Hirsch shared.
Hirsch and colleagues designed the study to assess the feasibility of applying fMRI as an indicator of language disability in autism.
The researchers analyzed data from fMRI scans acquired between 2008 and 2010 for 15 non-autistic children (mean age, 12.3 years) and 12 autistic children (mean age, 12.4 years) who were imaged while alert. They also examined fMRI datasets of another 27 autistic children (mean age, 8.62 years) who were sedated during scanning.
Hirsch and colleagues presented prerecorded parents’ voices as stimuli, a technique that was pioneered more than 10 years ago for neurosurgical planning.
The researchers concentrated on BOLD activation within the primary auditory cortex (A1) and superior temporal gyrus (STG). They established and quantified differentiation between autistic and control subjects by calculating the standard deviations away from the mean of the control group for two measures: signal spread and amplitude. Hirsch and colleagues used anatomically defined regions of interest applied to individual activation patterns.
Activity in the A1 region of the brain did not differ between autistic and control patients. However, activation within the STG was greater for control children relative to autistic patients.
“These findings first tell us that the autistic children in our study appeared normal with respect to the primary auditory system,” Hirsch said. “But it appears that the STG in the autistic brains was not as sensitive to the language narratives as was the STG in the brains of the typical children.”
When the researchers completed a similar analysis of the sedation-adjusted values from the control group, they identified 96 percent of sedated autistic patients.
Hirsch noted that one method—either signal spread or amplitude—would suffice in clinical practice, adding, “Future studies are required to determine optimal diagnostic criteria.” She also referred to the possible utility of propofol sedation, which enables functional imaging in children as young as one or two years.
“If fMRI is to be applied as a diagnostic tool for use in children with developmental delay, sedation would likely be necessary in most cases.”
The researchers acknowledged several limitations to the study, such as “the question of age and how these findings would apply to younger children, who are at any age at which an objective medical diagnostic procedure would be most useful for the purpose of early intervention.”
Hirsch suggested paths for future research, including determining how fMRI might help differentiate between autism and other developmental disorders and applying the concept to children as young as one or two years of age. Another need, she continued, is a larger sample size.
Regardless of which avenues researchers explore in the future studies, Hirsch and colleagues have succeeded in a critical first step—transferring foundational work employed in neurosurgical planning to demonstrate the proof of concept for employing fMRI in the diagnosis of children with autism.