Home > News & Opinion > News > 2012 > 'Noisy' brain signals could underlie autism, study says

'Noisy' brain signals could underlie autism, study says

Popularity tracker
Virginia Hughes
24 September 2012

Group theory: Averaged across many individuals, brain activity in adults with autism (orange) does not look different than in controls (blue), but varies widely in a given individual with the disorder.

The glow of a lamp, the ring of a musical note, the tickle of a feather — each sensation stimulates a different part of the cortex, the outer layer of the brain. In an adult with autism, these responses vary much more than in someone without the disorder, according to a study published 20 September in Neuron1.

Using functional magnetic resonance imaging, the researchers exposed participants to the same stimulus — a sight, sound or touch — repeatedly over dozens of trials.

When they averaged brain activity across all the trials, they found no difference between the autism and control groups. But when they looked at a single participant's response from one trial to the next, they found that, for those with autism, the same person may show a strong response to a certain stimulus in one trial and a weak response in the next.

Sensory signature:

The researchers say the ‘noise’ in autism brains could be a neural signature of the disorder. “It could ultimately explain the whole slew of altered behaviors that we see in autism,” Behrmann says.

The researchers also did not explore whether children with other developmental disorders or intellectual disability show the same type of variability.

"If you’re going to make this argument, then it’s really important to show that [the variability] is specific to autism,” says Kevin Pelphrey, associate professor of child psychiatry at the Yale Child Study Center, who was not involved in the new study. “You have to explain, developmentally, how would you get the autism phenotype by starting out with a noisy brain?"

Behrmann and her collaborators scanned the brains of 14 high-functioning adults with autism and 14 controls while they performed a difficult task. The participants watched letters flash on a screen and pressed a button when they saw one repeat.

During the task, the researchers exposed participants to various sensory stimuli: beeps in both ears, a cluster of moving dots on one side of the screen, or puffs of air on the back of their hands. Each of these stimuli activates a different part of the cortex — the auditory cortex, the visual cortex and the somatosensory cortex, respectively. 

In each area, the researchers found robust and similar brain responses in both groups of participants. But individuals in the autism group showed a lower signal-to-noise ratio, the strength of their brain signals divided by the background noise from each trial.

“We’re way outside of the social brain areas, in regions that do the most basic cortical computations that one can imagine,” Behrmann says. 

Heads up:

The study, first reported at the 2011 Society for Neuroscience annual meeting in Washington, D.C., is one of only a handful in the autism literature to investigate trial-to-trial variability in a single participant.

For example, a 2008 behavioral study found that, compared with controls, children with autism have more variable reaction times from one trial to the next2.

Each of these studies uses a different technique. “There’s some real scientific convergence, suggesting that it’s a reliable finding,” says Milne, lecturer in cognitive psychology at the University of Sheffield in the U.K. “It’s exciting because it provides evidence that there is a kind of general and pervasive disruption in neural processing.”

However, most of these studies have not examined whether the noisy patterns are specific to autism.

The new study may also suffer from head motion artifacts, which are common in brain scans.

The researchers performed some checks to account for this possibility. They found no difference in head motion between the autism and control groups. They also adjusted for individual movements, and note that they saw more variable responses in the cortex than in deeper regions. Head-motion artifacts typically appear all over the brain.

“If it was purely a motion effect, you’d expect it to permeate every response they looked at,” says Steve Petersen, professor of cognitive neuroscience at Washington University in St. Louis, who was not involved in the new study. “At least part of their result is probably a real result.”

Others, however, say this difference in the cortex is a red flag. “It reminds me of exactly what you’d see if you had a difference in movement,” Pelphrey says. That’s because when lying in a scanner, the outer regions of the brain — farther from the stability of the neck and spine — are most susceptible to body movements, he says.


1: Dinstein I. et al. Neuron 75, 981-991 (2012) Abstract

2: Geurts H.M. et al. Neuropsychologia 46, 3030-3041 (2008) PubMed

3: Milne E. Front. Psychol. 2, 51 (2011) PubMed


Name: Marlene Behrmann
25 September 2012 - 9:28PM

We thank our colleagues for their comments on this paper and take this opportunity to respond to two specific aspects.

One of the commentators raised the issue of specificity and noted that it is really important to show that the intra-individual variability we observe is specific to autism. This is a valuable point and, in the manuscript, we consider this issue directly (see p8, section entitled "Specificity of Poor Response Reliability to Autism." We have begun to put in place a similar study to examine the consistency of neural responses in adult individuals with schizophrenia and we hope that the findings (and the comparison with the autism data) will shed light on the specificity of the variability.

A second point that was raised concerned the possibility that the variability we observed might simply be a result of motion artifact. One of the commentators suggested that the pattern we report is consistent with this motion explanation as the variability is evident in the outer regions of the brain “farther from the stability of the neck and spine”. Motion artifact is a large problem in data interpretation in imaging studies and it is for this reason that we spent much time ensuring that this artifact could not account for the pattern of data. On p4 of the paper, we wrote:
"We performed several control analyses to ensure that larger trial by-trial fMRI variability in the autism group was not caused by more variable head motion, heart rate, respiration, or eye fixation during the experiments. The variability of all six head motion estimates, derived during 3D motion correction, was statistically indistinguishable across groups as was the mean frame-by frame displacement (Figures S7A and S7B). Furthermore, we reanalyzed the fMRI responses after removing head motion parameters using orthogonal projection (Fox et al., 2006) and found that signal-to-noise ratios remained significantly smaller in autism (Figures S7C–S7E)."

In addition, we demonstrated that the increased variability is not attributable to head motion as we only see this in the evoked response in sensory areas and not in frontal cortex or any other of the 40 cortical ROIs during those same scans (see Fig 4). Head motion would be expected to generate variability throughout, not just in the areas driven by the stimulation. As noted by Steve Petersen in his comment
“If it was purely a motion effect, you’d expect it to permeate every response they looked at.” It did not permeate every response nor every cortical area.

Taken together, we believe we have been very careful in addressing an obvious potential motion confound and that we have ruled out a motion artifact as the explanation for the inconsistency in the neural responses in autism.

Marlene Behrmann
Ilan Dinstein
David Heeger

Add a Comment

You can add a comment by filling out the form below. Plain text formatting.

Question: What is 10 + 4 ?
Your answer: