Humans have five senses. Hearing, vision, touch, smell, and taste. Although they're all processed in different parts of the brain initially (that would be the auditory cortex, the visual cortex, the somatosensory cortex, the olfactory cortex, and the gustatory cortex respectively), our senses interact to give us a unified perceptual experience.
As with some of the visual illusions mentioned in an earlier blog, these illusions may be the occasional price we pay for having a perceptual system that, more often than not, is able to successfully integrate the incoming sensory stimulation.
Which brings us to autism:
It's long been argued that people with autism have sensory integration difficulties. However, the evidence relating to multi-sensory integration is actually pretty mixed. A number of studies have reported that the McGurk effect is reduced in autism. However, others have reported that people with autism show a normal sound-induced flash illusion. As far as I'm aware, nobody has looked at the rubber hand illusion and we know very little about how the sense of touch is integrated with other senses.
In an article published this week in Autism Research, Natalie Russo and colleagues at Albert Einstein College of Medicine reported a study of auditory-somatosensory (touch) integration in autism. Seventeen children with autism were asked to watch a film with the sound turned down. During the film, short tones were presented through speakers and a pad held in one hand vibrated. There were three kinds of stimuli: (1) Tones on their own; (2) Vibrations on their own; and (3) Simultaneous tones and vibrations.
Throughout all of this, the participant's electrical brain responses were measured using EEG (electroencephalography) - a cap with electrodes placed strategically all over the head.
To analyse the EEG data, the researchers added together the brain response to the tone on its own and the brain response to the vibration on its own. They then compared this to the brain responses to the bimodal stimulus (i.e., the simultaneous tone and vibration).
If the two modalities are processed separately by the brain and are not integrated then the response to the bimodal stimulus should be precisely the sum of its parts. Indeed, this is pretty much what they found for the autism group.
The researchers also tested a control group of typically developing children, who were the same ages as the autistic children. For these children, the brain response to the bimodal stimulus was more than the sum of its parts. This effect was particularly strong at around 150 milliseconds after the start of the stimulation at electrodes overlying the parietal lobes.
What does it all mean?
The results of the study are pretty intriguing. It seems as though the control children's brains integrated the tone and the vibration when they occurred together, but this cross-modal sensory integration was pretty much absent in the autism group.
Russo et al suggest that their results may reflect a weakening of the connections between the different sensory regions of the brain. And they argue that this would be consistent with the more general theory that the brains of people with autism are 'under-connected'. In other words, the different parts of the brain are off doing their own thing, rather than talking to each other.
What's nice about the study is that the subjects didn't actually have to do anything. They just sat and watched the movie while their fingers were vibrated and their ears were assailed by beeps. It's difficult, therefore, to explain away the results as just a consequence of the autistic children not paying attention or not understanding what they had to do.
However, the authors note an important limitation of their study. The children covered a wide age range from 6 to 16 years and its known that brain responses to these kinds of stimuli tend to change (get faster) with age. This meant there was quite a lot of variation even in the control group and this may have masked interesting differences.
It's also worth pointing out that some other recent EEG studies have reported normal audio-visual integration in autism. It's not entirely clear why the results conflict, so it would be interesting in future to look at integration of different combinations of senses within the same group of participants.
While the current results indicate that cross-modal integration is reduced in autism or (and I always have to add this caveat) at least in some individuals with autism, they don't tell us where in the brain this occurs or what the underlying neural process is. The cross-modal integration effect measured by the electrodes at the scalp could potentially originate in the auditory cortex, the somatosensory cortex, or some other brain area that takes inputs from these two primary sensory regions. Unfortunately, the spatial resolution of EEG is very poor and it may require a different technique such as MEG (magnetoencephalography) to address this question.
The 'how' question is potentially even more intriguing. The EEG (or MEG) response measured on the surface of the head actually reflects the combined activity of millions of neurons. The cross-modal integration effect in typically developing children could be caused by a change in brain activity (more neurons firing somewhere in the brain). But it could also reflect an increase in neural synchronization (the same neurons firing but all doing so at the same time). Fortunately, there are some rapidly developing analytical techniques that should allow researchers to differentiate between these possibilities and provide further insights into the neurobiological mechanisms of autism.