Everyone knows what noise is, but only because normally developed brains perceive it as such, i.e., senseless, sometimes loud, un-patterned sounds. ("Noise" can also mean sensory confusion absent sound; deaf children can be autistic, too.) But what would happen if the sounds of a language were perceived as the equivalent of static or if the words on a page were just another distraction in an external word that was all irritating distraction? What if you couldn't distinguish between a barking dog and an oncoming train, a ringing telephone or your mother's voice?

These are the types of questions Merzenich and Rubenstein repeatedly asked themselves as they assembled what they believe is autism's most likely developmental scenario. Merzenich does not reflexively genuflect toward his own creation. But with each new experiment in his own lab or in collaboration with scientists elsewhere, he becomes more convinced. "The fact that 30 percent of autistic children have seizures and 50 to 70 percent have abnormal brain activity while sleeping suggest that these children have noisy and unstable cortical networks."

But what does he mean by "noisy"? And in the most ideal of circumstances, how does the brain learn to hear?

"We know the brain is plastic. It changes with experience," Merzenich explains. The changes are not grossly structural; you don't suddenly get a new lobe when you take up piano, for instance. But the changes are subtly structural in the sense that signals, moving from one brain circuit to another, do create pathways that both underlie and support ongoing cognitive abilities. Learning and memory are prime examples.

Think of it this way. To work normally, the brain has to be wired a certain way. Like everything else in the human body, the blueprint for this self-organizing machine is ordered by the genes while the actual construction is carried out by proteins, which guide cells from their I-can-be-anything-I-want-to-be state to their adult stage, where they differentiate into special types. Since a new genetic template is formed at fertilization, thanks to the mixing of genes carried by the chromosomes of the father and the mother, there is always the possibility that the ideal genetic instructions will be altered. Such is nature. A change for the worse here or there might not matter much. But as Rubenstein -- who studies the genetic expression that directs the development of different regions of the brain -- explains, a collection of mutations that, say, rev up the release of glutamate, a neurotransmitter responsible for exciting synapses, could mean trouble. Similar problems that affect the amount or capacity of other neurotransmitters, such as acetylcholine, dopamine and serotonin would amplify the impact.

For Merzenich, the alarm most frequently sounds first in the brain's nascent auditory system. "What you don't want is a brain that is hypersensitive to sound, which overloads the circuits." In essence, the brain needs to be able to distinguish, sort, classify and selectively focus on information pouring through its auditory filtering systems. Merzenich labels this flash point in neural development, during which so much is up for grabs, the "critical period." It is the point at which sound patterns are discerned and the "testing, testing" stage ends. With the patterns set, the system of circuits that enables single neurons to react to specific sounds is then made more reliable and permanent by the release of a chemical known as brain-derived neurotrophic factor (BDNF), the genes' final sign-off on its own creation.

But because what the brain hears helps to shape it, if it can't hear signals properly -- because of the flaws in the genetic blueprint -- faulty circuits form. Disorder reigns, but in degrees determined by the depth of the genetic flaws and, it seems, by susceptibility to environmental influences. More about those later.

For example, in experiments with rats whose developing forebrains were bombarded with sequenced bursts of noise at different frequencies (simulating the consequences of the inherited genetic weakness in humans), Merzenich and his colleagues found that the critical period had ended too early. In other words, the brains had not developed with a full repertoire of plastic circuits -- the very circuits that underlie the perception, memory and cognitive abilities that support a child's language development. In related experiments, during which developing rats were exposed to continuous, nonpulsating sound, akin to white noise, a similar problem occurred, if for a different reason. "We found that critical period stayed open too long. Differentiation was delayed indefinitely, leaving the brain more susceptible to cortical epilepsy."

Merzenich is unsure which of the two possible scenarios might help to better explain autism. In his mind, both could be responsible. And both are catastrophic because the auditory information pouring through the brain is undifferentiated and nonselective. "An autistic child's brain takes in everything simultaneously and records everything together. The cortical maps are all wrong. Instead of order, there is chaos, overexcitation and hypersensitivity."

Toxin Tocsin

If Merzenich and Rubenstein are right, a child with a mild or moderate genetic tendency toward autism might develop normally if nothing else happens to transform the genetic wobble into freefall. Sadly, a growing list of candidates, led by such old poisons as PCBs (polychlorinated biphenyls) and their newer cousins, PBDEs (polybrominated diphenyl ethers), have been shown in animal experiments to scramble brain wiring in much the way the two UCSF researchers have theorized.

In fact, in experiments led by University of California, Davis researcher Isaac Pessah -- a member of that university's M.I.N.D. Institute -- it was found that exposing perinatal rats to relatively low levels of PCBs resulted in brains whose auditory centers were completely disorganized. "Keep in mind," says Pessah, "that rats are not auditory creatures like humans. These kinds of changes in the brains of kids would be a huge problem."

PCBs, of course, have been banned since the mid-1970s, so any potential threat should be decreasing, correct? Wrong. It turns out, says Pessah, that naturally occurring microorganisms have stripped some of the PCBs of their chlorine molecules, meaning that current field tests do not adequately measure the existence of these "non-coplanar" PCBs in our air, water and food. Pessah estimates that several hundred of these variants still persist, representing a silent and growing threat to human health. Understanding how and where the PCBs do their damage are questions Pessah hopes to answer in ongoing research. "Right now, it seems that the PCBs affect a huge protein complex that serves as an integration point for cell signaling."

PBDEs, flame retardants found in everything from children's pajamas to computer casings, seem ready to join the list of major health hazards as well, thanks to a series of breast-milk studies in North American and European women that have revealed dangerously high concentrations. Indications are that these concentrations are doubling every four to five years. In response, California has banned PBDEs, starting in 2008. The European Union has already done so. Merzenich is not ready to state unequivocally that PCBs and PBDEs are behind the rise in autism worldwide, but their insidious effects on the developing nervous system have raised alarms among toxicologists and public health advocates everywhere. Moreover, they fit the profile of what might be fueling the rise, since they are now lodged in our ecosystem and in our fat cells, while remaining as ubiquitous as air.

Other possible factors are less easy to quantify and, because they stray into the sociology of parental practices, place Merzenich on that fragile border between messiah and pariah. His target: ambient noise in modern life, particularly in childrearing environments. "There is so much more noise exposure for infants than there used to be. Even little things can add to the risk and push a baby across the threshold," he warns. The risk, he believes, comes not for the profoundly autistic, who are born with severe and recognizable problems, but in the genetically weakened, whose developing brains teeter delicately between normal ability and impairment. For this group, he believes, the constant insult of a blaring television, blasting radio or even a loud air conditioner or fan can cause the brain to organize around a poor signal. In essence, noise becomes its language, leaving the child with brain wiring that is less attentive to the differing information that sound carries.

The danger in saying this too loudly is that parents will overreact and tuck their infants into soundless rooms, which Merzenich asserts would be a disastrous misreading of what is really a cautionary prescription. "Parents need to talk and read to their babies all the time. That is absolutely essential. Singing or playing lullabies and other structured music is a good idea too. In general, it is the richness and variety of clear sounds that matter most." Indeed, Merzenich's success in using computer programs to retrain the brains of children with less serious learning disorders gives him hope that in time, the more responsive cadre of autistic children might reclaim some of what has been lost.

Merzenich has been adept at turning theory into practice before, so no one is betting against him this time, least of all Bryna Siegel, director of UCSF's Autism Clinic. Yet impressed as she is with the elegance of the Merzenich-Rubenstein theory, she considers it too narrow to explain the variety of problems -- particularly the social deficits -- she finds in autistic children. This despite Merzenich's belief that the developmental train wreck they describe is likely affecting many dimensions of brain function, to say nothing of other brain regions that are struggling to organize normally. "Over-excitation of the brain and language deficits are only part of what I see," says Siegel. "And I don't know of any clinical studies that show that children raised in noisy environments are more prone to autism. Autism is an astoundingly complicated syndrome." Moreover, even if theory led to a cure tomorrow, hundreds of thousands of autism patients would remain. And it is due partly to Siegel's efforts at finding these children through improved screening methods that the numbers continue to rise.

Page 1 - Grasping Autism

Page 3 - Screening Is Believing