The human calamity we call autism has been underplayed by the media. The usual portrayal of autistic behavior in films and books is useful to increase public awareness, but such portrayals are almost always of relatively mild cases. Here are some elements of an actual case diagnosed as severe autism in a Nigerian boy 13 years of age:

We'll call him John. He was abandoned in a refuse dump a few days after birth and taken into a home for destitute children when his parents could not be found. As a child he rarely played with other children. He was unable to develop speech--no verbal communication. He screamed when distressed or in need of attention. He avoided eye contact, often staring into space. He never turned around when his name was called. He was disruptive with other children and often snatched their food away after finishing his own meal.

John is now 13 years old. He never utters a word. He sometimes shouts and screams without apparent reason. He hardly ever responds to instructions. He appears distant when attempts are made to interact with him. But he responds to the word "take" if you look at him while you hold a biscuit or some other snack. He'll then snatch the biscuit and eat it at once. He'll sometimes run around the room in circles, apparently with glee, and he'll stop running only when he's physically forced.

This boy has no history of any major medical illness that might have influenced his neurological development. He has a hypomelanotic skin condition (oculocutaneous albinism), but no vision or hearing impairments and no motor abnormalities. His developmental deficits are apparently restricted to communication, cognition, and social interactions.

Such is a case of severe autism. No matter all the details that continue to emerge from clinics and laboratories, autism is an enigma--maybe the main psychiatric enigma currently preoccupying the American public.

Anecdotal descriptions of cases with probably some form of autism exist as early as the 18th century, but the term "autism" did not come into general use until the middle of the 20th century when it was introduced to describe aberrant infant behavior: lack of affective contact, a desire for sameness, fascination with objects, and mutism or non-communicative language before 30 months of age.

Today, autism terminology includes "pervasive development disorder" (PDD), or "autism spectrum disorder" (ASD), which in turn comprises five diagnostic categories: autism, Rett syndrome, Asperger syndrome, childhood disintegrative disorder, and "pervasive developmental disorder not otherwise specified" (PDD-NOS). PDD is also called "broad" autism, while the more severe autism per se is called "narrow" autism, the label for John's case.

Rett syndrome is a special case that illustrates the confusion in diagnostics. It's an X-chromosome neurological disorder seen only in girls. It appears suddenly at about 6 months of age, leads to decreased brain and body growth with autistic behavior, severe dementia, seizures, and early death. The identical-twin concordance is 100 percent. Rett syndrome is considered part of the "pervasive developmental disorder" group, but it's probable that sooner or later it will be moved out of this diagnostic group when more is known about its etiology.

Here, let's use ASD or autism to mean broad autism and use the phrase "severe autism" to refer to the narrow form.

At the present time, ASD is defined by significant impairments in social interaction and communication and the presence of apparently unusual behaviors and interests. Many people with ASD also have unusual ways of learning, of giving attention, or of reacting to various sensations. Their thinking and learning abilities may vary from mental retardation to extreme intelligence. ASD usually begins before the age of three and remains an essentially permanent condition. The condition occurs in all racial, ethnic, and socioeconomic groups, and is four times more likely to occur in boys than in girls.

As late as the 1980s, autism was thought to be a rare disorder with a prevalence of about .08 percent (8 per 10,000 persons). In 1996, the prevalence of autism in metropolitan Atlanta was reported by the CDC as 0.34 percent. As reported by the CDC, the autism prevalence rate in the United States in 2002, in a survey of over 400,000 children in 14 areas, was 0.66 percent. The prevalence today is more than 10 times greater than the 1980s figure, and there isn't much agreement about why this is so. An epidemic? More accurate diagnosis? Greater media attention to the disorder? To date, the reasons are unclear.

What is clear is that, like many psychiatric diagnoses, ASD and severe autism are not disease entities. Rather, they are man-made labels used to denote or categorize a particular constellation of symptoms. Not all of these symptoms occur together; many of them can involve various causes; and some or all of the symptoms can be caused by qualitatively different effects on the brain. As a result, the labeling of autism forms is something of a moving target, subject to change nearly every decade in the absence of a cohesive understanding of their complex etiology.

Indeed, given the complexity of autistic behavior, it's reasonable to assume that more than a single cause might be involved in different individuals: genetic factors, prenatal impacts, postnatal toxins, or some combination of these causes. The question before us here is to what degree might the clinical entities called autism spectrum disorder (broad autism) and its severe form (narrow autism) involve prenatal impact on the developing embryo or fetus?

These days the fashion pendulum has swung from introspective psychodynamics to the mystique of DNA. Gene-mongering in the media is as rampant about autism as about any other aspect of human behavior, ordinary or aberrant. In 2007, in a piece about autism, ABC News quoted a pediatric neurologist:

"Identical twins most of the time will both have autism. The rate of a fraternal twin having autism is zero to 10 percent, whereas for identical twins it is 80 to 90 percent. So that means that genes have a lot to do with it."

Not quite. Autism in identical twins can be caused by a shared fetal environmental impact rather than by direct genetic transmission. The often-quoted concordance rate of 90 percent among identical twins is accurate for broad autism but not for severe (narrow) autism, for which identical-twin concordance is only about 60 percent. Also in contrast to the ABC News report, the concordance among fraternal twins can be as high as 23 percent for broad autism.

Studies of monozygotic (MZ, identical) twins can be important if we want a suggestion about the possible role of heredity in a condition or behavior. But the details are important. Monozygotic twins may or may not share a chorion, one of two membranes (the other is the amnion) that surround an embryo during early gestation, and the developmental consequences in each case are usually not predictable. The local prenatal environment for monozygotic twins can be the same or different. The concordance of autism in monozygotic twins varies with the study and ranges from 36 percent to 90 percent. Any significant concordance may imply genetic factors acting alone or genetic factors affecting susceptibility to a shared fetal environmental impact. It's also possible that given a fetal environmental impact, genetic factors determine the neurological consequence--in one case autism, in another instance ADHD, and so on.

At least 60 different genetic, metabolic, and neurologic disorders have been associated with autism, including fragile X syndrome, Down syndrome, fetal valproate embryopathy, and congenital German measles (rubella). At the present time no single gene or cluster of genes has been identified as a genetic cause of a majority of cases of autism. Nor is familial autism sufficient evidence of a genetic component, since if the diagnosed familial autism is a consequence of environmental impact, familial incidence is expected. In addition, even apparent familial autism is complicated by the possibility of multiple etiologies. Given the frequency of autism in the general population, it's possible for siblings to have the syndrome for different reasons.

Bottom line: At the present time in the United States, the consensus is that autism is a consequence of a development diversion that occurs before birth. In about 1 percent of cases there's evidence that the diversion is already programmed by inherited genes, but there's no evidence for complete genetic determinism of the disorder in 99 percent of cases.