Jeff and Cathy Romanczuk,
lukate@charter.netChapter 2
Although the main research material used for the study is secondary source bibliographic abstracts, the bulk of the literature review conducted for this chapter more broadly investigates primary source material about autism causes. A working understanding of the various causal hypotheses being researched and the rationale behind them is a prerequisite to following the bibliometrics involved. The researcher’s assumption is that readers of this thesis are likely to be more familiar with Information Sciences than they would be with autism etiology research. Therefore, this chapter concentrates on autism literature while bibliometrics sources will be introduced as they explain and support the methods used, results, and discussion in the following chapters. This background on the etiology of autism highlights four main causal categories: psychological, neurological, genetic, and environmental, as well as a mix of these causes.
More will be written about this in Chapter 5’s “Discussion and Implications,” but a technical note regarding the use of “autis?” is useful here. Despite the many names for autism over the years—“early infantile autism,” “childhood autism,” “late onset autism”—all but “pervasive developmental disorder” use “autism” somewhere in their label. Even “Kanner Syndrome” and “Asperger Syndrome” are generally linked to the keyword “autism.”
Investigations of bibliographic references at one point in time are synchronous studies. These contrast with diachronous studies, which follow a pattern of use for specific items over a certain time span. The latter type of study would be more conducive to this work, since autism causation literature over the last fifty years is driving this research.
Not Autism Literature, But Sources About Autism Cause Literature
This thesis focuses not on the autism literature itself, but on the literature about this body of work. That is, online and print medical, psychological, and scientific bibliographies are the primary information base. Most of the sources characterized are from abstracts provided by vendors of biomedical databases. However, a preliminary search was conducted to find if studies similar to the work at hand have already been done.
Although the initial literature search revealed no bibliometric studies of autism, bibliometric research in general (along with nonbibliometric autism research) revealed good ways of addressing the issues involved. A search of Dissertation Abstracts Online using “autis? and bibliom?” yielded no matches, with 668 autism entries and 105 for bibliometrics. Adding “caus?” to the “autis?” set resulted in 26 hits. Since the search of “autis?” and “bibliom?” found no matches, a broadened query was used next, AND-ing “bibliom?” to “(neuro? or genet? or enviro? or immun?).” This search yielded eight hits, which offered no specific pointers to similar research, but did generate year-span and publication measures that could be incorporated into an autism scientometric, that is, the “quantitative analysis of scientific productivity” in autism research.
A search of Library and Information Science Abstracts (LISA) Online under “autis?” revealed a few useful finds, especially an Autism Research Database maintained by Britain’s National Autism Society. The researcher has been granted permission to use this database.
Using “bibliometric? and (study or research)” plus “genetic?” in LISA resulted in 12 hits. One citation of note represents a doctoral thesis comparing “[two] ways of measuring obsolescence within the subject of human and medical genetics.” This work compares synchronous studies with diachronous studies. The bibliometric research undertaken here is diachronous, so tracking this type of obsolescence (or the lack of it, that is, the overlap and longevity of etiologies) could be a worthwhile measure.
Although nearly 1,259 abstracts are in the sample, exact or even close matches to the research at hand were limited. Titles such as “Etiology of Autism: A Review of Recent Biogenic Theories,” “Autism and Genetics: A Decade of Research,” “Autism and Medical Disorders: A Review of the Literature,” and “Autism: Fifty Years on from Kanner,” were not as useful for this research as expected. What these articles detail is not a broad bibliometric review of autism’s etiology over the years but their case for one dominant cause of autism in the present. Unlike the LISA and Dissertation Abstracts Online searches already mentioned, these articles come from medical journals. Even so, to find only four that come close to this topic highlights the void this research is attempting to fill.
One problem with attempting such a study is also one of the strongest arguments for undertaking the work. The definition of autism has been refined over the years since 1943, when Kanner first labeled the condition. Even though 1943 may sound recent as a beginning of autism’s evolution, Hart points out that only the name is new; “this disability has been part of the human experience for ages.” One obvious problem this causes for an inclusive mapping of the literature is what autism was called and what were thought to be its causes before the term was applied.
The questions about what autism was called and what were thought to be its causes before the label existed can be answered in part by how autism was characterized in its early years, mainly as a psychological affliction. In “Psychoses of Early and Late Childhood,” the authors provide excellent background on what those exhibiting psychosis were called in various centuries. Until the sixteenth century, some version of “clouding of consciousness” was used. During the seventeenth century, “stupidity” labeled what became “idiocy” in the eighteenth century. “Dementia” began to be used in the nineteenth century, and “schizophrenia” in the early twentieth century. By the middle of this century, clinical descriptions influenced by organic trends and gradual detection of specific symptoms resulted in the “birth” of autism (and a host of other exact names for previously more generalized characterizations of mental illness).
While it is discouraging that little is confirmed about autism in the 1990s, looking at the early years makes clearer the later revelations and breakthroughs. Consider that the disease was labeled during the Freudian heyday of the mid 1940s and that psychologists spent the next twenty years treating autistic individuals (and their parents) with psychotherapy. The “psychogenic” cause stemmed from the belief that parents who were “distant,” or at best uneven in their parenting skills, caused autism. One early experimenter contrasted the “bright-eyed and alert look” of autistic individuals with the “dull” faces of those similarly mentally handicapped as proof of a psychogenic etiology.
The hopeful side of the psychogenic argument is that it implied reachability, that these individuals could be cured. However, this argument began to weaken in Great Britain and the United States when the psychoanalysis and drug treatments successful with mood disorders were largely ineffective in treating autism. As the data collected for this research will show, however, many countries still believe and invest much research in confirming a psychological basis for autism. For those countries where this avenue of research had been losing its following, the notion that autism was emotional in origin lost further backing after studies confirming that parents of autistic children exhibit the same range of skills as do any parents. Added to these findings were data showing children raised in “old-fashioned institutions” (where they were genuinely emotionally deprived) did not develop autism.
Hart goes into great detail to demarcate the differences between autism and mental illness. Although diseases of the mind and diseases of the body are not always as clearly distinct as once believed, mental illness does not usually occur in children. But the fundamental difference is that the brain’s neurotransmitters work erratically in the mentally ill. In autistic individuals the neurotransmitters work one (wrong) way, or simply do not work at all. Rapin adds to this distinction that evidence of a high correlation of autism with rubella and epilepsy indicate a biological, not a psychogenic etiology. However, she also points out that—even distinguished from purely psychological disorders—there is still a stigma attached to autism:
Autism is incurable. . .and, of all the developmental disorders of brain function, it is most often associated with behaviors that are extremely difficult to manage and socially unacceptable. The diagnosis of autism is the schizophrenia or cancer of developmental disorders. It is a diagnosis to whisper, avoid altogether, or couch in less-damaging terms.
In fact, early studies correlated infantile autism to adult schizophrenia, but these results were not replicated by follow-up studies.
Rapin makes the point that research into neurobiological causes of autism did not begin until the psychogenic theories about autism’s etiology ended (approximately the early 1970s for most countries). She indicates that the specific deficit areas now being researched include: sensory, cognitive, social, and behavioral. Even so, deficits for these diagnostic areas are still assessed by clinical observation, with no dominant theory as to autism’s underlying causes. However, the neurological researchers assert with near certainty that the relationship between the behavioral and cognitive characteristics of autism with those of abnormal brain structure and function implicate the central nervous system (CNS). Hart points out that new tests of blood and spinal fluid, as well as recent advances in ways of examining living brains (like magnetic resonance imaging [MRI] and electroencephalographs) are making inroads. However, more detailed findings are frustratingly inconsistent. Computer tomographs and MRIs, recent technologies that can “map” a living brain, sometimes indicate physical brain structure differences for autistic individuals and sometimes do not. The next obvious step was autistic brain chemistry studies using positron emission tomographs and magnetic resonance spectroscopy. While these results have been “more promising,” to date they are also inconclusive.
Even so, the CNS has been a focus of autism research from the beginning, and more so since technological improvements of the past thirty years have allowed for neurological studies on living brains. Bauman and Kemper preface their 1994 book by stating, “It has become apparent that autism is a developmental disorder of the central nervous system.” Rapin adds that even Kanner and Asperger in the earliest autism research “speculated that early infantile autism was the sign of abnormal brain function.” However, Gerlach credits Rimland’s 1964 book, Infantile Autism: The Syndrome and its Implications for a Neural Theory of Behavior, with debunking “the assumption that autism was a result of ‘bad’ parenting, and it helped to establish autism as a neurobiological disorder.”
The reason researchers can be so sure autism is neurobiological without knowing its specific causes is that any dysfunction in each brain system produces predictable behavioral deficits. Bauman’s research is guided by the hypothesis that those with similar symptoms may have in common a dysfunction in the same brain region, which may then allow for similar interventions. Bauman and Kemper close their book with a recapitulation of how the previous decade’s epidemiological, clinical, and biologic research has provided insight into “the underlying neurologic substrate” for autism and hypotheses about the characteristic features of autism.
Autism is characterized by many developmental deficit areas, most notably: sensory, attention, language, cognitive, and social. There are many (sometimes conflicting) theories about the brain area or brain function responsible for each dysfunction. Atypical responses to external stimuli indicate a vestibular dysfunction, often characterized by extremes in auditory responsiveness (either obliviousness to sounds or hypersensitive hearing). Vestibular abnormalities may also contribute to the “savant” traits sometimes noted about those with autism: visual/spatial abilities above the average for their peer group. The uneven attention spans symptomatic of autism (and many other disorders) are the result of a subcortex dysfunction. Language deficits, the most frequent complaint attendant to autism, have not been linked to any specific brain region’s abnormalities but may be a cerebellar malfunction or a byproduct of the communicative and social deficits associated with the disorder. Social deficiencies—the “most salient feature of autistic children”—appear to have a hypothalamic-limbic pathology, though some researchers suspect this atypical behavior is caused by a more general “right hemisphere” dysfunction.
Whether the physical composition of these brain regions is faulty or their transmissions are malfunctioning, the case for most etiologies of autism being neurobiologic is strong. This is especially true considering the frequency with which autism occurs with other CNS disorders.
Especially in children who seem to develop appropriately to a point—only to have their progress curtailed—abnormalities of the limbic system, cerebellum, and cerebellar circuits have been implicated. Bauman has also noted differences in the neuron size and growth rate in the cerebelli of autistic individuals. This would suggest abnormal relays between the cerebellum and other regions of the brain.
This brain research also suggests autism is congenital, though the symptoms may not be obvious during the first few years. Also supporting a prenatal etiology is the similarity of autistic traits to “those found in other conditions which have clear physical causes, such as congenital aphasia, or congenital deafness and blindness produced by German measles in the mother during pregnancy.”
That autism is a neurobiological disorder does not preclude other etiologies, especially genetic ones. In fact, in the 1940s Asperger noted behavioral similarities between his clients and some of their parents, and speculated about possible genetic factors. However, during the intervening years of belief in the psychological basis of autism, this avenue was not pursued. As late as 1976, researchers were concluding that “the cause was unlikely to be genetic.” It has been only in the last fifteen years that research into this pathology regained momentum.
In six studies compiled by Smalley in 1988, the prevalence of autism in siblings was found to be 2.7 percent. Although this percentage seems small, using 4 or 5 cases of autism in 10,000 births from the general population, this 2.7 percent is about 75 times greater than the norm. Piven and Folstein point out that the risk of autism in siblings could be even higher, since parents of autistic children often limit their plans to have subsequent children.
Not only the siblings, of course, but all relatives of affected individuals have “a higher mean liability” than does the general population. This was Falconer’s conclusion while researching a “multifactorial polygenic inheritance” model. Hart summarizes a multigenerational study jointly conducted by the University of Utah and the University of California, Los Angeles. This research found not only a statistically significant prevalence of autism in the extended families of autistic members, but also the autism rate for the children of an autistic parent (there were 11 autistic parents in the study) was 46 percent.
Other family conditions, such as cognitive deficits (usually in speech and language) and “personality characteristics” showed a higher prevalence in families with an autistic member. In a 1950s study, Kanner and Eisenberg noted that “a number of parents of these children were serious-minded, perfectionistic individuals with intense interests in abstract ideas, who appeared to lack interest in developing relationships.” Although parenting style itself is generally no longer believed to cause autism, “autistic traits” are still reported more often in the parents of autistic individuals than in parents in the control groups.
Twin studies in autism to date have produced inconclusive results which serve mainly to show how complex the disability is. While a few studies have found a higher concordance of autism in monozygotic twins than in dizygotic (82 percent versus 10 percent in one study), the small sample sizes used confound the results. Further, Steffenburg noted “perinatal stress” to be “much more common” in the autistic co-twin of discordant pairs and suggested that complications during birth “increase the risk of autism in individuals with a genetic liability for this disorder.”
Hart explains the theory behind the twin studies this way: since identical (monozygotic) twins are “genetic clones,” they have the same neurological structure. If autism were purely genetic, either both would have it or neither. Fraternal (dizygotic) twins have “genetic programs” as different as any siblings’ and share only the womb for nine months. So if fraternal twins both develop autism, this would suggest an environmental cause, such as perinatal exposure to some virus or chemical pollutant. He then relates the results of a twin study conducted by Ritvo which confounds this logic. Ritvo used 23 pairs of identical twins and 17 sets of fraternal twins. The fraternal sets showed few similarities, with several pairs containing one autistic individual and one twin with typical development. While 22 of the monozygotic pairs did develop similar patterns of autism, one pair had an autistic individual and a twin not diagnosed with anything. Hart interprets these results as suggesting that while autism has a strong genetic component, it may not be “purely genetic” but certain preconditions increase susceptibility. “Maybe autism cannot be inherited, but the susceptibility to autism might plague some families, just as cancer runs in families.”
Piven and Folstein point out a complication in autism studies not present in cancer studies, though. Autism is a “behavioral syndrome defined by clinical features.” They cite a Vandenberg report that further complicates the issue: the “search for a single underlying genetic mechanism in autism assumes homogeneity of genetic and environmental influences.” Even so, the heritability of autism, after factoring in the variance attributable to the average effects of genes, is estimated to be as high as 90 percent, with the other 10 percent of cases caused by environmental factors.
“Environmental” is something of a “catch all” category that includes autoimmune weaknesses (which make an individual more prone to many diseases), allergies, reactions to chemicals (including immunizations), and the presence of other diagnoses. Five disorders often seen individually in connection with autism are Rett's Syndrome, phenylketonuria, tuberous sclerosis, neurofibromatosis, and especially Fragile X (or Martin-Bell) Syndrome. Rett's Syndrome is listed among the Autistic Spectrum Disorders in the DSM-IV as pertaining to women only. Fragile X is a form of mental retardation (MR) linked to the X chromosome and the other three are all presumed to be the result of distinct single gene defects. Mental retardation generally has a high correlation with autism, with seventy percent of all autistics testing in the MR range.
Both MR and autism are associated with a higher rate of perinatal complications than are experienced by the general population. This softens the line between genetic and environmental etiologies. However, pregnancy complications (first trimester bleeding, for example) in autism are often “minor” compared to those preceding other MR births. Both MR and autism are associated with infantile spasms and congenital rubella exposure; however, MR is associated with many more disorders that autism is not (like cerebral palsy and Down syndrome). Hart points out that although autism is considered a birth defect, there are no chemical or chromosomal tests yet to predict it. Although it is considered a developmental disability, we still do not know which effects to label “primary” and which to call “secondary.” He adds that although there is a genetic link, we still do not know what that link may be and what added environmental factors, if any, trigger autism’s onset. He relates the belief that autism is caused by “modern life-styles,” specifically the pollutants of modern times, citing reports that incidents of autism have been increasing. However, he feels autism is not increasing any more than all birth defects are, and it is very likely that more traits are now being called autistic that were not previously.
ASA’s “How They Grow” indicates some early illnesses—like encephalitis or prolonged high fever—have been shown to produce cases of autism. However, autism is mainly an organic brain defect. There is some generally abnormal biochemistry and/or a physical malformation of the brain. Even if the brain areas are of normal structure, their neurotransmitters do not function properly. Hart indicates that autism’s broad range of indications and possible etiologies makes the broad term “autistic syndrome” more appropriate. “Researchers call conditions that show hereditary as well as environmental factors syndromes. . . Accordingly, many professionals use the term ‘autistic syndrome’ when discussing the biology and causes of this disorder.”
There are many potentially genetic etiologies, but also many non-genetic ones. One school of thought is that exposure in utero or in infancy, to other diseases (like rubella) may trigger autism. However, this is not supported. Complicating this hypothesis is an aspect unique for autism, but not exclusively the domain of this disease; it is a disability which is congenital, but whose symptoms do not appear (or are not reported) until the toddler years. As mentioned in the previous chapter, there are no metabolic or chromosomal tests to predict autism. Although there are many nonpsychological theories about what causes autism, the main three are that: (1) it is a neurobiological disorder, (2) it has a genetic component, and (3) because it often occurs in conjunction with other handicapping conditions like epilepsy, there may be an autoimmune/environmental etiology. However, none of these causal “theories completely explains autism. At best, some contain hints and clues that may one day answer the questions of how it occurs, how the body’s chemistry is affected, and what parts of the brain are involved.”
Bauman and Kemper posit that part of the trouble with zeroing in on autism is that it is manifested in a spectrum of symptoms that vary in both appearance and severity. There may be many subtypes of autism. ASA’s “How They Grow” booklet calls autism “a class of disabilities,” not one, and adds that it may occur with other brain disturbances. A third of autistic individuals develop epileptic-like seizures in their teen years and a majority of autistics test in the mentally retarded intelligence range. To use Rapin’s nicely put phrase, one of the “less-damaging terms” for autism is “pervasive developmental disorder” (PDD). The DSM-IV lists five PDDs, including the open-ended “PDD-Not Otherwise Specified.” Despite the damage-control aspect of this range of labels, it is probably more accurate in that PDD indicates a spectrum of disabilities.
Further complicating the disorder, according to Gerlach, is that there is no test to diagnose autism, only to rule out similar diagnoses (like mental retardation and Fragile X syndrome). Autism itself is diagnosed behaviorally and is often viewed therefore as a behavioral disorder. However, obvious evidence of genetic predisposition, as well as correlation with some pre- and perinatal infection exposure, introduce biological factors.
Conclusions Drawn from the Literature Review
Autism’s etiology is unclear. A mixture of preconditions could trigger it. It could be biologically based. It could be congenital, or environmental, or none of these. While this chaotic uncertainty is not promising for the future of autism research, it is an ideal framework from which to map a bibliometric research project. The range of etiologies being investigated, the international scope and number of countries publishing articles on the topic, and its relatively short evolution over the last half of the twentieth century are promising components for a structured application of bibliometrics.
The next chapter lays out the design of this bibliometric study. From choosing sources, through deciding which facets of the sample would be analyzed, to describing the methods used to count, compare, and contrast the facets, the methods chapter provides the founding structure from which observations can be made and on which results can be built.
lukate@charter.net