5 Explaining autistic conditions: the biological level
5.1 Are there genetic factors in autistic spectrum conditions?
Section 4 focused on explaining the characteristic symptoms of ASDs in terms of socio-cognitive functioning. In this section the focus shifts to the biological level: what biological influences might both trigger and maintain atypical functioning in areas like theory of mind, global information processing and emotional relatedness?
As was emphasised in Section 1, biological perspectives on ASDs reflect several ‘sub-levels’. These sub-levels offer a complex mix of complementary, conflicting or co-existing accounts, both within themselves and in relation to other levels of explanation covered in the course. We start by considering genetic factors that might explain why ASDs often affect more than one member of the same family.
In investigations of whether genetic factors affect behaviour patterns, particular interest focuses on comparisons between identical twin (mono-zygotic or MZ) and non-identical twin (di-zygotic or DZ) pairs. If there is a genetic influence, the concordance rate for MZ twins should be particularly high, because both members of twin pairs have the same genetic material. For DZ pairs the genetic relationship between the twins is the same as that between ordinary siblings. Concordance rates for DZ twins and for siblings should be similar: higher than in the general population, but much lower than for MZ twins.
Studies documenting higher rates of concordance for autistic spectrum symptoms among MZ twins compared to DZ twins are described in Box 8.
Box 8: Twin studies of autism
Folstein and Rutter (1978) investigated 21 same-sex pairs of twins, including 11 MZ pairs and 10 DZ pairs between the ages of 5 and 23. Each pair included one member diagnosed as autistic. Of the 11 MZ twin pairs, 4 were concordant for classic autism, i.e. both twins had autism. Of the 10 DZ pairs, none were concordant for autism. However, the concordance rates rose considerably when all autistic spectrum symptoms were taken into account. Seven out of 11 MZ twins unaffected by classic autism had some autistic type symptoms, particularly involving language. This was true for only one of the unaffected members of a DZ pair.
Folstein and Rutter's finding have been extensively replicated. Bailey et al. (1995) re-contacted all participants in the earlier study. They re-checked diagnostic and medical assessments and augmented the overall sample, providing data on a total of 25 MZ and 20 DZ same-sex twin pairs. The findings, which are summarised in Table 1, confirmed and extended those of Folstein and Rutter. The overall MZ concordance rate for classic autism in this combined study is 60 per cent. However, this concordance rises to 92 per cent if twins showing a broader spectrum of autistic-type symptoms are taken into account. The autism concordance rate for DZ twins is 0 per cent but rises to 10 per cent when autistic spectrum symptoms are included.
|MZ % concordance||DZ % concordance|
|Both twins autistic||60||0|
|One twin autistic; other with spectrum symptoms||32||10|
The markedly raised concordance for full autism in MZ twins has been interpreted as evidence for a genetic predisposition. The presence of autistic-type difficulties in most of the non-autistic identical twins, and one of the non-autistic DZ twins, is consistent with the idea of an autistic spectrum, and suggests a genetic basis for this spectrum.
What other explanatory factors might be considered when interpreting these twin studies?
Some have argued that the concordant MZ pairs developed autism because they were exposed to damaging social influences during childhood that did not affect the DZ pairs to the same extent. This argument is difficult to sustain. No convincing model has been offered to explain how such difficulties could arise purely from social influences that have such a profound and early impact on identical twins but not on non-identical twins.
If the genetic interpretation of MZ concordance rates is correct, one puzzling question is ‘why are there so many identical twin pairs who are not fully concordant?’ (that is, they share spectrum difficulties, but are not equally severely affected). Folstein and Rutter proposed that the more profoundly affected member of these pairs might have been exposed to additional ‘environmental’ hazards in the womb or during birth. They examined birth records for all twin pairs in their study, for evidence of problems such as a delay in breathing of more than five minutes, or a convulsion, which would be likely to cause brain damage. In a majority of cases the more seriously affected twin had suffered an additional birth hazard. This led Folstein and Rutter to propose a ‘threshold’ model of causation in which a genetic abnormality makes a child vulnerable to developing an autistic spectrum condition, and a birth hazard interacts with this predisposition to ‘push’ the child over the threshold into fullblown autism. Pursuing this argument further, Folstein and Rutter speculated that in some cases (for instance non-concordant DZ twins) brain damage caused by a birth hazard alone might be sufficiently strong to produce autism.
This is an important but controversial model, since it suggests that different cases of autism might arise from different causal influences, working either together or separately. This has been accepted by researchers such as Peeters and Gillberg (1999), who argue that the biological causes of different cases of ASDs are multiple, with only a proportion being genetically triggered.
However, Bailey et al. (1995) provided arguments for a different interpretation of Folstein and Rutter's data showing an effect of birth hazard. They produced evidence that birth hazards such as a delay in breathing are a result of earlier ‘sub-optimal’ development due to autism, not a cause of autism. In other words, the members of twin pairs who experienced birth traumas did so becausedamage that would later result in autism was already affecting robustness and responsivity in the womb. They concluded that autism is a ‘strongly genetic disorder’, in which an initial genetic fault triggers atypical development of the brain and nervous system, which in turn leads to the observed behavioural symptoms and socio-cognitive deficits.
This discussion highlights two different models that identify genetics and brain damage as separable biological influences. Folstein and Rutter's model sees these influences as ‘adding together’ or interacting, whereas Bailey et al. see them as part of a single chain of influence leading from genes to brain damage to behaviour. Both models allow for environment: the first sees an unfavourable environment in the womb as something that adversely influences the baby before birth; the second sees the baby's own ‘sub-optimal’ development as influencing his/her environment, for instance by reducing the baby's intake of oxygen before birth. But this model begs the question of why MZ twins with the same genetic material (and therefore the same genetic ‘faults’) should differ in their foetal robustness. It seems difficult to avoid the conclusion that a two-way interaction between the foetus and his/her pre-natal environment leads to more or less severe outcomes. This indicates the complexity existing among models at the biological level of explanation.
Another question raised by the twins studies is how to interpret the concordance rate for the DZ twins. According to a genetic hypothesis, the rate of concordance in DZ twins should, like that for family members, be higher than predicted by the incidence of ASDs in the general population. Bailey et al.'s data support this prediction. Extrapolating further, families with one autistic member should be relatively likely to have others with ASDs in the immediate or wider family tree.
This has been investigated in a range of research studies. For instance, Bolton et al. (1994) compared the incidence of ASDs in the families of individuals with autism and in control families. The results indicated a significant clustering of autism and autistic type conditions in relatives of individuals with autism, with an overall rate of 20 per cent, very similar to that quoted for DZ twins in the twin studies. Once again, the distribution of symptoms within family members supported the idea of a spectrum ranging from classic autism in some family members to extremely subtle symptoms in others. Gillberg (1991) carried out a similar study in which he looked at the incidence of Asperger's syndrome and ASDs across three generations of certain families. One of the family patterns is shown in Figure 7.
(1) was the original patient. He is an unmarried man of 33 with Asperger's syndrome. He works as a lawyer. (2) is the mother of (1). She is described as highly intelligent with borderline Asperger's symptoms – pedantic and friendless. (3) is the eldest brother of (1). He was diagnosed with classic autism at the age of four and lives in a group home. (4) is the middle brother of (1). He has borderline Asperger's symptoms, including odd pedantic speech. He is married despite his social gaucheness. (5) is the first-born son of (4), aged three. He is described as showing signs of classic autism.
Controversially, Baron-Cohen et al. (2002) suggest a possible evolutionary basis for family patterns of ASDs. Their model embraces the notion introduced in earlier sections, of a cognitive phenotype for autism, characterised by poor understanding of how minds work, coupled (at least in high-functioning individuals) with very good understanding of domains governed by physical laws, such as physics and engineering. Baron-Cohen et al. point to the obsessions that many autistic children have with machines, and provide evidence for precocious understanding of how mechanisms work, among children with ASDs. They also cite survey evidence that professions such as engineering and science predominate among the parents of people with ASDs. They argue that, expressed in a mild form, this way of engaging with the world might have had selective evolutionary advantages. The drawbacks of one or more members of a community having poor social understanding would be offset if these individuals had an enhanced understanding of physical causality, since this would enable them to fulfil useful functions such as constructing robust dwellings, or predicting the path of approaching storms. The full implications of these intriguing ideas have yet to be evaluated.
Overall, the discussion in this section points strongly to a genetic influence in ASDs, but probably does not imply that there is a ‘gene for autism’. Although certain inherited disorders such as phenylketonuria are known to be due to a single gene fault, the twin and family pattern in ASDs are most likely to indicate influences that are polygenetic, due to the combined effects of multiple genes. Ideas about which genes, on which chromosomes, might be involved, and whether these are the same genes in all cases of autism are extremely controversial. Equally, the mechanisms by which genetic and/or chromosomal abnormalities play a predisposing role in autism are not understood. However, it is highly likely that genetic influences have organic effects – particularly on the early development and functioning of the brain and nervous system.
Concordance rate: Measure of how frequently a phenomenon or condition co-occurs in two sets of individuals, particularly those who are related, such as twins.
Phenylketonuria: A metabolic disorder in which excessive amino acid levels in the blood cause brain damage if untreated. Caused by a known fault on a single recessive gene.
Polygenetic: The combined influences of a number of different genes acting together, as opposed to the influence of a single gene.
Organic effects: Generic term for influences affecting body organs and systems.