Abstract
Keywords
The combined 48th Annual Meeting of the American Academy of Child and Adolescent Psychiatry and the 14th Annual Meeting of the Faculty of Child Psychiatry of the Royal Australian and New Zealand College of Psychiatrists is a historic occasion for a number of reasons. It is the first combined meeting of the two organizations. It is also the first Julian Katz oration since Julian passed away in August 2001. As the first full-time academic in Australian child psychiatry, Professor Katz made a significant contribution to the teaching and practice of child psychiatry in Australia and New Zealand. His interest in psychology was first stimulated at high school by reading Freud, so it is perhaps fitting that the present paper should remember Freud's Project for a scientific psychology.
In 1995, two hundred neuroscientists, cognitive scientists, analysts, philosophers and geneticists gathered under the auspices of the New York Academy of Sciences to mark one hundred years since Freud [1] published his Project for a scientific psychology. The proceedings were subsequently published as a monograph in the Annals of the New York Academy of Sciences [2]. According to Trupp [3], Freud's extensive experience of biology and neurology lay to the left of the Project and psychoanalysis to the right. Solms [4] emphasized that Freud was aware that the state of neuroscientific knowledge in his time barred excessive confidence in his physiological speculations. Karl Pribram [5] emphasized Freud's visionary perspicacity in anticipating many details central to today's investigations such as pre- and postsynaptic potential gradients, and a double feedback between basal forebrain and cortex to produce the attentional processes necessary for reality testing. When Freud wrote his monograph, he inferred that at some future date our knowledge of brain function would allow the development of a brain-based psychology. He stated ‘consciousness is the subjective side of all psychical events, and is thus inseparable from the physiological mental process’.
According to Andrew Brook [6], Freud transformed his mechanistic neuroscientific explanations of the mind into a metapsychology of the unconscious, but there remained in Freud's explanations a duality between interpretations aimed at uncovering unconscious meanings, and his mechanistic metapsychology. He claims that this divide between mechanistic models out of the scientific image, and intentional explanations out of the manifest image continues to this day. For example, Pickering and Chater [7] argue that cognitive science and folk psychology, meaning intuitive psychology, treat entirely different domains and should remain separate endeavours. Paul and Patricia Churchland [8, 9] believe that intentional psychology, deriving from common sense psychology, has never got anywhere as a science, and that what we have to do is return to Freud's dream of a complete neuroscientific understanding of the mind in terms of patterns of neural activity.
Solms [4] argues that Freud's ‘some day’ has arrived, namely the neuroscience of the mind has developed to such an extent in the decades since Freud's death in 1939, that we are in a position to construct a model of the mental apparatus on the basis of detailed neuroscientific knowledge. Perhaps the developmental, genetic, and neuropsychological insights from the study of ADHD are the beginnings of such a model. However, it could be argued that child psychiatry still grapples with the dualist dilemma. Leon Eisenberg [10] has often been quoted as saying that psychiatry suffered during the psychoanalytic era from being ‘brainless’ and during the biological era from being ‘mindless’.
A recent landmark paper by Rutter and Sroufe [11] entitled ‘Developmental Psychopathology: Concepts and Challenges’ outlined three key issues: (i) the understanding of causal processes, (ii) the concept of development, and (iii) the continuities between normality and pathology. Rutter and Sroufe describe the rising influence of behaviour genetics on our understanding of causality. Also, issues in the progressive reorganization of behaviour as the organism transacts with the environment during development. Finally, the delineation of what is involved in the continuities and discontinuities between normality and pathology is a central concern of developmental psychopathology. The authors point out that regardless of whether or not the underlying liability to psychopathology is dimensional with a continuum spanning normality and pathology, categorical decisions will often be required for practical decision-making purposes.
Clinical research development
My own interest in hyperactivity began during my training, when I treated a middle-aged mother, who had coped well with raising poodles, but became depressed at having to cope with her 3-year-old hyperactive twins. This observation was the beginning of a life-long interest in the phenomena and psychopathology of what is now called Attention Deficit Hyperactivity Disorder.
After returning to Australia, to what subsequently became the Avoca Clinic at the Prince of Wales Hospital in Sydney, I became involved in a study of the Feingold Diet, an exclusion diet, alleged to treat hyperactivity, by an allergist, Ben F. Feingold from Kaiser Permanente in California [12]. I was fortunate to obtain the advice of John Werry, who suggested measures for each symptom of the Hyperkinetic Syndrome. These included a Continuous Performance Task (CPT) for attention and a stabilometric chair for activity, and a Draw-A-Line Slow task for inhibition. We developed a rather primitive CPT, using a video recorder, and developed norms for the task, by testing children in preschools and schools. There followed a number of papers based on the CPT, including normative [13], social class [14], diagnostic [15], as well as a ‘diet’ paper [16]. The ‘normative’ paper ‘showed a marked increase in capacity for sustained attention and inhibition around four years of age with significant social class effects. We also showed that an age-normalized function discriminated 60% of a group of conjointly clinically diagnosed hyperkinetic children from matched normals [15]. Keith Conners’ modifications, reported at this meeting (Institute on Neuropsychological Assessment in Child and Adolescent Psychiatry), have increased sensitivity and specifity to over 80%.
In 1988, we conceived a study, which showed that haloperidol blocked the beneficial effects of methylphenidate on CPT performance in children with ADHD [17]. As haloperidol is primarily a dopamine-blocking agent, this pointed to the dopamine hypothesis. A subsequent reanalysis of the haloperidol data showed that the effect was more pronounced in the second half or block of the test, suggesting that methylphenidate was having an effect on capacity to sustain attention [18]. A visit to Jim Swanson's laboratory school at Irvine, California, where we talked about the dopamine hypothesis of Attention Deficit Hyperactivity Disorder gave rise to a paper in 1991 on the Dopamine theory of Attention Deficit Hyperactivity Disorder, at a time when most interest was focused on noradrenergic theories. Recently, Jim Swanson and myself have co-authored an update on the dopamine theory, which, continues to exert an important heuristic effect on much current research [19].
Brain mapping
A further application of our CPT work has been our collaboration with the group at Swinburne University of Technology with Richard Silberstein and Maree Farrow [20, 21]. They have utilized a technique known as Steady State Probe Topography, which enables examination of disturbances in the spatial distribution and dynamics of brain electrical activity. Unlike many brain mapping techniques, it provides very good resolution of brain activity over time. The investigators examined brain activity over the course of an A–X CPT task, in which children are asked to press a button when they see an X preceded by an A. There were 60 male ADHD subjects who met criteria for ADHD and were medication free; they underwent 2 recording sessions on the same day. The second session was conducted 90 minutes after administration of methylphenidate (0.3 mg/kg). The steady-state visually evoked potential (SSVEP) was recorded from 64 scalp electrodes, while subjects performed the AX version of the continuous performance task. Differences in SSVEP amplitude and latency during the A–X interval between before and after methylphenidate conditions were examined. During the appearance of an A, there was increased parieto-occipital activation, after methylphenidate, compared with before, and reduced latency in right frontal and posterior regions. The right parieto-occipital changes were most significant. After disappearance of the A, increased parieto-occipital activation was still evident until the appearance of the target X, as was reduced posterior latency. Thus, the most significant effect that methylphenidate produced was increased activation and reduced latency in the right parieto-occipital region, that was sustained throughout most of the A–X interval. Transient frontal and temporal latency reductions also occurred after methylphenidate. The results were thought to suggest that prefrontal and right parietal processes are involved in the maintenance of activation over the course of the CPT, allowing a ‘working’ retention of the parameters necessary for correct task performance. That is, there was sustained fronto-parietal activity between the A cue and the X target (Farrow, in press).
These data are consistent with recent theories about the importance of working memory deficits in ADHD [22] and are also consistent with the work of Patricia Goldman-Rakic [23] on brain areas involved in the delayed-response task in primates. Joseph Ledoux [24] has emphasized the importance of working memory in integrating long-term memories with current tasks or problems, and we could speculate about the role of working memory in various theories of executive function in ADHD.
Behaviour genetics
The pursuit of a biological paradigm was renewed by the idea of a twin study of ADHD, utilizing the Australian Twin Registry. With the help of David Hay, a psychologist/ behaviour geneticist, who had been working with twins and twin studies as long as I had worked on ADHD, we designed a DSM-III-R based questionnaire for circulation to families of twins and their siblings. Influenced by my clinical experience of comorbidity, and the importance of early language and reading problems, we included questions on Oppositional Defiant Disorder (ODD), Conduct Disorder (CD), Speech and Language development and Reading problems, as well as pre and perinatal problems. Comorbidity has turned out to be of great interest in later studies. The Australian Twin Study of ADHD has now been going since 1990. We were initially funded for one year by the Government Employees Medical Research Fund, and subsequently have had three grants from the National Health and Medical Research Council of Australia.
We now have three four-year waves of data, as well as an adult twin cohort, and an associated Western Australian population twin sample. The study has played a major role in understanding the underlying causes of ADHD and comorbid conditions. We have recently edited a text on the behaviour and molecular genetics of ADHD ‘Attention Genes and ADHD’ [25], which has begun to show how quantitative and molecular genetics can contribute to the understanding of phenomenology. We have been among the first to show:
1. ADHD is more common in twins, with 10–11% incidence for DSM-III-R ADHD in twins and 8% in siblings. Also, ADHD is often associated with the speech, language and reading problems that frequently occur in multiples. Speech and reading problems are as predictive of ADHD as gender and twin or sibling status [26]. 2. Despite this, the determinants are very similar in twins and their singleton siblings as well as in males and females. Rhee [27, 28] has used our twin data to compare the polygenic multiple threshold model with the constitutional variability model. The former asserts that probands of the less frequently affected sex (females in the case of ADHD) should have more affected relatives than should the probands of the more frequently affected sex, while the latter assumes that different causal factors are important. Rhee confirmed the multiple threshold model for males and females by showing that cotwins and cosiblings of female probands had a higher number of ADHD symptoms on average (mean = 5.54, SD = 4.35) than did cotwins or cosiblings of male probands (mean = 4.24, SD = 4.0). 3. We have also shown that ADHD is inherited as a continuum throughout the population, so that the cutoff scores on symptom number may be arbitrary but necessary, and that the heritability is high (75–90%) irrespective of how ADHD is defined [29]. This paper, which received an award from the American Academy of Child and Adolescent Psychiatry in 1997, illustrates Rutter and Sroufe's point [11] about the distinction between arbitrary cut-off points and underlying causal processes, suggesting that a trait varies throughout the population. This does not mean that at the extremes, causality may not differ. Rutter cites the example of IQ which functions dimensionally as a predictor of educational attainment or social competence right across the range, extending from profound retardation to above-normal levels of functioning. On the other hand, for the most part, the causes of severe and profound retardation are quite different from variations in IQ within the normal range. 4. We have shown that the reading problems, often seen in ADHD, may have a genetic association with ADHD, rather than being a cause or a consequence as discussed by Stevenson in his chapter on Comorbidity of Reading/Spelling Disability and ADHD [30]. While the model is somewhat complex, it shows that all of the genetic variance in inattention is shared with genes affecting reading, while hyperactivity/ impusivity is partly genetically independent. This work together with international studies, such as Wilcutt et al. [31] promises to tease out the somewhat controversial relationship between reading and ADHD. 5. The work of Irwin Waldman and colleagues at Emory University on our twin data has demonstrated that ODD is not an alternative diagnosis to ADHD, but is strongly related at the genetic level, as outlined in his chapter in our book ‘Causes of the overlap among symptoms of Attention Deficit Hyperactivity Disorder, Oppositional Defiant Disorder and Conduct Disorder’ [32]. 6. We also plan to examine pre- and perinatal problems including minor physical abnormalities, motor coordination and dermatoglyphics, as well as family management style and expressed emotion in monozygotic and dizygotic twins discordant for ADHD, identified from our large cohort, as a test of environmental effects.
A further application of our behaviour genetic work relates to comorbidity with oppositional defiant and conduct disorders. As already noted, Irwin Waldman [32] has shown that ODD shares considerable genetic influences with Attention Deficit Hyperactivity Disorder. His analysis of our twin data, using a Cholesky model which decomposes the covariance of the conditions, indicates shared additive genetic influences of 95% on ADHD, 68% on ODD and 51% on CD. On the other hand, specific common environment was greater, 58% for CD only. Shared environmental influences contributed minimally to ODD symptoms (c 2 = 0.01) and not at all to ADHD symptoms. However, genetic and shared environmental influences were both moderate for CD symptoms (h 2 = 0.51, c 2 = 0.34). Thus, the results indicated that while there was considerable overlap in the genetic and the environmental influences on the three disorders, and the shared environmental influences on ODD and CD overlapped completely, the influences of shared environment was much greater on CD than ODD. This finding may go some way to explaining the somewhat unexpected findings of the Multimodal Treatment Study of Children With ADHD (MTA) [33]. The study found only small differences between carefully controlled medication treatment alone and combined medication, behaviour therapy treatments for ADHD comorbid with ODD [34]. The overlap of aetiological genetic influences for ODD and ADHD, with minimal or no shared environmental effects may help explain why both respond to medication, while CD requires structured family and behavioural interventions.
A further potential of the twin work is to examine the effects of genes and environment over time. As described by David Hay in his chapter on developmental genetics [35], we can model the ‘continuity’ model, where the same influences apply at all ages; the ‘discontinuity’ model, where there are different determinants of behaviour at each life stage; the ‘environmental development’ model, where initial genetic influences become less important with environmental experience; and the ‘genetic development’ model, where after strong early influences of the family environment genetic effects become more apparent. For example, Plomin [36] has described how the heritability of IQ, general intelligence increases from 20% in infancy to 40% in childhood to 60% in adulthood. Another example in our twin study showed there were specific common environmental effects on hyperactivity/impulsivity at time one (age 4–12 years) but these were not apparent 4 years later at time two.
Genes and environment
We have thus established a unique database, which hopefully will continue to be of interest. It has been an exciting and rewarding effort, but our study and numerous other twin studies raise an important basic question. Rowe [37] points out that for many problems of childhood, the role of common family environment in twin studies appears to be small. Plomin and Daniels [38] emphasized the importance of non-shared rather than shared environment in explaining differences between siblings. In the ‘continuum or category’ analysis of DSM-III-R ADHD, ‘common family environment’ accounted for only 13% of the variance. Similar findings caused many investigators [37, 39] to conclude that variation in shared rearing experiences is a weak source of trait variation. However, these findings contrast with clinical experience that poverty, family stress and violence are powerful accompaniments of child psychopathology, and epidemiological findings suggesting a relatively rapid increase over time in overall population rates of psychopathology [40]. Rutter [41] in another review states that while the above twin study findings mounted a serious challenge, the rejection of psychosocial influences was unwarranted. First, there is an implicit assumption that the origin and mode of mediation of a risk factor are synonymous, and act similarly on all syndromes. That this is not so is illustrated by differences in the reason why people smoke, and the risk factors involved in the effects of smoking. The finding above of difference in common environmental effects on ODD and CD is another example. Rutter also questions whether the equal environments assumption applies equally to MZ and DZ twins, and whether there may in some circumstances be an environmental effect within MZ twins, again leaving room for environmental effects.
Another approach to the question of genes and environment may be provided by the work of James Flynn, a New Zealand psychologist and his colleague William Dickens of the Brookings Institute in Washington DC, reported in the Psychological Review [42]. Their model attempts to explain the paradox that while the studies of Jensen [43] have found a low influence of environment on IQ, a worldwide survey of IQ trends over time found that the current generation outscores the previous generation by between 9 and 20 IQ points. The size and speed of these IQ gains virtually dictates an environmental explanation, because genes just don't change that fast. The Flynn model assumes that a small genetic advantage (or disadvantage) for a particular trait will become matched with environments for that trait. Thus, if you are born tall and fast, and find you are good at basketball, you will spend more time playing basketball, and perhaps end up a Michael Jordan. That is, thanks to the powerful multiplying effects between talent and environment, a modest genetic advantage may turn into a huge performance advantage. Flynn and Dickens call this effect a ‘social multiplier’ effect. From our perspective as child psychiatrists, the importance of the model is to suggest that the child participates in creating his or her environment, and that we need to understand how particular traits create particular environments, in order to enhance or decrease social multiplier effects. This could be interpreted as social Darwinism, but at least it brings the child back into the attachment equation.
Rutter [44] has described at least five different mechanisms involved in the operation of genetic risk factors. First, because genetic effects are on proteins, protein products could be directly implicated in the causal processes for a disorder. Genes that function in this way have not yet been identified for psychopathology. Second, genes may have relatively direct effects on partfunctions that when combined with other functions make up the disorder. This could be so for components of dyslexia, or ADHD components. Third, genes may influence temperamental (or other) dimensions that, in themselves, do not constitute the disorder, but which serve indirectly to increase the risk of disorder when combined with other risk factors. For example, the role of neuroticism in the increased risk for affective and anxiety disorders. It may also be the case for noveltyseeking in the risk for ADHD. Fourth, genes may act through their role in increasing (or decreasing) environmental risk exposure. According to Rutter, this may come about through passive, active or evocative geneenvironment correlations. Passive correlations reflect the fact that parents pass on genes, as well as create environments for upbringing, for example in relation to personality disorder. Here genes increase risk through their influence on environmental risk exposure as discussed above. Fifth, genes may exert their effects through influences on susceptibility to environmental risks, both physical and psychosocial.
Molecular genetics
The Human Genome Project has demonstrated that we are part of the natural world, and our behavioural systems have evolved from that world. While the implications are yet to evolve I believe that our behavioural systems can be understood in both biological and behavioural terms. To quote from the Human Genome Sequencing Consortium ‘the rediscovery of Mendel's laws of heredity in the opening weeks of the 20th century sparked a scientific quest to understand the nature and content of genetic information that propelled biology for the last hundred years’. We now have the enormous potential of the human genome sequence in public databases. What does this mean for behavioural disorders? [45]
We know that the ‘one gene, one diagnosis’ hypothesis is wrong other than for a few conditions such as Huntingtons Disease. Generally multiple genes are involved in behaviour, and these genes interact with the environment. Nearly all behaviours that have been studied show moderate to high heritability, and because of its high heritability ADHD should benefit from molecular genetic advances.
So far, two candidate dopamine genes, the dopamine transporter (DAT1) [46] and dopamine receptor (DRD4) [47] have been investigated for associations with ADHD. The polymorphisms of the dopamine genes are defined by variable numbers of tandem repeat sequences (VNTR) of bases on the genes. The most common variants of the DAT1 gene are specified by nine or ten repeats of the 40-bp sequence, and the most common variants of the DRD4 gene are specified by two, four or seven repeats (copies) of the 48-bp sequence. Swanson [48] reported that three published studies of the DAT1 gene reported an association with ADHD, while four of five published data on DRD4 reported positive associations. He speculated that the 7-R allele of the DRD4 gene may produce a receptor that is ‘subsensitive’ to dopamine, while the 10-R allele of the DAT1 gene may be associated with a dopamine transporter that is abnormally efficient at the re-uptake process. There have been both replications and non-replications in recent years, but Faraone [49] has done a meta-analysis of DRD4 studies, which shows that despite some non-replications, the association with the DRD4 receptor gene is significant with an odds ratio of around 1:4. While this is relatively small it is significant, and suggests that other genes or quantitative trait loci will be involved. The Dopamine Transporter gene was the first of the dopamine system genes shown by Cook [46] to be associated with ADHD. This has now been replicated in a number of studies [50–52].
Arnsten [53] has also suggested a role for noradrenergic neurotransmitters, which has created interest in α-adrenergic genes. Also, comorbidity with learning disability has given rise to further interest in ‘reading’ genes. For example, Grigorenko et al. [54] described an association of Chromosome 6 with phonetic awareness, and Chromosome 15 with whole word reading. An award-winning study by Winsberg and Comings [55] found that homozygosity of the 10-repeat allele of the dopamine transporter gene was characteristic of nonresponse to methylphenidate. This was the first ‘pharmacogenomic’ study in Child Psychiatry and predicts future approaches.
A number of ethical issues are discussed by Sir Michael Rutter [44] in the final chapter of our book. First, the concern that genetic research will result in neglect of social research. Here, according to Rutter, it is clearly important that we research both genetic and environmental risk factors. We know that poverty, social inequity, racial discrimination, family breakdown, poor parenting, child neglect and abuse are environmental risk factors, but what we do not know is how a particular genetic make-up interacts with environmental risk or protective factors. Second, the issue of genetic determinism, which probably becomes a non-issue when one considers that genetic influences on behaviour are multifactorial and act in combination with or through environmental risk factors. Third, Rutter discussed the notion that genetic understandings will medicalize social problems, and factors such as poverty and neglect will be minimized. Here again we should understand that aetiology and treatment are not necessarily the same thing. For example we now believe there are strong genetic determinants for autism, but the treatment so far is social and educational. There are many causes for fractures but treatments are entirely remedial. Treatment is always carried out in a social context, even when biological treatments are quite powerful. The issue of stigma has also been raised, but, if anything, genetic understandings should reduce stigma as we come to understand the relationship between genes and environment and the fact that everyone probably carries risk genes for some conditions under some circumstances. The question of insurance may be important, but again the multifactorial nature of disease makes it more difficult to misuse genetic information, and according to Rutter, society has a responsibility in deciding who bears the cost of genetic risk, the individual or society. The issue of free will and genes has sometimes been raised in relation to ADHD. It is important to understand that genetic effects are probalistic rather than deterministic, and exert their effects in particular environmental circumstances, so the notion of a genetic fate map is premature, and hopefully, at least in part, fanciful.
Conclusions
For our group, the study of ADHD moved from a depressed mother of hyperactive twins, to objective measures of hyperactivity and to the dopamine hypothesis, as well as brain mapping, and quantitative and molecular genetic studies.
To go back to Freud's project, Allan Schore [56] has argued in a paper entitled ‘A century after Freud's project, is a rapprochement between psychoanalysis and neurobiology at hand?’. I am sure the biology of childhood trauma is important area of rapprochement. Joseph Le Doux [24] in his book on ‘The Emotional Brain’ has a delightful section called ‘Little Albert meets Little Hans’.
For child psychiatry, the concept of biological subsystems [57, 58] which are amenable to study by neuroscience, promises to rescue us from the dualism begun by Freud when he abandoned his Project, and which haunts us today when we divide into those who classify and those who interpret. If we are to attract imaginative young graduates into child psychiatry, we need to revisit Freud's questions in terms of evolutionary biology, genetics and neuroscience and realize his vision of a basic biological science with an associated metapsychology. The 21st century should let all flowers bloom, and offer us many exciting views of ourselves.