Psychiatric disorders as defined by the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) [1
] have been hypothesized as reflecting the extreme end of underlying, continuously distributed traits [2
]. In line with this, the behaviour of individuals suffering from psychiatric disorders differs quantitatively but not necessarily qualitatively from the behaviour of individuals without psychiatric problems. The threshold of what is and what is not abnormal is to a certain extent arbitrarily determined, but patients have in common that their behaviour interferes with their normal life and cause the patient (and his/her environment) to suffer. Why certain people pass this threshold and are diagnosed with a disorder and others do not is determined by additive and interacting genetic and environmental risk factors [5
]. Studies have shown that psychiatric disorders have genetic and environmental underpinnings which probably contribute to certain neurocognitive abnormalities that, in turn, lead to abnormal behaviour. It is theorized that these neurocognitive abnormalities form underlying, continuously distributed, vulnerability traits (endophenotypes) that heighten the risk for developing a disorder (phenotype) [7
]. In this context, neurocognitive abnormalities refer to mental functions that are mediated by brain processes; these mental functions are not directly observable, but may be manipulated and measured using experimental paradigms. Multiple endophenotypes interact to determine the finally observable behavior, the phenotype, which might be abnormal. In this context, the phenotype refers to directly observable symptoms of a disorder (Figure ).
The relationship between genetic and environmental risk factors, endophenotypes, and phenotype in ADHD.
Studying endophenotypes may have certain advantages over studying phenotypes. It has been proposed that endophenotypes may be more suitable for detecting risk genes, because endophenotypes are genetically less complex than phenotypes (i.e. related to fewer genes than phenotypes) and, therefore, probably stronger linked to these disease genes than phenotypes [7
]. Endophenotypes may also be useful in exploring different pathways leading up to the disorder: Patients having the same diagnosis may differ strongly in the number and severity of symptoms they portray, suggesting heterogeneity in the causal pathways [10
]. Creating more homogeneous subgroups of patients based on their endophenotypic functioning, may facilitate unravelling these differential causal pathways.
In the last two decades, substantial attention has been given to studying endophenotypes of psychiatric disorders. This has led to the discussion of what exactly constitutes an endophenotype and what criteria must be met for a neurocognitive function to be useful as candidate endophenotype [7
]. Since an endophenotype forms a link between susceptibility genes and the disorder, it follows that: (1) the neurocognitive dysfunction is heritable (and familial), in which at least partly the same genes influence the endophenotype and phenotype; (2) the neurocognitive dysfunction is associated with the disorder; (3) the neurocognitive dysfunction is observable in non-affected first-degree relatives of an affected individual, because first-degree relatives are likely to carry some of the susceptibility genes of the disorder.
Thus far, several psychiatric conditions have been targeted for candidate endophenotypes and considerable knowledge has been gathered on the usefulness of different endophenotypes in for example, bipolar disorder, schizophrenia, substance abuse, and depression [15
]. This study targeted Attention-Deficit/Hyperactivity Disorder (ADHD), since it is one of the most prevalent child psychiatric disorders, yet the knowledge on the usefulness of ADHD endophenotypes is still limited.
Two studies have failed to find neurocognitive impairments in parents of children with ADHD [20
] and two studies found no conclusive evidence of cognitive dysfunctioning in non-affected siblings [22
]. However, one study did find evidence for cognitive functions as endophenotypes [24
] and other studies, specifically targeting inhibition or interference control, found evidence for these functions as endophenotypes [13
]. In a previous study, we also found evidence for inhibition as well as visuo-spatial and verbal working memory as endophenotypes for ADHD [29
]. Moreover, it appeared that deficits in the various cognitive functions partly arose from the same genetic and/or environmental risk factors. Furthermore, evidence has been found for time reproduction as ADHD endophenotype, a function related to the sense of time [30
]. Less attention has been given to studying functions outside the cognitive domain, although ADHD is frequently associated with motor deficits in Caucasian and non-Caucasian subjects [31
]. In previous studies, we and others have shown that non-affected siblings have subtle problems, similar to their affected siblings, in motor timing, motor control, motor speed and variability, and speed of oculomotor control [33
], suggesting endophenotypes for ADHD may also lie inside the motor domain. Moreover, it appeared that some of the cognitive dysfunctioning in children with ADHD and their non-affected siblings may be related to problems already apparent on a simple reaction time task [31
Taken together, the majority of studies have found support for cognitive and motor endophenotypes for ADHD [24
]. These studies have in common that they administered one or a few measures tapping into a single domain. It is unlikely, however, that one such measure/domain mediates the relation between genotype and phenotype and can predict the phenotype, because it is unlikely that all children with ADHD and their non-affected siblings will show this endophenotype given the causal heterogeneity of ADHD [13
]. It is more likely that multiple endophenotypes mediate the relation between genotype and phenotype, and together are more powerful in predicting the phenotype. Therefore, the first aim of our study was to investigate if an endophenotypic construct, encompassing a broad battery of both cognitive and motor endophenotypes, is predictive of the ADHD diagnosis.
Practically all studies of ADHD endophenotypes conducted thus far have reported the same type of results: The group of non-affected siblings performs in between the affected siblings group and the normal control group. At a phenotypic level, non-affected siblings do not (or apparently to a lesser extent) deviate from the controls. The reverse appears to be true for affected children: Their phenotypic deficits are more pronounced than one would expect based on their cognitive and motor dysfunctioning. Therefore, the second aim of the paper was to confirm apparent observations that children with ADHD show more severe ADHD symptoms than one would expect based on their endophenotypic dysfunctioning, whereas non-affected siblings show less ADHD-like behaviour than one would expect based on their endophenotypic vulnerabilities.
The question that automatically arises if these observations can be confirmed, is why endophenotypic vulnerabilities are not proportionally related to deviations at a phenotypic level? It might be that certain factors moderate and/or mediate the relation between endophenotype and phenotype. Moderation would imply that the relation between endophenotype and phenotype is not comparable across different levels of the moderating factor [37
]. Mediation would imply that (a part of) the relation between endophenotype and phenotype can be explained through their correlation with a third factor. When this third factor is taken into account, the relation between endophenotype and phenotype disappears or weakens [37
]. The third aim of this study was to explore whether four factors (gender, age, IQ, and rater bias) had a (moderating and/or mediating) effect on the relationship between endophenotype and phenotype.
With respect to the first factor gender: ADHD is more frequently diagnosed in boys than girls [38
], probably because boys are more vulnerable to the disorder. Since affected children are more often boys than girls, whereas the gender ratio is more or less equal in the group of non-affected siblings, moderating and/or mediating effects of gender may possibly be accountable for the apparent non-comparable magnitude of group differences at the endophenotypic and phenotypic level. With respect to the second factor age: It is known from several studies that the severity of ADHD symptoms appears to decline to some extent with age [39
] and that the same might be true for underlying neurocognitive vulnerabilities [30
]. Yet, others have failed to find diminishing neurocognitive vulnerabilities with age [40
]. It may thus be possible that age has a moderating and/or mediating influence on the relation between endophenotype and phenotype. With respect to the third factor IQ: A frequently reported finding is that children with ADHD have on average a lower IQ than controls [44
]. It has been suggested that this lower IQ may underlie cognitive dysfunctioning, or vice versa, that cognitive dysfunctioning is at the heart of a lower intelligence, or that there is no hierarchical relation between both domains but both domains share common causes [47
]. Either way, since IQ seems both associated with the neurocognitive dysfunctioning as well as with ADHD behaviour, a mediating effect of IQ may be expected on the relation between neurocognitive dysfunctioning and ADHD.
An additional factor that may influence the relation between endophenotype and phenotype may be rater bias: Parents may underestimate the severity of ADHD symptoms in their undiagnosed/non-affected child and overestimate the severity of ADHD symptoms in their diagnosed/affected child. Teachers, however, may be less likely to be affected by this contrast effect, since they often do not have siblings of the same family in their class due to age differences between the siblings. Rater bias explaining the non-comparable magnitude of group differences at the endophenotypic and phenotypic level may thus be investigated by comparing the relation between endophenotype and phenotype as observed by parents and teachers.
The first aim was to investigate if an endophenotypic construct, encompassing a broad battery of both cognitive and motor endophenotypes, is predictive of the ADHD diagnosis. The second aim was to confirm apparent observations that children with ADHD show more severe ADHD symptoms than one would expect based on their endophenotypic dysfunctioning, whereas non-affected siblings show less ADHD-like behaviour than one would expect based on their endophenotypic vulnerabilities. The third aim of this study was to explore the (moderating and mediating) effects of four factors on the relationship between endophenotype and phenotype: gender, age, IQ, and rater bias.