Page 124 - Read Online

De Francesco et al. J Transl Genet Genom 2024;8:102-18 https://dx.doi.org/10.20517/jtgg.2023.51 Page 110

The comorbidity between Internalizing and Externalizing disorders: the p factor
The high frequency of comorbidity rates among psychopathologies has progressively directed research
towards a reformulation of their structural model, leading to a thorough reconceptualization of psychiatric
disorders. They are no longer regarded as independent entities but rather as portions of a continuous
[40]
diagnostic spectrum . Furthermore, these epidemiological studies have revealed two different types of
comorbidities depending on the experimental design adopted by researchers. Early studies have highlighted
a type of comorbidity that faithfully adheres to its definition, meaning the co-occurrence of two or more
disorders at the same time. Subsequent longitudinal studies, on the other hand, have allowed the
identification of sequential comorbidity, indicating an individual’s propensity to develop a specific disorder,
as they exhibit signs of a pathological condition within the same diagnostic spectrum at the time of
measurement . The categories of INT and EXT seem to account for sequential comorbidity. It has been
[40]
demonstrated that the portion of variance explained by these transdiagnostic factors is the one that remains
stable throughout the lifespan, unlike the specific variance associated with each disorder, which would
[41]
undergo more pronounced fluctuations over time . Moreover, the estimation of co-occurrence between
disorders from different domains has been examined over the years, both through cross-sectional studies
and through longitudinal research designs. The former showed that the correlation coefficient between
INT-EXT disorders was approximately 0.5, a substantial indicator of the coexistence of symptoms belonging
to the two transdiagnostic factors at the time of data collection . Furthermore, longitudinal studies have
[42]
made it possible to observe the persistence of this correlation over time, thus delineating a pattern of
comorbidity referred to as heterotypic. The discovery of this correlation between the overarching category
of INT and EXT has led to the hypothesis that there may be a common underlying susceptibility for both
transdiagnostic factors. This susceptibility is characterized by a broad set of common etiological factors that
would act independently of those specifically associated with each disorder . This hypothesis regarding the
[40]
presence of a common latent factor underlying susceptibility to various disorders is not entirely new in the
realm of the study of the mind. In fact, when tracing the history of psychometric research on intelligence,
the hypothesis of the existence of a single overarching factor influencing each cognitive ability, known as the
"g" factor, had already been proposed [26,40] .

Specifically, as is well known, specific items within psychometric tests measuring Intelligence Quotient
correspond to each of the abilities. While variations in scores on each subtest are attributed to specific
factors, the correlation between various items is attributed to the "g" factor [26,40] . Likewise, by applying the
same concept to the psychopathological domain, the correlation between various disorders is thought to be
influenced by the “p” factor, a term first introduced in a study by Caspi et al. in 2014 . In this study, the
[40]
authors confirmed the hierarchical structure of the model with the “p” factor at the top, even when
incorporating a third symptom category alongside INT and EXT, namely Thought Disorders. These
Thought Disorders were typically excluded in previous research because the symptoms associated with
them were considered uncommon in the general population .
[40]

The contribution of twin studies in the analysis of the comorbidity patterns between Internalizing
and Externalizing domains
Twin studies have given great support to the exploration of the underlying factors behind the frequent
comorbidity between INT-EXT domains . However, the literature exhibits significant variability in
[43]
findings mainly for two reasons. First, results are greatly affected by the instruments used by the authors to
assess the symptoms under study. For example, a finding that recurs throughout the twin study literature
reveals that generally, when parent-report measures are used, there is an overestimation of genetic
influences on the traits analyzed, a trend not found when self-report measures are used [2,44] . Secondly,
estimates can also differ based on the age of the sample under consideration. It is known that human life
encompasses various developmental stages, and the environmental context gradually changes. During