To analyze heterogeneity in ADHD experts in last decade advised to look beyond the lists of existing symptoms towards phenotypic measures that can be represented dimensionally and have well-theorized relationships with neurobiological systems, (Sonuga-Barke & Halperin, 2010; Insel et al, 2010; Fair D, Bathula D, Nikolas M, Nigg JT, 2012; Georgiades S, Szatmari P, Boyle M, 2013; Sanislow CA, Pine DS, Quinn KJ, et al, 2013). This is the nucleus of RDoC aims because children and adolescents with ADHD can be characterized in terms of several features that are best represented as dimensions and have well-theorized relationships to biological systems (Cuthbert & Insel, 2013).

Recently evidence suggests that measures of child temperament may predict ADHD symptoms (Einziger et al., 2018). Temperament is a characteristic of personality (Buss & Plomin, 1984; Crowell, 2016) and refers to individual, neurobiologically-based difference in reactivity, self-regulation and cognition (Eisenberg, 2012).

It has been previously emphasized (Nigg J., 2016), that Hypotalamic-Pituitary-Adrenal (HPA) axis, through the cortisol hormone, may represent a powerful biological measure of behavioural self-regulatory systems, activity level, cognition, temperament (Stadler et al, 2011; Martel et al., 2009; Sonuga-Barke, 2005) and arousal (Snoek, Van Goozen, Matthys, Buitelaar, & Engeland, 2004). Cortisol is released from the surrenal gland by means of the HPA axis activation, in response to catecholaminergic neurotransmitters (Ulrich-Lai & Herman, 2009). Cortisol is involved in the regulation of a wide range of body functions, including emotion processing (Skosnik, Chatterton, Swisher, & Park, 2000), awakening (Fries, Dettenborn, & Kirschbaum, 2009) and stress responses (Chrousos, Kino, & Charmandari, 2009).

Both cortisol and temperament may share self-response regulatory processes (Martel et al., 2008; Nigg, 2016; Ulrich-Lai & Herman, 2009), and one study on 70 healthy pre-schoolers indicates that children with SE temperament has higher morning salivary cortisol levels during their first week of a new primary school year (Davis, Donzella, Krueger, & Gunnar, 1999), hypothesized as a stress-induced effect.

To date, no studies have been conducted to study heterogeneity starting from the self-response regulatory processes between temperament and cortisol in children and adolescents with ADHD.

As for clinical application of HPA-axis and cortisol level in ADHD diagnosis and dimension, available studies are, as yet, either inconsistent (Bonvicini, Faraone, & Scassellati, 2016; Freitag et al, 2009, Corominas et al, 2012) or suggestive, but not significant, upon dimensional stratification of ADHD symptoms (Pinto et al., 2016).

The aims of this study to contribute to the issue of clinical heterogeneity of ADHD, analysing whether ADHD symptoms and co-morbidity traits simultaneously link to both cortisol level and temperament dimensions, as biomarkers of arousal and inhibited behaviour. To pursue our aim we formulated five specific research questions: i) Do temperament dimensions and awakening cortisol level differ between children and adolescents with ADHD and TDC? ii) Do temperament dimensions and awakening cortisol levels correlate with dimension of ADHD core symptoms? iii) Does awakening cortisol level associate with any of the three-temperament dimension? iv) Do parent ratings of the oppositional and defiant traits and anxiety traits affect the association between temperament dimensions and cortisol levels with ADHD core symptoms in the ADHD group?

Inclusion Criteria:

1. ADHD group:

   • The assessment was completed as part of their routine psychiatric assessment using unstructured interview, screening of ADHD and other psychiatric symptoms through parent questionnaires, and direct observation with the participants, which led to a DSM-5-based diagnosis of ADHD;
   • Signed informed consent from parents or legal guardian and/or assent for youth aged >12 y/o.

2. TDY (Typically Developing Control) - group:

   • Signed informed consent from parents or legal guardian and/or assent for youth aged >12 y/o;
   • TDY were recruited from pre-school, primary and secondary schools in the local catchment area.
   • Assessment of the TDY was carried out by a clinician in schools through an unstructured interview with the parents and teachers.

Exclusion Criteria:

1. ADHD group:

   • Not signed informed consent from parents or legal guardian and/or assent for youth aged >12 y/o;
   • IQ below 70;
   • neurological and other psychiatric disorders;
   • genetic and/or medical conditions mimicking ADHD symptoms;
   • treatment with psychotropic medications other than ADHD medications;
   • TDY were also excluded if they presented with any learning or behavioural difficulties as reported in the parent and teacher interviews, as well as mild, moderate or severe symptoms of ADHD in the clinician-led rating scale.

2. TDY (Typically Developing Control) - group:

   • Not signed informed consent from parents or legal guardian and/or assent for youth aged >12 y/o;
   • IQ below 70;
   • neurological and other psychiatric disorders;
   • genetic and/or medical conditions mimicking ADHD symptoms;
   • treatment with psychotropic medications other than ADHD medications.

We enrolled 67 individuals with ADHD and 70 typically developing youths (TDY) aged between 4 and 16 years. Individuals with ADHD were recruited from the Child and Adolescent Neuropsychiatry Unit, University of Sassari, Italy.

Statistical Analyses were performed using the software R (version R 3.4.1). Analysis of variance covaring for sex (ANCOVA) was used to compare the two groups on the three temperament dimensions and on the awakening cortisol levels. A series of linear regression model analyses were used to test the association of temperament dimensions and cortisol levels with ADHD traits and the oppositional-defiant and anxiety traits. To test the difference between cortisol levels on- and off- methylphenidate treatment in participants with ADHD, we performed repeated measures analysis of variance. Before analysis, cortisol levels were log-transformed to normal. Analyses were then carried out using standardized scores for all measures.