Characterizing Underlying Cognitive Components of ADHD Presentations and Co-morbid Diagnoses: A Diffusion Decision Model Analysis

Abstract

Objective: To Explore whether subtypes and comorbidities of attention-deficit hyperactivity disorder (ADHD) induce distinct biases in cognitive components involved in information processing. Method: Performance on the Integrated Visual and Auditory Continuous Performance Test (IVA-CPT) was compared between 150 children (aged 7 to 10) with ADHD, grouped by DSM-5 presentation (ADHD-C, ADHD-I) or co-morbid diagnoses (anxiety, oppositional defiant disorder [ODD], both, neither), and 60 children without ADHD. Diffusion decision modeling decomposed performance into cognitive components. Results: Children with ADHD had poorer information integration than controls. Children with ADHD-C were more sensitive to changes in presentation modality (auditory/visual) than those with ADHD-I and controls. Above and beyond these results, children with ADHD+anxiety+ODD had larger increases in response biases when targets became frequent than children with ADHD-only or with ADHD and one comorbidity. Conclusion: ADHD presentations and comorbidities have distinct cognitive characteristics quantifiable using DDM and IVA-CPT. We discuss implications for tailored cognitive-behavioral therapy.

A diffusion decision model analysis of the cognitive effects of neurofeedback for ADHD.

“…It decomposes performance into components (quantified by model parameters) that have established psychological interpretation and that can be studied separately: general response cautiousness (boundary separation, a ), start of decision processes (starting point, z ), the tendency toward premature decisions (starting point bias, z/a ), the efficiency of integrating stimulus information (drift rate, v ), context sensitivity (drift bias, c v ), and the time for task preparation, forming a neural representation, and response execution (nondecision time component, T er ; Forstmann et al, 2016; Ratcliff, 1978; Ratcliff & McKoon, 2008). The DDM is attractive because it is a well-established model with clinical populations (Caulfield & Myers, 2018; Forstmann et al, 2016; Ging-Jehli et al, 2022; Pe et al, 2013; Weigard & Sripada, 2021; Wiecki et al, 2015; Zeguers et al, 2011), utilizes more information than conventional performance measures, and it is theoretically founded in decision theory that integrates our current understanding of brain dynamics and functioning (Cohen & Kohn, 2011; Forstmann et al, 2016; Gold & Shadlen, 2001, 2007; Hanes & Schall, 1996; Philiastides et al, 2006; Ratcliff et al, 2003; Wong et al, 2007).…”

“…Research (see for a review: Ging-Jehli et al, 2021; Weigard & Sripada, 2021) that uses DDM analyses consistently found that children with ADHD have a poorer efficiency of integrating stimulus information (lower drift rate, v ) than those without ADHD across different tasks intended to assess cognitive concepts including: cognitive flexibility (e.g., Metin et al, 2013; Salum et al, 2014); inhibitory control (e.g., Huang-Pollock et al, 2017; Mowinckel et al, 2015); selective attention (e.g., Merkt et al, 2013; Mulder et al, 2010; Weigard & Huang-Pollock, 2014); sustained attention (e.g., Huang-Pollock et al, 2012); time perception (e.g., Shapiro & Huang-Pollock, 2019); working memory (e.g., Weigard & Huang-Pollock, 2017); and reinforcement learning paradigm (e.g., Fosco et al, 2017). Moreover, recent DDM applications (Ging-Jehli et al, 2022) to the data of CPTs have shown that ADHD is not only associated with lower drift rates but also with deficient drift biases ( c v ) and longer nondecision times ( T er ). However, these differences were only detectable when considering comorbidities and DSM -defined presentations.…”

“…There is preliminary evidence (e.g., Ging-Jehli et al, 2022; Merkt et al, 2013; Salum et al, 2014) that the severity of ADHD symptoms is positively associated with lower drift rate ( v ). Studies that applied neurocognitive tests, together with a DDM analysis, examined the cognitive characteristics of ADHD by comparing performance between children with and without ADHD.…”

“…The primary purpose of this DDM analysis was to assess whether low drift rate ( v ), a purported characteristic of ADHD in prior research (Ging-Jehli et al, 2021, 2022; see also Sripada & Weigard, 2021), improves in response to NF treatment for ADHD. Specifically, the hypothesis was that phi v (introduced in the Method section: Diffusion decision model, subsection: Model parameters under examination) would improve significantly more with NF than with control treatment.…”

Race Lévy flights: A mathematically tractable framework for studying heavy-tailed accumulation noise

“…The DDM has been applied in different domains of theoretical (Forstmann et al, 2016;Ratcliff et al, 2016) and clinical (White et al, 2010;Voss et al, 2013) psychology. For example, this model has been utilized for better understanding the information processing in developmental disorders such autism spectrum disorder (Pirrone et al, 2017(Pirrone et al, , 2020, attention deficit hyperactivity disorder (Pedersen et al, 2017;Ging-Jehli et al, 2021;Iuculano et al, 2020;Nejati et al, 2022), and dyslexia (Zeguers et al, 2011;Manning et al, 2021Manning et al, , 2022. In addition, the DDM has shed light on the underlying mechanisms of different areas of decision-making, such as perceptual (Ratcliff and McKoon, 2008), value-based (Fontanesi et al, 2019a;Pedersen et al, 2017;Fontanesi et al, 2019b), and lexical (Ratcliff et al, 2004a;Wagenmakers et al, 2008;Tillman et al, 2017a).…”