Early-onset obsessive-compulsive disorder (OCD) is substantially less researched than adult-OCD, resulting in prevalent equivocation surrounding the neurocognitive profile of child-OCD. Research into this area is pivotal as population studies report that youths with OCD struggle significantly in academic settings. In the General Introduction of this thesis, I reviewed existing literature and found that strikingly, young patients do not show impairment on features that are considered both hallmarks of adult OCD and tightly linked to disorder symptomatology, such as response inhibition and cognitive flexibility. Among the characteristics that are thought to be present in children and adolescents with OCD are abnormal decision-making under uncertainty and impaired learning, and I decided to focus on these features as they may be driving poor academic attainment in young people with the disorder. In addition, I sought to investigate other cognitive processes that have not been well-researched in adolescent-OCD but are found to be robustly altered in adult OCD such as goal directed/model-based reasoning, meta-cognition, and feedback sensitivity. I aimed to delineate these various processes using a battery of suitably complex cognitive tasks. Moreover, I highlighted that majority of past studies fail to find differences between young patients and controls due to behavioural signatures being too subtle to be uncovered by standard statistical analyses. Hence, I employed computational modelling of cognitive task data to disentangle latent decision-making processes displayed by adolescents with OCD. In Chapter 2, I modelled data from the Wisconsin Card Sorting task, a frequently used paradigm of cognitive flexibility, and confirmed that youths with OCD show equivalent performance on the task to controls. Only patients on serotonergic medication showed increased response latencies and a tendency to make unique errors (choosing a deck associated with no rule present on the test card). Next, in Chapter 3, I sought to understand instrumental and Pavlovian learning, and whether adolescents with OCD show increased punishment sensitivity on a novel aversive Pavlovian-to Instrumental Transfer paradigm. Once again, patient performance was equivalent to that of controls. Hence, the remaining chapters were dedicated to probing behaviour on probabilistic paradigms. In Chapter 4, I formally investigated model-based and model-free learning using a well-validated two step decision-making task, and fit a reinforcement learning drift diffusion model to both choice and reaction time data. Patients showed increased exploration on the task as well as faster and more erratic decisions compared to controls. Nonetheless, model-based learning was equivalent between groups. In the penultimate chapter, I demonstrate on a predictive-inference task that patients with OCD update their choices more frequently compared to controls independent of prediction error magnitude. Finally, in Chapter 6, I administered a probabilistic reversal learning paradigm to a large sample of 50 adolescent patients and 53 matched controls. Standard analyses revealed a significant reversal learning deficit in patients with OCD, wherein they displayed more errors and a lower propensity to repeat choices following positive feedback during the post-reversal phase. Crucially, computational modelling revealed striking group differences where adolescents with OCD displayed elevated reward learning and lower punishment learning, increased exploration, and decreased perseveration compared to controls. In the General Discussion, I emphasise that atypical learning and decision-making in adolescent-OCD are more pronounced on probabilistic tasks, where task environments are more volatile. Results are partly discussed in the context of the uncertainty model of OCD, where subjective feelings of doubt experienced by patients drive compulsive behaviours such as checking and certainty-seeking in daily life, alongside excessive exploration on probabilistic tasks. I also consider various explanations for cognitive distinctions between adult- and adolescent OCD. More general implications of the findings are discussed for understanding OCD in the context of adolescent development and for treatment/support strategies.