Data related to brain function may have the potential to improve the reliability and validity of assessments for the aetiologically and clinically heterogeneous syndrome of attention-deficit/hyperactivity disorder (ADHD). This study investigated associations between questionnaire assessments of behavioural features of adults with ADHD and an aspect of neurocognitive performance which has been reported to be impaired in adults with ADHD.

Fifty-nine adult patients with a DSM-IV diagnosis of ADHD, and their informants, completed questionnaires related to aspects of severity of ADHD. Associations were examined between questionnaire ratings and performance on a computer-administered task of spatial working memory (SWM).

Correlations between ratings of ADHD and SWM indicated moderate but significant correlations for patients' ratings, but not for informants' ratings. Also, patients who reported a past history of ‘self-harm’ (N = 33) had a significantly worse mean performance on both measures of SWM (p = 0.004, 0.003).

The results indicate that aspects of impulsivity, i.e. self-ratings of ‘emotive’ behaviour (involving rapid response to stimuli and marked reactivity of mood) and of past ‘self-harm’, show relatively strong associations with SWM performance in adults selected on the basis of an ADHD diagnosis. A profile of neurocognitive performances may have a role in the assessment of ADHD.

Childhood-onset attention-deficit hyperactivity disorder (ADHD) in adults is clinically heterogeneous and commonly presents with different patterns of cognitive deficits. It is unclear if this clinical heterogeneity expresses a dimensional or categorical difference in ADHD.

We first studied differences in functional connectivity in multi-echo resting-state functional magnetic resonance imaging (rs-fMRI) acquired from 80 medication-naïve adults with ADHD and 123 matched healthy controls. We then used canonical correlation analysis (CCA) to identify latent relationships between symptoms and patterns of altered functional connectivity (dimensional biotype) in patients. Clustering methods were implemented to test if the individual associations between resting-state brain connectivity and symptoms reflected a non-overlapping categorical biotype.

Adults with ADHD showed stronger functional connectivity compared to healthy controls, predominantly between the default-mode, cingulo-opercular and subcortical networks. CCA identified a single mode of brain–symptom co-variation, corresponding to an ADHD dimensional biotype. This dimensional biotype is characterized by a unique combination of altered connectivity correlating with symptoms of hyperactivity-impulsivity, inattention, and intelligence. Clustering analyses did not support the existence of distinct categorical biotypes of adult ADHD.

Overall, our data advance a novel finding that the reduced functional segregation between default-mode and cognitive control networks supports a clinically important dimensional biotype of childhood-onset adult ADHD. Despite the heterogeneity of its presentation, our work suggests that childhood-onset adult ADHD is a single disorder characterized by dimensional brain–symptom mediators.