Artificial intelligence (AI) is increasingly being explored and adopted across the research lifecycle, from idea generation and literature discovery to data analysis, manuscript preparation, and editorial and peer-review processes. This perspective provides an overview of AI's role across the stages of scientific research. We describe emerging tools and workflows, illustrating how AI can assist researchers by aggregating and synthesizing a large body of work across various domains, supporting methodological implementation, and facilitating communication and publication. We also discuss their shortcomings, including surface-level reasoning, the fabrication of plausible but incorrect outputs, and the challenges posed by the fact that researchers new to a field may not know which questions to ask or which nuances to interrogate. In addition, we discuss recent advances toward more agentic and end-to-end AI systems, highlighting both their technical feasibility and the challenges they pose for validation, oversight, and responsible use. For each stage of the research lifecycle, we outline key limitations of current AI systems and propose practical considerations, what researchers should and should not do to support rigorous and ethical integration of AI into scientific workflows. This integration requires coordinated frameworks across the ecosystem. Journals, funding agencies, universities, and policymakers play essential roles in defining standards for transparency and accountability, while individual researchers remain responsible for methodological rigor and validity of reported results.

Recent evidence suggests the brain’s major excitatory neurotransmitter, glutamate, plays a key role in attention-deficit/hyperactivity disorder (ADHD). Here we ask if glutamate also plays a role in the variable clinical courses of ADHD. While some children ‘grow out’ of ADHD by adolescence, others experience persistent symptoms into adulthood. Prior work implicates structural and functional differences in medial prefrontal cortex as pivotal in these different ADHD symptom courses, and we now ask if glutamate developmental change also contributes. Given the role of glutamate in neurotransmission, we also investigate potential impacts on the brain’s intrinsic connectivity. Using a glutamate-specific magnetic resonance spectroscopy sequence at 3 T, we analyzed 241 spectra on 161 participants, including 69 with persistent ADHD, 20 with remitting ADHD, and 72 never affected controls. Intrinsic functional connectivity was also assessed in a subset of 104 participants with 141 functional MRI scans. Using linear mixed models, we found an age-related increase in medial prefrontal glutamate in the persistent ADHD group which differed significantly from an age-related decrease among those who remitted and the never affected controls. Furthermore, altered prefrontal glutamate concentrations were associated with changes in intrinsic connectivity between the default mode network (which includes medial frontal cortex) and subcortical regions. These findings may indicate altered maturation of glutamate in the medial prefrontal cortex in youth with persistent ADHD.

Both brain and behavior undergo considerable changes during adolescence, a time synonymous with the emergence several psychiatric disorders, including mood and psychotic disorders. Multiple studies have explored the interaction between brain and behavior during childhood and adolescence, with both mixed and convergent findings between the studies. One challenge is that different longitudinal models can be applied to both the same or similar dataset and yet yield different results. Without knowing the ground truth, the best modeling approach to understanding the brain/behavior interplay during development remains unclear. Within this context, simulated datasets can provide a mechanism to test many different neurodevelopmental features, not only regarding the interplay between brain and behavior, but also how specific epidemiologic factors can influence study findings. Simulated datasets have the ‘ground truth’ relationships embedded within the models used to create these datasets. Thus, the goal of this study was to assess the epidemiologic and clinical characteristics of 15 large simulated longitudinal neuroimaging datasets, independently created from five research groups from three different continents. We found that while most of the neuroimaging and behavioral trajectories in the simulated datasets reflected growth patterns identified in the literature, there was also considerable variability between how the groups modeled the epidemiologic, clinical, and neurodevelopmental growth patterns. The conclusion is that while simulation studies are an underused strategy to test assumptions related to neurodevelopmental disorders, simulation studies require iterative approaches, testing simulated against real-world data and challenging the different assumptions associated with the neurodevelopmental models, until there is convergence between actual and simulated datasets.

Objective:The association between family environment factors, such as family conflict and parental monitoring, and attention-deficit/hyperactivity disorder (ADHD) behaviors is well established. However, it is unclear whether ADHD behaviors precede changes in familial functioning, whether certain family dynamics increase ADHD behaviors, or if the association is reciprocal. Method:Data from 7,425 children across 4 waves of the Adolescent Brain and Cognitive Development Study were analyzed. ADHD behaviors were evaluated using the Child Behavior Checklist. Family conflict and parental monitoring were assessed using the Family Environment Scale and the Parental Monitoring Survey, respectively. Random-intercept cross-lagged panel modeling was employed to explore potential within-subject temporal associations between ADHD behaviors and the family environment. Results:At the within-subject level, increases in ADHD behaviors were associated with subsequent increases in family conflict (B = .02, 95% CI [0.002, 0.04], p = .03, adjusted p = .04) and reductions in parental monitoring (B = -.006, 95% CI [-0.01, -0.0006], p = .03, adjusted p = .04). Similar findings emerged when restricting analyses to 1,315 participants meeting diagnostic criteria for ADHD. These findings were not moderated by sex, race/ethnicity, household income, or medication status. Among participants with ADHD, cross-lagged associations involving ADHD behaviors remained significant after controlling for time-varying externalizing and internalizing behaviors, which did not show significant cross-lagged associations with family conflict or parental monitoring. Conclusion:These findings indicate that worsening ADHD behaviors can negatively impact the family environment. Despite small effect sizes, this study bolsters the case for early detection and treatment to help young people with ADHD, potentially mitigating impact on family dynamics.

Although psychostimulant treatment for attention-deficit/hyperactivity disorder (ADHD) is often effective, the medication does not work for all and some experience adverse side effects. As a result, a desire for effective non-pharmacological options emerges as a top priority for those with lived experience. One such approach is cognitive training focused on the neurocognitive differences associated with ADHD. Its results have been mixed, partly as it is delivered using computers, and sustained engagement with this format is difficult for children with attentional difficulties. To investigate further, we developed a cognitive training tool delivered in a gamified, fully virtual reality (VR) platform (Floreo TM), and here report on an open-label pilot trial of its feasibility, tolerability, and preliminary outcomes. Thirty youth with ADHD (mean age = 12.5 years; 33% female) were enrolled and completed up to 20 VR training sessions over a 4- to 6-week period. The intervention targeted four key cognitive domains implicated in ADHD: impulse control, working memory, temporal processing and processing speed. Clinician-rated inattention symptoms and associated functional impairments, based on interviews with parents, showed significant pre–post improvements (symptoms: t(25.90) = -5.60, p < .001; d = -1.05; impairment: t(25.68) = -3.29, p = .003; d = -0.62). The intervention was well tolerated, with only minor adverse events such as occasional headaches and brief frustration. Usability ratings were high for parents and moderate for youth. Significant near-transfer effects were observed on tasks of impulse control (t (23.49) = -2.86, p = .009; d = -0.56) and processing speed (t(24.06) = 3.40, p = .002; d = 0.66). These findings suggest that VR-based cognitive training is a feasible, acceptable, and potentially effective approach for children with ADHD. Future randomized controlled trials are warranted to confirm efficacy and examine durability of effects.

Autism Spectrum Disorder (autism) and Attention Deficit/Hyperactivity Disorder (ADHD) often co-occur, although it remains unclear if these conditions share common neurobiological foundations or exhibit distinct alterations in resting-state brain connectivity. We conducted a cross-sectional mega-analytic comparison of functional connectivity patterns linked to autism and ADHD traits in children and adolescents (ages 6-19 years; n=10,168), with follow-up analyses considering autism (n=764 autistic; n=893 neurotypical) and ADHD diagnoses (n=2,026 ADHD; n=2,409 neurotypical). In total, 12,732 unique child and adolescent participants were included: 3,528 in both analyses, 6,640 in the trait analysis only, and 2,564 in the diagnostic analysis only. Autism traits and diagnosis were associated with reduced connectivity between the thalamus, putamen, salience/ventral attention, and frontoparietal networks, while ADHD traits showed the opposite pattern. Hyperconnectivity between the default mode and dorsal attention networks was observed in both autistic and ADHD groups relative to neurotypical individuals and associated with ADHD traits. Despite frequent co-occurrence, autism and ADHD traits exhibit distinct neural signatures, with small effect sizes indicating subtle associations.

OBJECTIVE:A large body of functional MRI research has examined a potential role for subcortico-cortical loops in the pathogenesis of attention deficit hyperactivity disorder (ADHD), but has produced inconsistent findings. The authors performed a mega-analysis of six neuroimaging data sets to examine associations between ADHD diagnosis and traits and subcortico-cortical connectivity. METHODS:Group differences were examined in the functional connectivity of four subcortical seeds in 1,696 youths with ADHD diagnoses (66.39% males; mean age, 10.83 years [SD=2.17]) and 6,737 unaffected control subjects (47.05% males; mean age, 10.33 years [SD=1.30]). The authors examined associations between functional connectivity and ADHD traits (total N=9,890; 50.3% males; mean age, 10.77 years [SD=1.96]). Sensitivity analyses were used to examine specificity relative to commonly comorbid internalizing and non-ADHD externalizing problems. The authors further examined results within motion-matched subsamples, and after adjusting for estimated intelligence. RESULTS:In the group comparison, youths with ADHD showed greater connectivity between striatal seeds and temporal, fronto-insular, and supplementary motor regions, as well as between the amygdala and dorsal anterior cingulate cortex, compared with control subjects. Similar findings emerged when ADHD traits were considered and when alternative seed definitions were adopted. Dominant associations centered on the connectivity of the caudate bilaterally. Findings were not driven by in-scanner motion and were not shared with commonly comorbid internalizing and externalizing problems. Effect sizes were small (largest peak d, 0.15). CONCLUSIONS:The findings from this large-scale mega-analysis support established links with subcortico-cortical circuits, which were robust to potential confounders. However, effect sizes were small, and it seems likely that resting-state subcortico-cortical connectivity can capture only a fraction of the complex pathophysiology of ADHD.

In this chapter, we review scientific findings that form the basis for neuroimaging and neurophysiological biomarkers for ADHD diagnosis and treatment. We then highlight the different challenges in translating mechanistic findings into biomarkers for ADHD diagnosis and treatment. Population heterogeneity is a primary barrier for identifying biomarkers of ADHD diagnosis, which requires shifts toward dimensional approaches that identify clinically useful subgroups or prospective biomarkers that can identify trajectories of illness, function, or treatment response. Methodological limitations, including emphasis on group level analyses of treatment effects in small sample sizes, are the primary barriers to biomarker discovery in ADHD treatment. Modifications to clinical trials, including shifting towards testing biomarkers of a priori prediction of functionally related brain targets, treatment response, and side effects, are suggested. Finally, future directions for biomarker work are discussed.

Objective: Though exposure and response prevention (ERP) is a well-proven treatment for OCD across the lifespan, prior RCTs have not studied adolescent and adult patients with the same ERP protocol relative to an active comparator that controls for non-specific effects of treatment. This approach assesses differences in the effect of OCD-specific exposures in affected adolescents and adults and in response to ERP compared to a stress-management control therapy (SMT).Methods: This assessor-blinded, parallel, 2-arm, randomized, ambulatory clinical superiority trial randomized adolescents (aged 12-18) and adults (24-46) with OCD (N = 126) to 12 weekly sessions of ERP or SMT. OCD severity was measured before, during and after treatment using the child or adult version of the Yale-Brown Obsessive Compulsive Scale (C/Y-BOCS), depending on participant age. We predicted that ERP would produce greater improvement in OCD symptoms than SMT and that there would be no significant post-treatment differences across age groups.Results: ERP (n = 63) produced significantly greater improvements on C/Y-BOCS scores at post-treatment than SMT (n = 63) (Effect size = -0.72, CI = -0.52 to -0.91, p < .001). ERP also produced more treatment responders (ERP = 86%, SMT = 32%; chi(2) = 46.37, p < .001) and remitters than SMT (ERP = 39%, SMT = 7%; chi(2) = 16.14, p < .001). Finally, there were no statistically significant post-treatment differences in C/Y-BOCS scores between adolescents and adults assigned to ERP.Conclusion: A single ERP protocol is superior to SMT in treating both adolescents and adults with OCD. OCD-specific therapy is necessary across the lifespan for optimal outcomes in this highly disabling disorder, though non-specific treatments like SMT are still all-too-commonly provided.

ADHD is a highly heritable childhood disorder, with estimated h2 of 0.7. To characterize the genetic variants contributing to this heritability, we used family-based designs, studying families with many members affected by ADHD. Genome-wide SNP array data were obtained on two family cohorts: (1) NHGRI family cohort (359 nuclear families,1538 individuals); (2) NCR family cohort (25 extended and 132 nuclear families, 631 members). Family-based association tests (FBAT) were used to identify variants shared across affected individuals in all families, while linkage analyses were used to identify shared haplotypes and variants with potentially larger effects. Using neuroimaging data from the NCR families, we asked if genes associated with familial ADHD overlapped with genes associated with neural features. A meta-analysis of FBAT identified three genome-wide significant SNP associations that lay upstream from TFRC, intronic to TRIM31 and intronic to DFNA5. TRIM31 has been found associated to intelligence in individuals with ADHD, while variants in DFNA5 have been identified as pleiotropic loci associated with multiple psychiatric disorders. Genes associated with ADHD (at p2) and one significant linkage region (LOD > 3) on 19p13.2-13.11. The linkage region on 19p encompassed a genome-wide significant variant found in the FBAT of the NHGRI families (rs55741253, p=2.68 × 10-8). Additionally, two linkage regions that reached a LOD > 2 in our meta-analysis replicated significant linkages from prior studies. Genes associated with a feature of the brain's functional connectivity, capturing the interaction between the brain's default mode and task positive network, overlapped with genes associated with ADHD (p=0.04). Using family-based association methods, we report significant associations between three genes and familial ADHD, with one signal lying within a region of significant linkage.

BACKGROUND: While attention-deficit/hyperactivity disorder (ADHD) has been associated with differences in the structural connections formed by the brain's white matter tracts, studies of such differences have yielded inconsistent findings, likely reflecting small sample sizes. Thus, we conducted a mega-analysis on in vivo measures of white matter microstructure obtained through diffusion tensor imaging of more than 6000 participants from 5 cohorts. METHODS: In a mega-analysis, linear mixed models were used to test for associations between the fractional anisotropy of 42 white matter tracts and ADHD traits and diagnosis. Contrasts were made against measures of mood, anxiety, and other externalizing problems. RESULTS: Overall, 6993 participants (ages 6-18 years, mean age 10.62 years [SD 1.99]; 3368 girls, 3625 boys; 764 African American, 4146 non-Hispanic White, and 2083 other race/ethnicities) had measures of ADHD and other emotional/behavioral symptoms (N = 6933) and/or enough clinical data to allow a diagnosis of ADHD (n = 951) or its absence (n = 4884). Both the diagnosis and symptoms of ADHD were associated with lower fractional anisotropy of the inferior longitudinal and left uncinate fasciculi (at a false discovery rate-adjusted p , .05). Associated effect sizes were small (the strongest association with ADHD traits had an effect size of partial r =-0.14, while the largest case-control difference was associated with an effect size of d =-0.3). Similar microstructural anomalies were not present for anxiety, mood, or externalizing problems. Findings held when ADHD cases and control subjects were matched on in-scanner motion. CONCLUSIONS: While present across cohorts, ADHD-associated microstructural differences had small effects, underscoring the limited clinical utility of this imaging modality used in isolation.

Background Attention deficit/hyperactivity disorder (ADHD) is usually conceptualized as a childhood-onset neurodevelopmental disorder, in which symptoms either decrease steadily into adulthood or remain stable. A recent study challenged this view, reporting that for most with ADHD, diagnostic status fluctuates with age. We ask if such a 'fluctuating' ADHD symptom trajectory subgroup is present in other population-based and clinic based cohorts, centered on childhood and adolescence.Methods Cohorts were the population-based Adolescent Brain Cognitive Development (ABCD: N = 9735), Neurobehavioral Clinical Research (NCR: N = 258), and the Nathan Kline Institute-Rockland (NKI-Rockland: N = 149). All participants had three or more assessments spanning different age windows. Participants were categorized into developmental diagnostic subgroups: fluctuant ADHD (defined by two or more switches between meeting and not meeting ADHD criteria), remitting ADHD, persisting ADHD, emerging ADHD and never affected. Data were collected between 2011 and 2022. Analyses were performed between May 2022 and April 2023. Findings A subgroup with fluctuant child and adolescent ADHD diagnoses was found in all cohorts (29.3% of participants with ADHD in ABCD, 26.6% in NCR and 17% in NKI-Rockland). While the proportion of those with fluctuant ADHD increased with the number of assessments, it never constituted the dominant subgroup. Interpretation We provide further evidence in three cohorts for the existence of a fluctuant ADHD diagnostic subgroup during childhood and adolescence, albeit in a minority of cases. Such fluctuant child and adolescent ADHD diagnoses may suggest a natural history more akin to relapsing-remitting mood disorders and/or a marked sensitivity to environmental shifts that occur across development.

Data on the prevalence of childhood OCD is scarce, with estimates based on unrepresentative, local samples, or clinical registries. Furthermore, while obsessive and compulsive symptoms (OCS) are believed to be common, estimates have varied between 8% and 72%. Finally, while the stability of OCD from childhood to adolescence has been estimated at 41%, this estimate was based on OCD cases treated at specialist clinics, and rates in the general population are unknown. Therefore, we examined the stability rates and presence of childhood OCD and OCS using data from a US-based cohort study.

While childhood attention-deficit/hyperactivity disorder (ADHD) is more prevalent in males than females, genetic contributors to this effect have not been established. Here, we explore sex differences in the contribution of common and/or rare genetic variants to ADHD. Participants were from the Adolescent Brain and Cognitive Development study ( N = 1253 youth meeting DSM-5 criteria for ADHD [mean age = 11.46 years [SD = 0.87]; 31% female] and 5577 unaffected individuals [mean age = 11.42 years [SD = 0.89]; 50% female], overall 66% White, non-Hispanic (WNH), 19% Black/African American, and 15% other races. Logistic regression tested for interactions between sex (defined genotypically) and both rare copy number variants (CNV) and polygenic (common variant) risk in association with ADHD. There was a significant interaction between sex and the presence of a CNV deletion larger than 200 kb, both in the entire cohort (β = −0.74, CI = [−1.27 to −0.20], FDR-corrected p = 0.048) and, at nominal significance levels in the WNH ancestry subcohort (β = −0.86, CI = [−1.51 to −0.20], p = 0.010). Additionally, the number of deleted genes interacted with sex in association with ADHD (whole cohort. β = −0.13, CI = [−0.23 to −0.029], FDR-corrected p = 0.048; WNH. β = −0.17, CI = [−0.29 to −0.050], FDR-corrected p = 0.044) as did the total length of CNV deletions (whole cohort. β = −0.12, CI = [−0.19 to −0.044], FDR-corrected p = 0.028; WNH. β = −0.17, CI = [−0.28 to −0.061], FDR-corrected p = 0.034). This sex effect was driven by increased odds of childhood ADHD for females but not males in the presence of CNV deletions. No similar sex effect was found for CNV duplications or polygenic risk scores. The association between CNV deletions and ADHD was partially mediated by measures of cognitive flexibility. In summary, CNV deletions were associated with increased odds for childhood ADHD in females, but not males.

OBJECTIVECortical-subcortical hyperconnectivity related to affective-behavioral integration and cortical network hypoconnectivity related to cognitive control have been demonstrated in obsessive-compulsive disorder (OCD); the study objective was to examine whether these connectivity patterns predict treatment response.METHODSAdolescents (ages 12-17) and adults (ages 24-45) were randomly assigned to 12 sessions of exposure and response prevention (ERP) or stress management therapy (SMT), an active control. Before treatment, resting-state connectivity of ventromedial prefrontal cortical (vmPFC), cingulo-opercular, frontoparietal, and subcortical regions was assessed with functional MRI. OCD severity was assessed with the Yale-Brown Obsessive Compulsive Scale before, during, and after treatment. Usable fMRI and longitudinal symptom data were obtained from 116 patients (68 female; 54 adolescents; 60 medicated).RESULTSERP produced greater decreases in symptom scores than SMT. ERP was selectively associated with less vmPFC-subcortical (caudate and thalamus) connectivity in both age groups and primarily in unmedicated participants. Greater symptom improvement with both ERP and SMT was associated with greater cognitive-control (cingulo-opercular and frontoparietal) and subcortical (putamen) connectivity across age groups. Developmental specificity was observed across ERP and SMT treatments, such that greater improvements with ERP than SMT were associated with greater frontoparietal-subcortical (nucleus accumbens) connectivity in adolescents but greater connectivity between frontoparietal regions in adults. Comparison of response-predictive connections revealed no significant differences compared with a matched healthy control group.CONCLUSIONSThe results suggest that less vmPFC-subcortical connectivity related to affect-influenced behavior may be important for ERP engagement, whereas greater cognitive-control and motor circuit connectivity may generally facilitate response to psychotherapy. Finally, neural predictors of treatment response may differ by age.