ADHD classification based on a multi-objective support vector machine

ADHD (attention deficit hyperactivity disorder) is a neurodevelopmental disorder characterized by harmful levels of inattention, disorganization, and/or hyperactivity-impulsivity. In childhood, these symptoms often overlap with those of other disorders, and they tend to persist into adulthood, interfering with relationships and academic and work life. Diagnosis, traditionally made by assessing the patient, i.e., testing and listening to relatives and teachers, has already been aided by neuroimaging. However, the visual analysis of such images to make a psychiatric diagnosis is a complex and sometimes time-consuming task. For this reason, computer-aided diagnostic tools have increasingly evolved that, when combined with machine learning (ML) techniques, can accelerate, facilitate, and maximize the accuracy of diagnoses. Nevertheless, research evaluating ML models for classifying ADHD considering severity using images of the brain SPECT (Single Photon Emission Computed Tomography) is still very sparse. For this reason, this article aims to evaluate the performance of the ML methods: k-NN (k-Nearest Neighbors), Naive Bayes, Decision Tree, MLP (Multilayer Perceptron) and SVM (Support Vector Machine) in the classification of ADHD. The main goal of this analysis is to check whether the subjects have the disorder or not, and to classify the severity of those who have it using SPECT images. A database was created from SPECT images and diagnostic reports. After pre-processing these data, the best hyperparameters for the ML methods were searched, trained/tested and finally statistically compared. The best results were obtained with SVM and k-NN, with 98% accuracy. Although ADHD diagnosis by neuroimaging is not yet a standard clinical procedure, we argue that this study can contribute to ADHD diagnosis research and support methods for the development of CAD (computer-aided diagnosis) systems.

Attention deficit hyperactivity disorder (ADHD) is one of the most common psychiatric disorders in school-aged children. Its accurate diagnosis looks after patients’ interests well with effective treatment, which is important to them and their family. Resting-state functional magnetic resonance imaging (rsfMRI) has been widely used to characterize the abnormal brain function by computing the voxel-wise measures and Pearson’s correlation (PC)-based functional connectivity (FC) for ADHD diagnosis. However, exploring the powerful measures of rsfMRI to improve ADHD diagnosis remains a particular challenge. To this end, this paper proposes an automated ADHD classification framework by fusion of multiple measures of rsfMRI in adolescent brain. First, we extract the voxel-wise measures and ROI-wise time series from the brain regions of rsfMRI after preprocessing. Then, to extract the multiple functional connectivities, we compute the PC-derived FCs including the topographical information-based high-order FC (tHOFC) and dynamics-based high-order FC (dHOFC), the sparse representation (SR)-derived FCs including the group SR (GSR), the strength and similarity guided GSR (SSGSR), and sparse low-rank (SLR). Finally, these measures are combined with multiple kernel learning (MKL) model for ADHD classification. The proposed method is applied to the Adolescent Brain and Cognitive Development (ABCD) dataset. The results show that the FCs of dHOFC and SLR perform better than the others. Fusing multiple measures achieves the best classification performance (AUC = 0.740, accuracy = 0.6916), superior to those from the single measure and the previous studies. We have identified the most discriminative FCs and brain regions for ADHD diagnosis, which are consistent with those of published literature.

BackgroundEffective and accurate diagnosis of attention-deficit/hyperactivity disorder (ADHD) is currently of significant interest. ADHD has been associated with multiple cortical features from structural MRI data. However, most existing learning algorithms for ADHD identification contain obvious defects, such as time-consuming training, parameters selection, etc. The aims of this study were as follows: (1) Propose an ADHD classification model using the extreme learning machine (ELM) algorithm for automatic, efficient and objective clinical ADHD diagnosis. (2) Assess the computational efficiency and the effect of sample size on both ELM and support vector machine (SVM) methods and analyze which brain segments are involved in ADHD.MethodsHigh-resolution three-dimensional MR images were acquired from 55 ADHD subjects and 55 healthy controls. Multiple brain measures (cortical thickness, etc.) were calculated using a fully automated procedure in the FreeSurfer software package. In total, 340 cortical features were automatically extracted from 68 brain segments with 5 basic cortical features. F-score and SFS methods were adopted to select the optimal features for ADHD classification. Both ELM and SVM were evaluated for classification accuracy using leave-one-out cross-validation.ResultsWe achieved ADHD prediction accuracies of 90.18% for ELM using eleven combined features, 84.73% for SVM-Linear and 86.55% for SVM-RBF. Our results show that ELM has better computational efficiency and is more robust as sample size changes than is SVM for ADHD classification. The most pronounced differences between ADHD and healthy subjects were observed in the frontal lobe, temporal lobe, occipital lobe and insular.ConclusionOur ELM-based algorithm for ADHD diagnosis performs considerably better than the traditional SVM algorithm. This result suggests that ELM may be used for the clinical diagnosis of ADHD and the investigation of different brain diseases.

Background Lower parental education is linked to higher incidence of attention-deficit hyperactivity disorder (ADHD) diagnosis in offspring, as is having ADHD in family, but these associations are rarely studied together. We examined whether parental education predicts offspring ADHD diagnosis differently for children and adolescents with or without ADHD family members. Methods We used Finnish total population register-based data on individuals born between 1994-2003 with non-migrant background. We followed 228,247 boys and 217,866 girls for first ADHD diagnosis, measured from health care or medication registers, in childhood (4-12) and adolescence (13-17), and identified ADHD for biological parents or full siblings. We used Poisson regression to examine differences in diagnoses separately by offspring sex and age group. Results Preliminary results show that boys are more often diagnosed for ADHD in childhood and girls in adolescence, but incidence is higher for boys in both age groups. Among boys and girls without ADHD family members, lower parental education is associated with higher incidence in both childhood and adolescence. Among those with ADHD family members, lower parental education predicts higher incidence in childhood, but in adolescence the differences by parental education even out for boys, and for girls, higher parental education predicts higher incidence. Conclusions While lower parental education predicts higher ADHD diagnosis incidence in general, for adolescents in ADHD families, higher parental education seems to increase the likelihood of getting diagnosed -particularly for girls. This suggests that adolescents in ADHD families with lower parental education may remain without ADHD diagnosis if they are not diagnosed as children. Support measures should be developed to ensure equal access to diagnosis and care irrespective of age, gender and family background. Key messages • Lower parental education predicts higher ADHD diagnosis incidence in general. Yet, adolescents in ADHD families with lower educated parents may remain underdiagnosed. • Support measures should be developed to guarantee equal access to ADHD diagnosis and care regardless of family background.

ObjectivesThere is an ongoing debate about the restrictive inattentive (RI) presentation of attention deficit hyperactivity disorder (ADHD). The current study aimed to systematically investigate the clinical, neuropsychological, and brain functional characteristics of children with ADHD restrictive inattentive presentation.MethodsA clinical sample of 789 children with or without ADHD participated in the current study and finished clinical interviews, questionnaires, and neuropsychological tests. Those individuals with a diagnosis of ADHD were further divided into three subgroups according to the presentation of inattentive and/or hyperactive/impulsive symptoms, the ADHD-RI, the ADHD-I (inattentive), and the ADHD-C (combined) groups. Between-group comparisons were carried out on each clinical and neuropsychological measure using ANCOVA, with age and sex as covariates. Bonferroni corrections were applied to correct for multiple comparisons. Two hundred twenty-seven of the subjects also went through resting-state functional magnetic resonance imaging scans. Five ADHD-related brain functional networks, including the default mode network (DMN), the dorsal attention network (DAN), the ventral attention network, the executive control network, and the salience network, were built using predefined regions of interest (ROIs). Voxel-based group-wise comparisons were performed.ResultsCompared with healthy controls, all ADHD groups presented more clinical problems and weaker cognitive function. Among the ADHD groups, the ADHD-C group had the most clinical problems, especially delinquent and aggressive behaviors. Regarding cognitive function, the ADHD-RI group displayed the most impaired sustained attention, and the ADHD-C group had the worst response inhibition function. In terms of brain functional connectivity (FC), reduced FC in the DMN was identified in the ADHD-C and the ADHD-I groups but not the ADHD-RI group, compared to the healthy controls. Subjects with ADHD-I also presented decreased FC in the DAN in contrast to the control group. The ADHD-RI displayed marginally significantly lower FC in the salience network compared to the ADHD-I and the control groups.ConclusionThe ADHD-RI group is distinguishable from the ADHD-I and the ADHD-C groups. It is characterized by fewer externalizing behaviors, worse sustained attention, and better response inhibition function. The absence of abnormally high hyperactive/impulsive symptoms in ADHD-RI might be related to less impaired brain function in DMN, but potentially more impairment in the salience network.

Lifelong greenspace exposure and ADHD in Polish children: Role of physical activity and perceived neighbourhood characteristics

INTRODUCTION: Attention Deficit Hyperactivity Disorder (ADHD) and previous history of concussion (PHC) are being reported more frequently in college athletes; the presence of these co-morbidities and their effects on concussion assessment has been somewhat limited. PURPOSE: Determine if a relationship exists between ADHD and PHC and the effect of these co-morbidities on neuropsychological test scores. METHODS: This study was a retrospective cross-sectional between groups design. D-1 college athletes (N=1840, males N=1737, females N=103) with a mean age of 19.8+1.6 years completed a computerized neuropsychological test (ImPACT) to establish baseline assessment of concussion history during pre-participation physicals. Participants were divided into groups based upon self-reported diagnosis of ADHD and PHC. Self-report symptoms and ImPACT composite scores served as dependent variables. Comparisons of groups were calculated utilizing regression models and One-Way ANOVAs. All statistical analyses were conducted utilizing STATA 9.0. RESULTS: 102 (5.5%) participants self-reported a diagnosis of ADHD. 486 (26.5%) of the participants self-reported a history of concussion. There was a difference between ADHD and controls for self-reported symptoms (F (1, 1838) = 26.226, p > 0.001, Composite Verbal Memory (F (1, 1838) = 10.02, p = 0.002), Composite Motor (F (1, 1838) = 5.96, p = 0.015), Composite reaction time (F (1, 1838) = 3.94, p = 0.047). No statistical differences were found between the PHC group and normal counterparts. There was a difference for the interaction between ADHD and PHC on Composite Verbal Memory (F (5, 1834) = 2.44, p= 0.0325). CONCLUSIONS: The increase of media and clinician awareness of PHC and ADHD diagnoses has increased reporting during baseline assessment. Significant differences existed between groups with participants with ADHD and participants with ADHD and PHC performing below their "normal" counterparts. Previous literature supports the potential additive effect of learning disabilities and PHC; however this subgroup of ADHD did not replicate previous findings. The increase of ADHD and PHC diagnoses demands that health care practitioners understand the implications and effects of these disabilities on neuropsychological tests utilized in concussion assessment.

Attention Deficit Hyperactivity Disorder (ADHD) is an early onset clinical condition characterized by attention difficulties, hyperactivity and impulsivity which can persist across the lifespan, significantly influencing the evolutionary course and facilitating the rise of psychiatric comorbidities. The presence of different ADHD subtypes in adults is a heterogeneity factor to be recognized in order to orient prognosis and treatment, as indicated by studies that described differences in the characterization of different subtypes in relation to both severity and comorbidities. In the present study we evaluated the socio-demographic and clinical characteristics of a sample of adults with ADHD and the characteristics associated with the different disorder subtypes. We described 60 patients aged between 18 and 65 years (mean age 34.1) with primary diagnosis of ADHD consecutively admitted to the Regional Centre for diagnosis and treatment of ADHD in adults in Milan. We observed high severity of symptoms and low quality of life, in particular in the "life outlook" dimension. The subtypes distribution was the following: 18.3% inattentive subtype, 8.3% hyperactive/impulsive subtype and 70% combined subtype. The hyperactive/impulsive subtype showed a significantly higher frequency in females, while the inattentive subtype was more frequent in males. Patients with the hyperactive/impulsive subtype showed worse quality of life and more frequent anxiety disorders. Considering the different clinical profiles among various subtypes, these data add relevance to subtypes classification of adult ADHD.

Attention deficit hyperactivity disorder (ADHD) is a multifactorial neurodevelopmental disorder, yet the interplay between ADHD polygenic risk scores (PRSs) and other risk factors remains relatively unexplored. The authors investigated associations, confounding, and interactions of ADHD PRS with birth-related, somatic, and psychosocial factors previously associated with ADHD. Participants included a random general population sample (N=21,578) and individuals diagnosed with ADHD (N=13,697) from the genotyped Danish iPSYCH2012 case cohort, born between 1981 and 2005. The authors derived ADHD PRSs and identified 24 factors previously associated with ADHD using national registers. Logistic regression was used to estimate associations of ADHD PRS with each risk factor in the general population. Cox models were used to evaluate confounding of risk factor associations with ADHD diagnosis by ADHD PRS and parental psychiatric history, and interactions between ADHD PRS and each risk factor. ADHD PRS was associated with 12 of 24 risk factors (odds ratio range, 1.03-1.30), namely, small gestational age, infections, traumatic brain injury, and most psychosocial risk factors. Nineteen risk factors were associated with ADHD diagnosis (odds ratio range, 1.20-3.68), and adjusting for ADHD PRS and parental psychiatric history led to only minor attenuations. Only the interaction between ADHD PRS and maternal autoimmune disease survived correction for multiple testing. Higher ADHD PRS in the general population is associated with small increases in risk for certain birth-related and somatic ADHD risk factors, and broadly to psychosocial adversity. Evidence of gene-environment interaction was limited, as was confounding by ADHD PRS and family psychiatric history on ADHD risk factor associations. This suggests that the majority of the investigated ADHD risk factors act largely independently of current ADHD PRS to increase risk of ADHD.

Between 15% and 35% of children and adolescents with attention deficit hyperactivity disorder (ADHD) have learning disabilities or behavior disorders as coexisting conditions.1 Stimulant medication alone will fail to produce optimal improvement at home and school unless combined with academic and behavioral therapies. Parents as well as pediatricians may lack the expertise necessary to successfully advocate for the ADHD child or patient to receive psychological or academic testing and appropriate classroom accommodations. Despite their legal eligibility based on Section 504 of the Rehabilitation Act of 1973 or the Individuals with Disabilities Education Act (IDEA), students with ADHD may not receive the educational services essential for them to achieve their potential. Earlier studies have addressed this issue by describing methods to improve interaction between physicians and schools.2 Subsequent research has suggested an extensive multidisciplinary process to develop community consensus for the accurate diagnosis and comprehensive treatment of ADHD.3 The recent American Academy of Pediatrics (AAP) classification of ADHD as a chronic condition recommends that the medical home guiding principle of family-centeredness should be followed to promote strong family–school partnerships.4 Although both the AAP and the National Dissemination Center for Children with Disabilities have endorsed a strategy of letters to the school from the pediatrician and parent to improve collaboration with the school, the format and effectiveness has not been studied.5,6

Previous studies have suggested that Attention Deficit Hyperactive Disorder (ADHD) affect individuals across their lifespan, especially in relation to employment. The purpose of this review was to synthesize the results from studies examining the prospective association of ADHD diagnosis in childhood and later education, earnings and employment, compared to children without an ADHD diagnosis. A review protocol was prospectively registered with PROSPERO (ID = CRD42019131634). The findings were reported according to the PRISMA guidelines. The systematic review is based on a structured and preplanned analysis of original prospective longitudinal studies. A total of 2505 potential records were identified, two through backward search. Six papers met the inclusion criteria. One paper was assessed as good, four as fair and one as poor quality. The studies indicated that ADHD diagnosis affected the nature of the individual’s attachment to the labour market across different labour market attachment outcomes. Adults with persisting symptoms, had significantly more problems at work. Even if ADHD symptoms desist in adulthood, the negative impact of earlier ADHD symptoms can still be seen on occupational outcomes. Significantly fewer probands had a Bachelor’s degree compared to controls. Based on one good quality study and four fair quality studies, it is indicated that patients with childhood diagnosed ADHD, generally experience employment of lower quality compared with peers, in relation to income, education and occupational attainment. The overall level of evidence is rated as poor.

Introduction Most research on the association between insomnia and attention deficit hyperactivity disorder (ADHD) has been conducted in children. It has been established that insomnia and ADHD are frequently comorbid, but less is known about whether the association extends to individual ADHD symptoms, or any possible moderators of the association. This study aims to understand whether sex plays a key role in the moderation between insomnia and ADHD in adult military servicemembers, as well as to investigate the relationship between insomnia diagnosis and ADHD symptomology. Methods Data were obtained from the All Army Study of the Army Study to Assess Risk and Resilience in Servicemembers (STARRS; N=21292; age 18-61, 88.24% male). Participants completed a survey that included items assessing insomnia and ADHD status based on DSM-5 criteria. Chi-squares and logistic regression, including interactions with sex, were used to examine the relationship between insomnia diagnosis and ADHD symptomology. Results There is a significant association between insomnia and ADHD diagnoses (r = 0.30; <.0001), consistent with prior literature. There was a small but statistically significant moderation by sex of the relationship between insomnia and ADHD diagnosis (Chi square = 9.43; p = .002). There was also a strong association between insomnia diagnosis and certain symptoms of ADHD: keeping attention on repetitive work (Chi square = 3687.75; <.0001), remembering appointments (Chi square = 3186.735; <.0001), and getting things in order (Chi square = 2756.28; <.0001). Conclusion Insomnia diagnosis was associated with both ADHD diagnosis and symptomology in a nationwide sample of army servicemembers, with the relationships being stronger in males compared to females. These findings highlight the importance of assessment and treatment of insomnia in patients with ADHD given their interrelationships. Support (If Any) This publication is based on data obtained from Army STARRS (Inter-university Consortium for Political and Social Research, University of Michigan-http://doi.org/10.3886/ICPSR35197-v1). KEM’s time was supported by the U.S. Department of Veterans Affairs, Clinical Science Research and Development Service - IK2 CX001874. EAK’s time was supported by the U.S. Department of Veterans Affairs, Rehabilitation Research and Development Service-1IK2 RX001836.

(1) Background: Attention deficit hyperactivity disorder (ADHD) constitutes one of the leading mental health and behavioral disorders in childhood and adolescence. The main objective of this study was to analyze the time trend in the incidence of ADHD diagnoses in Navarre (Spain) from 2003 to 2019 in children and adolescents from 5 to 19 years old. Additionally, the seasonal trends of ADHD incidence and ADHD prevalence were determined. (2) Methods: A population-based observational retrospective study, which included people born between 1991 and 2011 and who attended compulsory education between 2007 and 2017 in Navarre (Spain), was carried out with data from both the Education and Health Department databases. (3) Results: The incidence rate increased from 4.18 cases per 1000 person-years in 2003 to 7.43 cases per 1000 person-years in 2009, before decreasing progressively to 2.1 cases per 1000 person-years in 2019. A peak incidence rate at 7–8 years of age was observed, which is consistent across the study period and for both genders. Males were more than twice as likely to be diagnosed with ADHD than females, with similar time trends in both. A seasonal pattern in ADHD diagnosis was found, with peaks in February–March and the lowest rates in the summer months. Inattentive cases were much more frequent than hyperactive cases, whereas combined cases remained low across the study period. (4) Conclusions: In this age-period-cohort analysis, a clear period and age effect was observed. We found a decreasing trend in the ADHD incidence rate since 2015. Further research is needed to confirm whether a change of trend is occurring globally.

The Northern Finnish Birth Cohort (NFBC) study of Attention Deficit Hyperactivity Disorder (ADHD) was initiated in 2001 to identify risk genes and gene–by–environmental interactions underlying ADHD in a population-based cohort of adolescents who have been followed prospectively since their birth in 1986. The cohort is drawn from northern Finland (pop. 630,000), where the population consists primarily of two groups—an early settlement and a late settlement. The late settlement descended from a small number of families (or founders) who migrated there in the 17th century and expanded locally. The small number of founders in a region that is fairly geographically isolated has resulted in a “population isolate” that is invaluable to genetic studies of single gene (i.e., monogenic) disorders as well as disorders which are thought to be due to several genes (i.e., polygenic). The first set of five papers on the NFBC study of ADHD was recently published in the December 2007 issue of the Journal of the American Academy of Child and Adolescent Psychiatry, and this article will summarize those detailed papers to present the main conclusions drawn about ADHD, based on one of the largest population-based studies of ADHD. Readers are directed to that journal for the methods, detailed results, and a more extensive discussion of the conclusions. The NFBC provides a unique opportunity to examine the developmental trajectory of ADHD in a social setting where medication is not the primary approach to treating ADHD. In fact, very few of the almost 457 (ADHD = 188, Controls = 269) adolescents assessed during the study had ever received any pharmacological treatment. Despite striking differences in treatment approaches between northern Finland and the United States, the clinical characteristics, psychiatric comorbidity, and cognitive deficits associated with ADHD are very much alike. Thus, one of the first major conclusions from this set of studies is that, despite the controversy regarding the validity of the ADHD diagnosis, ADHD is a clearly recognizable condition with a relatively stable prevalence worldwide. Using DSM–IV diagnostic criteria in a population-based cohort of adolescents (ages 16–18) in northern Finland, the prevalence is 8.5%. The clinical characteristics of ADHD in this population are strikingly similar to other ADHD samples drawn around the world with more boys than girls being affected, similar rates of psychiatric comorbidity, and similar rates of persistence of ADHD from childhood into adolescence (~2/3). Description of the ascertainment method and clinical characteristics of the sample were described in the paper by Smalley and colleagues (Smalley et al., 2007). One goal of the study was to examine ADHD symptoms within a large popu-

The Attention Deficit Hyperactivity Disorder (ADHD) is a complex mental health disorder that affects children's learning activities and also their relationships with other children. There are various types of ADHD symptoms, which some of them can be easily recognized as indicators of having ADHD and some of them are difficult to specify as symptoms. Generally, the average age for diagnosing ADHD in children is around seven years old. It is better to early discover these symptoms in children and classify their ADHD types so appropriate and necessary treatments can be effectively applied to improve ADHD children’s behaviors and learning activities. The ADHD diagnosis takes a lot longer for children under six years old. This research aims to improve the ADHD diagnosis and attempt to discover symptoms and behavioral activities that can be the important indicators for classifying ADHD types of the children. Today, many parents share knowledge and experiences about their children in social networks. These shareable contents can be useful to help identify some ADHD symptoms as well as recommended treatments or activities for the children. Thus, in this work, we propose an approach to extract contents from social media (e.g., Facebook contents) in order to discover the information regarding ADHD symptoms and use this extracted information to classify ADHD types based on the behavioral symptoms and activities. To verify our results, we use three sets of criteria from the Diagnostic and Manual of Mental Disorders (DSM) that are used by many ADHD specialists and medical doctors. The three sets of criteria include SNAP-IV, Vanderbilt Assessment Scale, and KUS-SI. Our research goal is to provide a way to help parents to find relevant information and learn about ADHD types and their related symptoms as well as recommended activities for improving ADHD conditions in children.