Outcome predictors in autism spectrum disorders preschoolers undergoing treatment as usual: insights from an observational study using artificial neural networks

Artificial neural networks reveal sex differences in gene methylation, and connections between maternal risk factors and symptom severity in autism spectrum disorder

Aim and methods: Artificial neural networks were used to unravel connections among blood gene methylation levels, sex, maternal risk factors and symptom severity evaluated using the Autism Diagnostic Observation Schedule 2 (ADOS-2) score in 58 children with autism spectrum disorder (ASD). Results: Methylation levels of MECP2, HTR1A and OXTR genes were connected to females, and those of EN2, BCL2 and RELN genes to males. High gestational weight gain, lack of folic acid supplements, advanced maternal age, preterm birth, low birthweight and living in rural context were the best predictors of a high ADOS-2 score. Conclusion: Artificial neural networks revealed links among ASD maternal risk factors, symptom severity, gene methylation levels and sex differences in methylation that warrant further investigation in ASD.

An Overview of Deep Learning Algorithms and Their Applications in Neuropsychiatry

Deep learning (DL) algorithms have achieved important successes in data analysis tasks, thanks to their capability of revealing complex patterns in data. With the advance of new sensors, data storage, and processing hardware, DL algorithms start dominating various fields including neuropsychiatry. There are many types of DL algorithms for different data types from survey data to functional magnetic resonance imaging scans. Because of limitations in diagnosing, estimating prognosis and treatment response of neuropsychiatric disorders; DL algorithms are becoming promising approaches. In this review, we aim to summarize the most common DL algorithms and their applications in neuropsychiatry and also provide an overview to guide the researchers in choosing the proper DL architecture for their research.

Predictors of Treatment Outcome in Preschoolers with Autism Spectrum Disorder: An Observational Study in the Greater Geneva Area, Switzerland

This study aims to identify predictors of treatment outcome in young children with ASD within a European context, where service provision of intervention remains sporadic. We investigated whether a child's age at baseline, intensity of the intervention provided, type of intervention, child's level of social orienting and cognitive skills at baseline predicted changes in autistic symptoms and cognitive development after 1 year of intervention, in a sample of 60 children with ASD. Our results strongly support early and intensive intervention. We also observed that lower cognitive skills at baseline were related to greater cognitive gains. Finally, we show that a child's interest in social stimuli may contribute to intervention outcome.

Pediatric Acute-Onset Neuropsychiatric Syndrome: A Data Mining Approach to a Very Specific Constellation of Clinical Variables

Objectives: Pediatric acute onset neuropsychiatric syndrome (PANS) is a clinically heterogeneous disorder presenting with: unusually abrupt onset of obsessive compulsive disorder (OCD) or severe eating restrictions, with at least two concomitant cognitive, behavioral, or affective symptoms such as anxiety, obsessive-compulsive behavior, and irritability/depression. This study describes the clinical and laboratory variables of 39 children (13 female and 26 male) with a mean age at recruitment of 8.6 years (standard deviation 3.1). Methods: Using a mathematical approach based on Artificial Neural Networks, the putative associations between PANS working criteria, as defined at the NIH in July 2010 (Swedo et al. 2012), were explored by the Auto Contractive Map (Auto-CM) system, a mapping method able to compute the multidimensional association of strength of each variable with all other variables in predefined dataset. Results: The PANS symptoms were strictly linked to one another on the semantic connectivity map, shaping a central "diamond" encompassing anxiety, irritability/oppositional defiant disorder symptoms, obsessive-compulsive symptoms, behavioral regression, sensory motor abnormalities, school performance deterioration, sleep disturbances, and emotional lability/depression. The semantic connectivity map also showed the aggregation between PANS symptoms and laboratory and clinical variables. In particular, the emotional lability/depression resulted as a highly connected hub linked to autoimmune disease in pregnancy, allergic and atopic disorders, and low Natural Killer percentage. Also anxiety symptoms were shown to be strongly related with recurrent infectious disease remarking the possible role of infections as a risk factor for PANS. Conclusion: Our data mining approach shows a very specific constellation of symptoms having strong links to laboratory and clinical variables consistent with PANS feature.

Towards a Multivariate Biomarker-Based Diagnosis of Autism Spectrum Disorder: Review and Discussion of Recent Advancements

An ever-evolving understanding of autism spectrum disorder (ASD) pathophysiology necessitates that diagnostic standards also evolve from being observation-based to include quantifiable clinical measurements. The multisystem nature of ASD motivates the use of multivariate methods of statistical analysis over common univariate approaches for discovering clinical biomarkers relevant to this goal. In addition to characterization of important behavioral patterns for improving current diagnostic instruments, multivariate analyses to date have allowed for thorough investigation of neuroimaging-based, genetic, and metabolic abnormalities in individuals with ASD. This review highlights current research using multivariate statistical analyses to quantify the value of these behavioral and physiological markers for ASD diagnosis. A detailed discussion of a blood-based diagnostic test for ASD using specific metabolite concentrations is also provided. The advancement of ASD biomarker research promises to provide earlier and more accurate diagnoses of the disorder.

Effects of Probiotic Supplementation on Gastrointestinal, Sensory and Core Symptoms in Autism Spectrum Disorders: A Randomized Controlled Trial

The microbiota-gut-brain axis has been recently recognized as a key modulator of neuropsychiatric health. In this framework, probiotics (recently named "psychobiotics") may modulate brain activity and function, possibly improving the behavioral profiles of children with Autism Spectrum Disorder (ASD). We evaluated the effects of probiotics on autism in a double-blind randomized, placebo-controlled trial of 85 preschoolers with ASD (mean age, 4.2 years; 84% boys). Participants were randomly assigned to probiotics (De Simone Formulation) (n=42) or placebo (n=43) for six months. Sixty-three (74%) children completed the trial. No differences between groups were detected on the primary outcome measure, the Total Autism Diagnostic Observation Schedule - Calibrated Severity Score (ADOS-CSS). An exploratory secondary analysis on subgroups of children with or without Gastrointestinal Symptoms (GI group, n= 30; NGI group, n=55) revealed in the NGI group treated with probiotics a significant decline in ADOS scores as compared to that in the placebo group, with a mean reduction of 0.81 in Total ADOS CSS and of 1.14 in Social-Affect ADOS CSS over six months. In the GI group treated with probiotics we found greater improvements in some GI symptoms, adaptive functioning, and sensory profiles than in the GI group treated with placebo. These results suggest potentially positive effects of probiotics on core autism symptoms in a subset of ASD children independent of the specific intermediation of the probiotic effect on GI symptoms. Further studies are warranted to replicate and extend these promising findings on a wider population with subsets of ASD patients which share targets of intervention on the microbiota-gut-brain axis.

CLINICAL TRIAL REGISTRATION

ClinicalTrials.gov, identifier NCT02708901.

Directional Relationship Between Vitamin D Status and Prediabetes: A New Approach from Artificial Neural Network in a Cohort of Workers with Overweight-Obesity

Objective: Despite the increasing literature on the association of diabetes with inflammation, cardiovascular risk, and vitamin D (25(OH)D) concentrations, strong evidence on the direction of causality among these factors is still lacking. This gap could be addressed by means of artificial neural networks (ANN) analysis.Methods: Retrospective observational study was carried out by means of an innovative data mining analysis-known as auto-contractive map (AutoCM)-and semantic mapping followed by Activation and Competition System on data of workers referring to an occupational-health outpatient clinic. Parameters analyzed included weight, height, waist circumference, body mass index (BMI), percentage of fat mass, glucose, insulin, glycated hemoglobin (HbA1c), creatinine, total cholesterol, low- and high-density lipoprotein cholesterol, triglycerides, uric acid, fibrinogen, homocysteine, C-reactive protein (CRP), diastolic and systolic blood pressure, and 25(OH)D.Results: The study included 309 workers. Of these, 23.6% were overweight, 40.5% were classified into the first class of obesity, 23.3% were in the second class, and 12.6% were in the third class (BMI > 40 kg/m ). All mean biochemical values were in normal range, except for total cholesterol, low- and high-density lipoprotein cholesterol, CRP, and 25(OH)D. HbA1c was between 39 and 46 mmol/mol in 51.78%. 25(OH)D levels were sufficient in only 12.6%. Highest inverse correlation for hyperglycemia onset was with BMI and waist circumference, suggesting a protective role of 25(OH)D against their increase. AutoCM processing and the semantic map evidenced direct association of 25(OH)D with high link strength (0.99) to low CRP levels and low high-density lipoprotein cholesterol levels. Low 25(OH)D led to changes in glucose, which affected metabolic syndrome biomarkers, first of which was homeostatic model assessment index and blood glucose, but not 25(OH)D.Conclusions: The use of ANN suggests a key role of 25(OH)D respect to all considered metabolic parameters in the development of diabetes and evidences a causation between low 25(OH)D and high glucose concentrations.

Artificial neural networks reveal individual differences in metacognitive monitoring of memory

Previous work supports an age-specific impairment for recognition memory of pairs of words and other stimuli. The present study tested the generalization of an associative deficit across word, name, and nonword stimulus types in younger and older adults. Participants completed associative and item memory tests in one of three stimulus conditions and made metacognitive ratings of perceptions of self-efficacy, task success ("postdictions"), strategy success, task effort, difficulty, fatigue, and stamina. Surprisingly, no support was found for an age-related associative deficit on any of the stimulus types. We analyzed our data further using a multilayer perceptron artificial neural network. The network was trained to classify individuals as younger or older and its hidden unit activities were examined to identify data patterns that distinguished younger from older participants. Analysis of hidden unit activities revealed that the network was able to correctly classify by identifying three different clusters of participants, with two qualitatively different groups of older individuals. One cluster of older individuals found the tasks to be relatively easy, they believed they had performed well, and their beliefs were accurate. The other cluster of older individuals found the tasks to be difficult, believed they were performing relatively poorly, yet their beliefs did not map accurately onto their performance. Crucially, data from the associative task were more useful for neural networks to discriminate between younger and older adults than data from the item task. This work underscores the importance of considering both individual and age differences as well as metacognitive responses in the context of associative memory paradigms.

International multicentre randomised controlled trial of improvisational music therapy for children with autism spectrum disorder: TIME-A study

BACKGROUND

Preliminary studies have indicated that music therapy may benefit children with autism spectrum disorders (ASD).

OBJECTIVES

To examine the effects of improvisational music therapy (IMT) on social affect and responsiveness of children with ASD.

DESIGN

International, multicentre, three-arm, single-masked randomised controlled trial, including a National Institute for Health Research (NIHR)-funded centre that recruited in London and the east of England. Randomisation was via a remote service using permuted blocks, stratified by study site.

SETTING

Schools and private, voluntary and state-funded health-care services.

PARTICIPANTS

Children aged between 4 and 7 years with a confirmed diagnosis of ASD and a parent or guardian who provided written informed consent. We excluded children with serious sensory disorder and those who had received music therapy within the past 12 months.

INTERVENTIONS

All parents and children received enhanced standard care (ESC), which involved three 60-minute sessions of advice and support in addition to treatment as usual. In addition, they were randomised to either one (low-frequency) or three (high-frequency) sessions of IMT per week, or to ESC alone, over 5 months in a ratio of 1 : 1 : 2.

MAIN OUTCOME MEASURES

The primary outcome was measured using the social affect score derived from the Autism Diagnostic Observation Schedule (ADOS) at 5 months: higher scores indicated greater impairment. Secondary outcomes included social affect at 12 months and parent-rated social responsiveness at 5 and 12 months (higher scores indicated greater impairment).

RESULTS

A total of 364 participants were randomised between 2011 and 2015. A total of 182 children were allocated to IMT (90 to high-frequency sessions and 92 to low-frequency sessions), and 182 were allocated to ESC alone. A total of 314 (86.3%) of the total sample were followed up at 5 months [165 (90.7%) in the intervention group and 149 (81.9%) in the control group]. Among those randomised to IMT, 171 (94.0%) received it. From baseline to 5 months, mean scores of ADOS social affect decreased from 14.1 to 13.3 in music therapy and from 13.5 to 12.4 in standard care [mean difference: music therapy vs. standard care = 0.06, 95% confidence interval (CI) -0.70 to 0.81], with no significant difference in improvement. There were also no differences in the parent-rated social responsiveness score, which decreased from 96.0 to 89.2 in the music therapy group and from 96.1 to 93.3 in the standard care group over this period (mean difference: music therapy vs. standard care = -3.32, 95% CI -7.56 to 0.91). There were seven admissions to hospital that were unrelated to the study interventions in the two IMT arms compared with 10 unrelated admissions in the ESC group.

CONCLUSIONS

Adding IMT to the treatment received by children with ASD did not improve social affect or parent-assessed social responsiveness.

FUTURE WORK

Other methods for delivering music-focused interventions for children with ASD should be explored.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN78923965.

FUNDING

This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 21, No. 59. See the NIHR Journals Library website for further project information.

Motor Skills as Moderators of Core Symptoms in Autism Spectrum Disorders: Preliminary Data From an Exploratory Analysis With Artificial Neural Networks

Motor disturbances have been widely observed in children with autism spectrum disorder (ASD), and motor problems are currently reported as associated features supporting the diagnosis of ASD in the current Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Studies on this issue reported disturbances in different motor domains, including both gross and fine motor areas as well as coordination, postural control, and standing balance. However, they failed to clearly state whether motor impairments are related to demographical and developmental features of ASD. Both the different methodological approaches assessing motor skills and the heterogeneity in clinical features of participants analyzed have been implicated as contributors to variance in findings. However, the non-linearity of the relationships between variables may account for the inability of the traditional analysis to grasp the core problem suggesting that the "single symptom approach analysis" should be overcome. Artificial neural networks (ANNs) are computational adaptive systems inspired by the functioning processes of the human brain particularly adapted to solving non-linear problems. This study aimed to apply the ANNs to reveal the entire spectrum of the relationship between motor skills and clinical variables. Thirty-two male children with ASD [mean age: 48.5 months (SD: 8.8); age range: 30-60 months] were recruited in a tertiary care university hospital. A multidisciplinary comprehensive diagnostic evaluation was associated with a standardized assessment battery for motor skills, the Peabody Developmental Motor Scale-Second Edition. Exploratory analyses were performed through the ANNs. The findings revealed that poor motor skills were a common clinical feature of preschoolers with ASD, relating both to the high level of repetitive behaviors and to the low level of expressive language. Moreover, unobvious trends among motor, cognitive and social skills have been detected. In conclusion, motor abnormalities in preschoolers with ASD were widespread, and the degree of impairment may inform clinicians about the severity of ASD core symptoms. Understanding motor disturbances in children with ASD may be relevant to clarify neurobiological basis and ultimately to guide the development of tailored treatments.

The role of carboxy-terminal cross-linking telopeptide of type I collagen, dual x-ray absorptiometry bone strain and Romberg test in a new osteoporotic fracture risk evaluation: A proposal from an observational study

The consolidated way of diagnosing and treating osteoporosis in order to prevent fragility fractures has recently been questioned by some papers, which complained of overdiagnosis and consequent overtreatment of this pathology with underestimating other causes of the fragility fractures, like falls. A new clinical approach is proposed for identifying the subgroup of patients prone to fragility fractures. This retrospective observational study was conducted from January to June 2015 at the Nuclear Medicine-Bone Metabolic Unit of the of the Fondazione IRCCS Ca' Granda, Milan, Italy. An Italian population of 125 consecutive postmenopausal women was investigated for bone quantity and bone quality. Patients with neurological diseases regarding balance and vestibular dysfunction, sarcopenia, past or current history of diseases and use of drugs known to affect bone metabolism were excluded. Dual X-ray absorptiometry was used to assess bone quantity (bone mineral density) and bone quality (trabecular bone score and bone strain). Biochemical markers of bone turnover (type I collagen carboxy-terminal telopeptide, alkaline phosphatase, vitamin D) have been measured. Morphometric fractures have been searched by spine radiography. Balance was evaluated by the Romberg test. The data were evaluated with the neural network analysis using the Auto Contractive Map algorithm. The resulting semantic map shows the Minimal Spanning Tree and the Maximally Regular Graph of the interrelations between bone status parameters, balance conditions and fractures of the studied population. A low fracture risk seems to be related to a low carboxy-terminal cross-linking telopeptide of type I collagen level, whereas a positive Romberg test, together with compromised bone trabecular microarchitecture DXA parameters, appears to be strictly connected with fragility fractures. A simple assessment of the risk of fragility fracture is proposed in order to identify those frail patients at risk for osteoporotic fractures, who may have the best benefit from a pharmacological and physiotherapeutic approach.

Applying machine learning to identify autistic adults using imitation: An exploratory study

Autism spectrum condition (ASC) is primarily diagnosed by behavioural symptoms including social, sensory and motor aspects. Although stereotyped, repetitive motor movements are considered during diagnosis, quantitative measures that identify kinematic characteristics in the movement patterns of autistic individuals are poorly studied, preventing advances in understanding the aetiology of motor impairment, or whether a wider range of motor characteristics could be used for diagnosis. The aim of this study was to investigate whether data-driven machine learning based methods could be used to address some fundamental problems with regard to identifying discriminative test conditions and kinematic parameters to classify between ASC and neurotypical controls. Data was based on a previous task where 16 ASC participants and 14 age, IQ matched controls observed then imitated a series of hand movements. 40 kinematic parameters extracted from eight imitation conditions were analysed using machine learning based methods. Two optimal imitation conditions and nine most significant kinematic parameters were identified and compared with some standard attribute evaluators. To our knowledge, this is the first attempt to apply machine learning to kinematic movement parameters measured during imitation of hand movements to investigate the identification of ASC. Although based on a small sample, the work demonstrates the feasibility of applying machine learning methods to analyse high-dimensional data and suggest the potential of machine learning for identifying kinematic biomarkers that could contribute to the diagnostic classification of autism.

Can biological components predict short-term evolution in Autism Spectrum Disorders? A proof-of-concept study

Background

The clinical and pathogenetic heterogeneity of Autism Spectrum Disorders (ASD) limits our ability to predict its short- and long-term evolution. Aim of this naturalistic study was to observe the clinical evolution of very young children with ASD for 12 months after first diagnosis, in order to identify those children who might develop a more positive trajectory and understand how a wide range of biological, clinical and familial factors can influence prognosis.

Methods

Ninety-two children were characterized in terms of family history, prenatal and perinatal variables, and clinical conditions. The sample was divided into four subgroups based on the association of 22 biological, clinical and family history variables. Developmental Quotient (DQ), determined using the Psychoeducational Profile Revised (PEP-R), and symptoms severity, measured by means of the Autism Diagnostic Observation Schedule (ADOS), were evaluated at baseline (T0) and after one year (T1), while receiving treatment as usual. Changes in DQ and ADOS between baseline and follow-up and differences in the short-term evolution of the four subgroups were analyzed.

Results

At T1, 55.4 % of the children demonstrated some gains either of autistic symptomatology or of developmental skills. Mean ADOS score was 13.63 ± 3.67 at T0 and 10.85 ± 4.10 at T1 and mean DQ was 0.64 ± 0.14 at T0 and 0.66 ± 0.15 at T1. At follow-up, 33.7 % of the children showed an improvement in DQ and 37 % presented a less severe symptomatology, measured by means of ADOS. Overall, 15.2 % of the sample displayed major improvements both on developmental quotient and ADOS severity score; these children presented less EEG abnormalities and familial psychiatric disorders. The four subgroups, based on biological, clinical and familial variables, showed differing trends in terms of evolution.

Conclusions

Categorizing very young children with ASD in terms of biological, clinical and familial variables can be instrumental in predicting short-term evolution. This exploratory study highlights the importance of a precise characterization and thorough analysis of interactions among biological and clinical variables, in order to predict the developmental evolution in children with ASD.

Gut to brain interaction in Autism Spectrum Disorders: a randomized controlled trial on the role of probiotics on clinical, biochemical and neurophysiological parameters

BACKGROUND

A high prevalence of a variety of gastrointestinal (GI) symptoms is frequently reported in patients with Autism Spectrum Disorders (ASD). The GI disturbances in ASD might be linked to gut dysbiosis representing the observable phenotype of a "gut-brain axis" disruption. The exploitation of strategies which can restore normal gut microbiota and reduce the gut production and absorption of toxins, such as probiotics addition/supplementation in a diet, may represent a non-pharmacological option in the treatment of GI disturbances in ASD. The aim of this randomized controlled trial is to determine the effects of supplementation with a probiotic mixture (Vivomixx®) in ASD children not only on specific GI symptoms, but also on the core deficits of the disorder, on cognitive and language development, and on brain function and connectivity. An ancillary aim is to evaluate possible effects of probiotic supplementation on urinary concentrations of phthalates (chemical pollutants) which have been previously linked to ASD.

METHODS

A group of 100 preschoolers with ASD will be classified as belonging to a GI group or to a Non-GI (NGI) group on the basis of a symptom severity index specific to GI disorders. In order to obtain four arms, subjects belonging to the two groups (GI and NGI) will be blind randomized 1:1 to regular diet with probiotics or with placebo for 6 months. All participants will be assessed at baseline, after three months and after six months from baseline in order to evaluate the possible changes in: (1) GI symptoms; (2) autism symptoms severity; (3) affective and behavioral comorbid symptoms; (4) plasmatic, urinary and fecal biomarkers related to abnormal intestinal function; (5) neurophysiological patterns.

DISCUSSION

The effects of treatments with probiotics on children with ASD need to be evaluated through rigorous controlled trials. Examining the impact of probiotics not only on clinical but also on neurophysiological patterns, the current trial sets out to provide new insights into the gut-brain connection in ASD patients. Moreover, results could add information to the relationship between phthalates levels, clinical features and neurophysiological patterns in ASD.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02708901 . Retrospectively registered: March 4, 2016.