Applied Machine Learning Method to Predict Children With ADHD Using Prefrontal Cortex Activity: A Multicenter Study in Japan.

Objective: To establish valid, objective biomarkers for ADHD using machine learning. Method: Machine learning was used to predict disorder severity from new brain function data, using a support vector machine (SVM). A multicenter approach was used to collect data for machine learning training, including behavioral and physiological indicators, age, and reverse Stroop task (RST) data from 108 children with ADHD and 108 typically developing (TD) children. Near-infrared spectroscopy (NIRS) was used to quantify change in prefrontal cortex oxygenated hemoglobin during RST. Verification data were from 62 children with ADHD and 37 TD children from six facilities in Japan. Results: The SVM general performance results showed sensitivity of 88.71%, specificity of 83.78%, and an overall discrimination rate of 86.25%. Conclusion: A SVM using an objective index from RST may be useful as an auxiliary biomarker for diagnosis for children with ADHD.

Age-related differences in frontal lobe function in children with ADHD.

BACKGROUND: The neural correlates of executive function disorders are thought to be predominantly localized within the prefrontal cortex (PFC). However, no study to date has investigated changes in this system across different age groups in children with attention deficit hyperactivity disorder (ADHD). Thus, this study aimed to explore changes in PFC function in children with ADHD. METHODS: Study participants included typically developing (TD) children (n = 140) and children with ADHD (n = 67) of primary school age. Behavioral executive functions and their neural basis were evaluated between the TD children and children with ADHD and also across different age periods (younger and older children). To examine executive function, inhibitory control was assessed using the reverse Stroop task, and PFC near-infrared spectroscopic measurements were used to investigate the neural mechanisms involved. RESULTS: Both ADHD symptoms and the ability to inhibit color interference improved with age. Compared to TD children, children with ADHD demonstrated decreased activation of the right and middle PFC across all age groups. Interestingly, the left PFC appeared to play a compensatory role. CONCLUSION: Children with ADHD exhibited changes in PFC function that varied with age. Longitudinal studies are required to assess the potential of using PFC function as an early biomarker of ADHD.

Lateralized frontal activity for Japanese phonological processing during child development.

Phonological awareness is essential for reading, and is common to all language systems, including alphabetic languages and Japanese. This cognitive factor develops during childhood, and is thought to be associated with shifts in brain activity. However, the nature of this neurobiological developmental shift is unclear for speakers of Japanese, which is not an alphabetical language. The present study aimed to reveal a shift in brain functions for processing phonological information in native-born Japanese children. We conducted a phonological awareness task and examined hemodynamic activity in 103 children aged 7-12 years. While younger children made mistakes and needed more time to sort phonological information in reverse order, older children completed the task quickly and accurately. Additionally, younger children exhibited increased activity in the bilateral dorsolateral prefrontal cortex (DLPFC), which may be evidence of immature phonological processing skills. Older children exhibited dominant activity in the left compared with the right DLPFC, suggesting that they had already acquired phonological processing skills. We also found significant effects of age and lateralized activity on behavioral performance. During earlier stages of development, the degree of left lateralization appears to have a smaller effect on behavioral performance. Conversely, in later stages of development, the degree of left lateralization appears to have a stronger influence on behavioral performance. These initial findings regarding a neurobiological developmental shift in Japanese speakers suggest that common brain regions play a critical role in the development of phonological processing skills among different languages systems, such as Japanese and alphabetical languages.

Facial identity recognition in children with autism spectrum disorders revealed by P300 analysis: a preliminary study.

BACKGROUND: To reveal the neural substrate of communication difficulties in children with autism spectrum disorders (ASDs), we investigated the P300 component of event-related potentials (ERPs) as represented by the average of electroencephalography findings time-locked to events and behavior. Because the P300 amplitude influences attentional resource allocation during discrimination, the component elicited during perception of known and unknown faces should indicate familiarity processing. METHODS AND RESULTS: Nine typically developing children (TD) and nine children with ASDs participated in this study (Experiment 1). The P300 amplitude in TD children was significantly larger during familiar face perception than during unfamiliar face perception (p<0.01). However, there was no evidence of familiarity effect in children with ASDs. In three children with ASDs, we also assessed the P300 amplitude during perception of a therapist's face one month before (baseline), a few days before and after social skills training (SST) sessions (Experiment 2). To evaluate the effect of familiarity on facial identity processing, we analyzed the therapist/unknown ratio of P300 amplitudes related to the face discrimination task as an index. The ratio was larger after SST sessions than before, but there was no difference in the ratio between baseline and before SST assessments. CONCLUSION: The P300 might be influenced by attentional resource allocation depending on the stage of learning face identification in children with ASDs. We speculate that this approach to evaluating brain responses during facial identity recognition could be used as a tool to clarify children's communication difficulties.

[Effectiveness of social skills training for children with developmental disorders: behavioral analysis using a two-dimensional motion capture system].

The current study sought to develop a new behavioral analysis methods to evaluate the effects of social skills training (SST). SST is known to be an effective method to improve the social skills of children with behavioral problems. However, current evaluation methods involve behavioral rating scales that are heavily dependent on evaluators' particular experiences they have had. To quantitatively examine the behavioral effects of SST, we examined subjects' head-movements related to social behavior, using a two-dimensional motion capture system (Kissei Comtec, Japan). Four children (three male, one female, 7-8 years of age) with pervasive developmental disorder (PDD) or attention deficit/hyperactivity disorder (AD/HD) participated in 16 sessions of SST. Before and after SST, head-coordinates on a two-dimensional plane were calculated using their behavior during a pair task, measured by four digital cameras. After SST, the number of communication behaviors was increased compared to before SST. In addition, children looked longer at another child within 30 degrees of the central visual field. Time-series analysis of the visual field during the detection of another child revealed significant auto-correlation from about -1.12 second. before to the beginning of communication behavior (p<0.05). The results suggested that our method can provide a quantitative index of characteristics related to skilled social behaviors. We conclude that a two-dimensional motion capture system would be useful for visualization of the interventional effects of SST, which would supplement assessments by the conventional observational strategies.

[Developmental changes of rapid automatized naming and Hiragana reading of Japanese in elementary-school children].

Two hundred and seven Japanese elementary school children aged from 6 (Grade 1) to 12 (Grade 6) years old were tested for their abilities to name numbers and pictured objects along with reading Hiragana characters and words. These children all showed typical development and their classroom teachers judged that they were not having any problems with reading or writing. The children were randomly divided into two groups, the first group was assigned to two naming tasks;the rapid automatized naming (RAN) of "numbers" and "pictured objects," the second group was assigned to two rapid alternative stimulus (RAS) naming tasks using numbers and pictured objects. All children were asked to perform two reading tasks that were written in Hiragana script: single mora reading task and four syllable word reading task. The total articulation time for naming and reading and performance in terms of accuracy were measured for each task. Developmental changes in these variables were evaluated. The articulation time was significantly longer for the first graders, and it gradually shortened as they moved through to the upper grades in all tasks. The articulation time reached a plateau in the 5th grade for the number naming, while gradual change continued after drastic change in the lower grades for the pictured object naming. The articulation times for the single mora reading and RAN of numbers correlated strongly. The articulation time for the RAS naming was significantly longer compared to that for the RAN, though there were very few errors. The RAS naming showed the highest correlation with the four syllable word reading. This study demonstrated that the performance in rapid automatized naming of numbers and pictures were closely related with performance on reading tasks. Thus Japanese children with reading disorders such as developmental dyslexia should also be evaluated for rapid automatized naming.

[Relationship between clinical symptoms and Hiragana reading ability in children with difficulties in reading and writing:usefulness of a clinical-symptoms-checklist].

We investigated the clinical symptoms of children with developmental dyslexia (DD) and evaluated the relationship between these symptoms and their Hiragana reading abilities. In order to detect the clinical symptoms of DD, we newly developed a clinical-symptoms-checklist (CL), which consisted of a total of 30 yes/no questions regarding symptoms linked to reading (15 questions) and writing (15 questions). Subjects were 98 Japanese school grade (1 to 9) children, aged 6 to 15 years old, with normal intelligence confirmed by the Wechsler Intelligence Test for Children (WISC-Ill) and they were divided into 2 groups according to their diagnosis. Twenty four children diagnosed as developmental dyslexia consisted the DD group, and the remaining 74 children were grouped in the non-DD group. CL showed significant construct validity (p<0.05) and inner consistency (reading: a =0.82, writing: a =0.72) after deleting two questions from the originals. The number of questions checked in the CL reading subcategory significantly correlated with the Hiragana reading ability of articulation time in all Hiragana reading tasks (p<0.001). More severe clinical symptoms and lower reading ability were observed in the DD group compared to the non-DD group. Receiver Operating Characteristics (ROC) analysis indicated that these two groups could be discriminated by the CL and the results of the reading task, and both sensitivity and specificity rate were approximately 80%. It was suggested that 7 or more positive checks in the CL and 2 or more abnormal scores in the reading tasks might discriminate DD from other conditions which cause difficulties in reading and writing in Japanese children.

[Developmental changes in reading ability of Japanese elementary school children--analysis of 4 kana reading tasks].

Five hundred and twenty-eight Japanese elementary school children aged from 6 (Grade 1) to 12 (Grade 6) were tested for their abilities to read Hiragana characters, words, and short sentences. They were typically developing children whom the classroom teachers judged to have no problems with reading and writing in Japanese. Each child was asked to read four tasks which were written in Hiragana script: single mora reading task, four syllable non-word reading task, four syllable word reading task, and short sentence reading task. The total articulation time for reading and performance in terms of accuracy were measured for each task. Developmental changes in these variables were evaluated. The articulation time was significantly longer for the first graders, and it gradually shortened as they moved through to the upper grades in all tasks. The articulation time reached a plateau in the 4th grade for the four syllable word and short sentence reading tasks, while it did so for the single mora and four syllable non-word reading tasks in the 5th grade. The articulation times for the four syllable word and short sentence reading tasks correlated strongly. There were very few clear errors for all tasks, and the number of such errors significantly changed between the school grades only for the single mora and four syllable word reading tasks. It was noted that more than half of the children read the beginning portion of the word or phrase twice or more, in order to read it accurately, and developmental changes were also seen in this pattern of reading. This study revealed that the combination of these reading tasks may function as a screening test for reading disorders such as developmental dyslexia in children below the age of ten or eleven years old.

Development of carefulness and quickness observed for a tray-carrying task.

The purpose of this study was to investigate the developmental change in performance of groups of children instructed to behave carefully and quickly in a tray-carrying task. The subjects were 69 nonhandicapped children from a kindergarten, ages 3-4 to 6 years, and 20 adult students who volunteered. Subjects were instructed to carry as fast as they could a tray with a glass of water for 3 m without a spill. The amounts of water spilled and the times taken were measured. The amounts of water spilled were hardly different within groups of children of the same ages or between age groups. But the times were different among groups: the younger ones took longer, and the difference in time between tray-carrying and normal walking was greatest for the youngest group. Children could carry the tray as carefully as adults but could not do so quickly. The nature of the representation of the instruction in relation to the behavior of children and the difference in strategy to guarantee carefulness between children and the adults are discussed.