Attention Deficit Hyperactivity Disorder (ADHD)
Conditions
Keywords
Attention Deficit Disorder with Hyperactivity, Artificial Intelligence, Posture, Machine Learning, child
Brief summary
The aim of this study is to investigate the potential of postural control and plantar pressure data in predicting Attention Deficit Hyperactivity Disorder (ADHD) in middle school students using machine learning methods. A total of 100 students will participate, including those identified with symptoms of ADHD and healthy controls. Participants will undergo non-invasive biomechanical assessments, including pedobarographic foot pressure measurement and mobile posture analysis. Behavioral data will be collected using DSM-IV-based rating scales developed by Atilla Turgay, completed separately by parents, teachers, and caregivers. All data will be used to develop predictive models using algorithms such as random forest, logistic regression, and support vector machines. The study is observational and cross-sectional.
Detailed description
This study aims to predict Attention Deficit Hyperactivity Disorder (ADHD) in middle school children by utilizing pedobarographic and postural parameters in combination with machine learning techniques. The study will include approximately 100 children aged 10-14, consisting of 50 children clinically diagnosed with ADHD and 50 healthy controls. Participants will be selected with permissions from the Eyüpsultan District Directorate of National Education and relevant school administrations in Istanbul. All participants will undergo anthropometric assessments, including height, weight, BMI, waist, neck, and hip circumferences, and skinfold thickness (triceps, subscapular, suprailiac, abdominal). Postural analysis will be conducted using the Mobile Posture Assessment App and the New York Posture Rating Test, while foot posture will be evaluated with the Foot Posture Index (FPI). Static and dynamic balance will be evaluated using the Flamingo Balance Test and the Y Balance Test, respectively. For pedobarographic measurements, the Metisens Static Pedobarography and Stabilometry System will be used. Children will stand barefoot for 20 seconds, and parameters such as plantar pressure distribution, contact area ratios, and Center of Pressure (COP) sway metrics (length, area, AP/ML) will be recorded. In addition, physical activity levels will be assessed using the International Physical Activity Questionnaire - Short Form (IPAQ-SF), which measures walking, moderate, and vigorous activities as well as sedentary time during the previous 7 days. Data will be converted into MET-minutes/week and categorized as Inactive, Minimally Active, or Highly Active according to standardized scoring protocols. ADHD symptoms will be assessed using the DSM-IV-based assessment scale developed by Atilla Turgay, with Parent and Teacher Forms. Data will be analyzed using statistical software (SPSS) to evaluate group differences and data distributions. Subsequently, machine learning and artificial intelligence algorithms will be employed to develop predictive models. Performance metrics such as accuracy, sensitivity, and specificity will be used to evaluate the model's success. This study represents a novel attempt to utilize foot biomechanics and postural parameters as input data for machine learning-based ADHD prediction. It aims to offer an accessible, cost-effective, and objective clinical support tool, potentially contributing to early diagnosis strategies in neurodevelopmental disorders.
Interventions
None listed
Sponsors
Biruni University
Study design
Observational model
CASE_CONTROL
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
10 Years to 14 Years
Healthy volunteers
Yes
Inclusion criteria
* Students attending a middle school located in Eyüpsultan district * Informed consent obtained from their parents * Students enrolled in full-time education * Children with age-appropriate motor development skills.
Exclusion criteria
* Children who have undergone orthopedic interventions due to lower extremity or spinal deformities * Children with congenital or acquired neuromuscular disorders * Children with significant visual or auditory impairments * Children with systemic diseases
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Prediction of ADHD Diagnosis Using Biomechanical and Postural Parameters | Baseline (Single assessment session) | Diagnostic accuracy (sensitivity, specificity, overall accuracy, AUC) of a machine learning model developed using postural, balance, pedobarographic, and anthropometric parameters will be evaluated in distinguishing ADHD and control children. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Postural Assessment - New York Posture Rating Test (NYPRT) | Baseline | Postural alignment will be evaluated with the New York Posture Rating Test. The test yields a total score from 13 to 65 based on visual ratings of multiple body segments in standing posture. Higher scores indicate better postural alignment; lower scores indicate poorer posture. |
| Plantar Pressure Distribution | Baseline | Plantar pressure distribution will be assessed using the Metisens Static Pedobarography system. The outcome is expressed as % of total load distributed across left vs right foot and forefoot, midfoot, hindfoot regions. Higher asymmetry indicates poorer postural control. |
| Foot Posture Assessment - Foot Posture Index (FPI-6) Total Score | Baseline | Foot posture will be assessed using the Foot Posture Index (FPI-6), which evaluates six criteria including talar head palpation, supra/inframalleolar curvature, calcaneal frontal plane position, talonavicular bulging, medial arch height, and forefoot abduction/adduction. Each item is scored on a 5-point scale (-2 to +2), yielding a total score from -12 to +12. Higher positive scores indicate increased pronation. Negative scores indicate supination. Scores around 0 indicate neutral foot posture. |
| Sway Path Length | Baseline | Sway path length of the Center of Pressure (COP) will be measured using the Metisens Stabilometry system during 20-second quiet stance. Greater path length indicates poorer postural stability. |
| Number of Sways | Baseline | Number of COP oscillations will be recorded using the Metisens Stabilometry system during quiet stance. Higher values indicate reduced stability. |
| Anteroposterior Stability Index (APSI) | Baseline | APSI will be calculated using the Metisens Stabilometry system, reflecting variability of COP in the anteroposterior direction. Higher values indicate poorer stability. |
| Postural Assessment via Mobile Posture App | Baseline | Postural alignment will be assessed using the Mobile Posture App in four views (frontal, sagittal-right, sagittal-left, and posterior). The application provides angular deviations in degrees (°) for multiple anatomical landmarks, including: Frontal view: ears, shoulders, hips, and trunk inclination (tilt from vertical midline). Sagittal views (right and left): trunk inclination (forward/backward tilt relative to vertical). Posterior view: ears, shoulders, hips, and trunk inclination (tilt from vertical midline). Higher values indicate greater postural deviation or asymmetry, while lower values indicate closer alignment to neutral posture. |
| Mediolateral Stability Index (MLSI) | Baseline | MLSI will be calculated using the Metisens Stabilometry system, reflecting variability of COP in the mediolateral direction. Higher values indicate poorer stability. |
| Balance Performance - Flamingo Balance Test (Static Balance) | Baseline | Static balance will be assessed using the Flamingo Balance Test, validated for use in children. Participants will perform the test on both the right and left legs. Each trial lasts 30 seconds, and the number of balance errors (e.g., touching the floor, losing balance) will be recorded. The test will be performed three times per leg, and the mean error count will be calculated. Higher number of errors indicates poorer static balance. |
| Stability Index (SI) | Baseline | Overall Stability Index (SI) will be calculated using the Metisens Stabilometry system, combining both AP and ML variability. Higher values indicate poorer overall balance. |
| Physical Activity Level - International Physical Activity Questionnaire, Short Form (IPAQ-SF) | Baseline (within the last 7 days before assessment) | Physical activity levels will be assessed using the International Physical Activity Questionnaire - Short Form (IPAQ-SF). The questionnaire records the frequency (days per week) and duration (minutes per day) of walking, moderate-intensity, and vigorous-intensity physical activity, as well as sedentary time, during the last 7 days. Data are converted into MET-minutes/week scores using standardized coefficients (walking: 3.3 METs, moderate: 4.0 METs, vigorous: 8.0 METs, sitting: 1.5 METs). Higher total MET-min/week indicates higher physical activity level. Participants are categorized as Inactive (Category 1), Minimally Active (Category 2), or Highly Active (Category 3) based on scoring guidelines. |
| ADHD Symptom Severity | Baseline | Measured using the DSM-IV-Based Assessment and Screening Scale for Behavioral Disorders in Children and Adolescents (Atilla Turgay), completed by both parents and teachers. |
| Balance Performance - Y-Balance Test (Dynamic Balance) | Baseline | Dynamic balance will be assessed using the Y-Balance Test. Participants will perform reach trials in the anterior, posteromedial, and posterolateral directions for both right and left legs. Each direction will be tested three times, and the mean reach distance (cm) of the three trials will be recorded for each leg and direction. Longer reach distances indicate better dynamic balance. |
Countries
Turkey (Türkiye)
Contacts
Primary ContactGüzin Kaya Aytutuldu, Asst prof.
Outcome results
None listed