Diagnosis of Attention Deficit / Hyperactivity Disorder with Fourth Wechsler Tool and integrated Version: Ranking of Effective sub Scale with Artificial Neural Network Analysis

Attention Deficit / Hyperactivity Disorder is a disorder with cognitive impairment that standardized tests such as Wechsler use to diagnose these defects. Recently, neural network research has received special attention in psychological research. The aim of this study was to rank Wechsler's effective fourth and integrated subscales in predicting this disorder by neural network method.

The research sample was the students who referred to Shahriar Counseling Centers and were selected by the sampling method and 162 people were selected by clinical interview based on DSM5 criteria in the academic year of 1996-97. Attention Deficit Hyperactivity, Hyperactivity, and Combination Scores were determined by teachers by completing the Swanson and Neupleham questionnaire, and finally the Wechsler Intelligence Test was performed in the fourth version. The data were analyzed by neural network and MATLAB software. 40 subscales were identified as network inputs with 40, an intermediate layer with 42 neurons, and an output layer as the best network pattern.

In the combined type of subscales 1-complete the picture, 2-Capacity of letter-to-song, 3-Similarities, 4- Options of visual vocabulary, 5-Accidental deletion, In the type of hyperactivity 1- complete the picture, 2-Deletion Random, 3- Direct spatial capacity, 4- Cube design process and 5- Similarities, in the type of failure, attention 1-picture concepts 2-Similarities 3- Symbolism, 4- Number-letter sequence 5- Calculations of the most effective subscales Were(P <0.05).

The results showed that neural networking is a method capable of predicting attention deficit / hyperactivity disorder, and some of Wechsler's subscales are superior to others in identifying cognitive impairments.