AbstractBackground and Objective. In a previous study, we showed a new EEG processingmethodology called Multi-Scale Ranked Organizing Map/Implicit Function AsSquashing Time (MS-ROM/IFAST) performing an almost perfect distinction betweencomputerized EEG of Italian children with autism spectrum disorder (ASD) andtypically developing children. In this study, we assessed this system indistinguishing ASD subjects from children affected with other neuropsychiatricdisorders (NPD). Methods. At a psychiatric practice in Texas, 20 childrendiagnosed with ASD and 20 children diagnosed with NPD were entered into thestudy. Continuous segments of artifact-free EEG data lasting 10 minutes wereentered in MS-ROM/IFAST. From the new variables created by MS-ROM/IFAST, only12 has been selected according to a correlation criterion. The selectedfeatures represent the input on which supervised machine learning systems (MLS)acted as blind classifiers. Results. The overall predictive capability indistinguishing ASD from other NPD cases ranged from 93% to 97.5%. The resultswere confirmed in further experiments in which Italian and US data have beencombined. In this analysis, the best MLS reached 95.0% global accuracy in 1 outof 3 classes distinction (ASD, NPD, controls). This study demonstrates thevalue of EEG processing with advanced MLS in the differential diagnosis betweenASD and NPD cases. The results were not affected by age, ethnicity andtechnicalities of EEG acquisition, confirming the existence of a specific EEGsignature in ASD cases. To further support these findings, it was decided totest the behavior of already trained neural networks on 10 Italian very youngASD children (25-37 months). In this test, 9 out of 10 cases have beencorrectly recognized as ASD subjects in the best case. Conclusions. Theseresults confirm the possibility of an early automatic autism detection based onstandard EEG.
