World J Psychiatry Ment Health Res | Volume 3, Issue 1 | Research Article | Open Access

Functional Neuroimages Analyzed by means of Artificial Neural Networks Can Help Psychiatrists to Diagnose Schizophrenia

Mario Enrique Molina1*, Manassero Carlos2, Cornejo Luciana1, Kotlik Silvina3, De Monte María Inés4 and Manzano Julieta1

1University of Mendoza, Argentina
2Department of Pharmacy, Carlos Pereyra School Hospital, Argentina
3Department of Psychiatry, Carlos Pereyra School Hospital, Argentina
4Department of Neurology, Carlos Pereyra School Hospital, Argentina

*Correspondance to: Mario Enrique Molina 

Fulltext PDF

Abstract

Introduction: Schizophrenia-related spectrum disorders show functional images that permit to infer the existence of underlying neurobiological features: Inter-hemispheric asymmetries in metabolic activity, abnormal regional glucose absorption in prefrontal and temporo-occipital areas, basal ganglia, and claustrum, among others. We developed Artificial Neural Network (ANN) aids to help us detect cerebral absorption patterns of glucose in subjects diagnosed with Schizophrenia.
Methods: Positron Emission Tomography (PET) images of seven patients diagnosed with schizophrenia, with predominantly negative symptoms (Schz group) and seven matching control volunteers (Control group) were accessed and processed by means of the developed ANN software. PET Scan: PET studies were acquired under resting conditions in Foundation School of Nuclear Medicine, Mendoza, Argentina.
Results: The diagnostic aid application was tested with ten randomized sets, formed by the two experimental groups, yielding highly compatible results with diagnostics obtained by trained psychiatrists. Test sensitivity=100%. Test specificity=95%. Conclusions: At present, neuroimaging methods are used with reluctance for the purpose of psychiatric diagnostics, as since until now it had been difficult to detect specific metabolic absorption patterns for different nosological entities. According to our results, the ANN algorithms appear as promising tools for helping diagnose schizophrenia, and presumably, for diagnosing other neuropsychiatric disorders.

Citation:

Molina ME, Carlos M, Luciana C, Silvina K, María Inés DM, Julieta M. Functional Neuroimages Analyzed by means of Artificial Neural Networks Can Help Psychiatrists to Diagnose Schizophrenia. World J Psychiatry Ment Health Res. 2019; 3(1): 1019.

Subscribe to Our Newsletter