Stratification of psychiatric disorders by using network discriminant and clustering analyses

• Grant number: 18/17996-5
• Support Opportunities: Scholarships abroad - Research
• Effective date (Start): July 15, 2019
• Effective date (End): November 14, 2019
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computational Mathematics
• Acordo de Cooperação: European Research Council
• Principal Investigator: André Fujita
• Grantee: André Fujita
• Host Investigator: Gunter Schumann
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Research place: King's College London, England

Abstract

There are at least three main reasons to the lack of reduction of the number of people with alcohol abuse, major depression, and attention deficit hyperactivity disorders (ADHD), namely: (i) the late diagnosis at a time when psychopathology is already well advanced, (ii) a classification based on patient self-report and behavioral observation, which do not reflect the underlying biological mechanisms of these disorders, and (iii) comorbidity (comorbidity between depression and ADHD in adolescents and adults is over 50%; comorbidities between alcohol abuse and ADHD in adolescents and adults are 12.9% and 61-64%, respectively). Thus, to reduce the frequencies of these disorders, we propose to identify biomarkers for early diagnosis, and also to obtain an objective stratification by taking into account the comorbidity. To this end, we will analyze the IMAGEN longitudinal data set (coordinated by Prof. Schumann) composed of over 2,000 individuals. First we will use CEM-Co to cluster the individuals by minimizing the comorbidity effect. Then, to identify discriminative markers and predictors of future psychopathology, we will apply, for each cluster identified by CEM-Co, methods designed specifically for brain networks analyses, namely network discriminant analysis and graph clustering expectation maximization algorithm. Finally, we will assess if brain network structures characterize subtypes within and if they correspond to comorbidity across these disorders. We hope our findings aid in the better comprehension, and early/objective diagnosis of alcohol abuse, depression, and ADHD.

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

GUZMAN, GROVER E. CASTRO; FUJITA, ANDRE. Convolution-based linear discriminant analysis for functional data classification. INFORMATION SCIENCES, v. 581, p. 469-478, DEC 2021. (18/17996-5, 18/21934-5, 20/01479-1)

FUJITA, ANDRE; LIRA, EDUARDO SILVA; SANTOS, SUZANA DE SIQUEIRA; BANDO, SILVIA YUMI; SOARES, GABRIELA ELEUTERIO; TAKAHASHI, DANIEL YASUMASA. A semi-parametric statistical test to compare complex networks. JOURNAL OF COMPLEX NETWORKS, v. 8, n. 2, APR 2020. (18/17996-5, 18/21934-5, 15/21162-4, 15/01587-0)

RAMOS, TAIANE COELHO; BALARDIN, JOANA BISOL; SATO, JOAO RICARDO; FUJITA, ANDRE. Abnormal Cortico-Cerebellar Functional Connectivity in Autism Spectrum Disorder. FRONTIERS IN SYSTEMS NEUROSCIENCE, v. 12, JAN 15 2019. (16/13422-9, 18/17996-5, 13/07375-0, 15/01587-0)