-
Transformer-based normative modelling for anomaly detection of early schizophrenia
Authors:
Pedro F Da Costa,
Jessica Dafflon,
Sergio Leonardo Mendes,
João Ricardo Sato,
M. Jorge Cardoso,
Robert Leech,
Emily JH Jones,
Walter H. L. Pinaya
Abstract:
Despite the impact of psychiatric disorders on clinical health, early-stage diagnosis remains a challenge. Machine learning studies have shown that classifiers tend to be overly narrow in the diagnosis prediction task. The overlap between conditions leads to high heterogeneity among participants that is not adequately captured by classification models. To address this issue, normative approaches h…
▽ More
Despite the impact of psychiatric disorders on clinical health, early-stage diagnosis remains a challenge. Machine learning studies have shown that classifiers tend to be overly narrow in the diagnosis prediction task. The overlap between conditions leads to high heterogeneity among participants that is not adequately captured by classification models. To address this issue, normative approaches have surged as an alternative method. By using a generative model to learn the distribution of healthy brain data patterns, we can identify the presence of pathologies as deviations or outliers from the distribution learned by the model. In particular, deep generative models showed great results as normative models to identify neurological lesions in the brain. However, unlike most neurological lesions, psychiatric disorders present subtle changes widespread in several brain regions, making these alterations challenging to identify. In this work, we evaluate the performance of transformer-based normative models to detect subtle brain changes expressed in adolescents and young adults. We trained our model on 3D MRI scans of neurotypical individuals (N=1,765). Then, we obtained the likelihood of neurotypical controls and psychiatric patients with early-stage schizophrenia from an independent dataset (N=93) from the Human Connectome Project. Using the predicted likelihood of the scans as a proxy for a normative score, we obtained an AUROC of 0.82 when assessing the difference between controls and individuals with early-stage schizophrenia. Our approach surpassed recent normative methods based on brain age and Gaussian Process, showing the promising use of deep generative models to help in individualised analyses.
△ Less
Submitted 8 December, 2022;
originally announced December 2022.
-
GeoSES -- um Índice Socioeconômico para Estudos de Saúde no Brasil
Authors:
Ligia Vizeu Barrozo,
Michel Fornaciali,
Carmen Diva Saldiva de André,
Guilherme Augusto Zimeo Morais,
Giselle Mansur,
William Cabral-Miranda,
João Ricardo Sato,
Edson Amaro Júnior
Abstract:
Objective: to define an index that summarizes the main dimensions of the socioeconomic context for research purposes, evaluation and monitoring health inequalities. Methods: the index was created from the 2010 Brazilian Demographic Census, whose variables selection was guided by theoretical references for health studies, including seven socioeconomic dimensions: education, mobility, poverty, wealt…
▽ More
Objective: to define an index that summarizes the main dimensions of the socioeconomic context for research purposes, evaluation and monitoring health inequalities. Methods: the index was created from the 2010 Brazilian Demographic Census, whose variables selection was guided by theoretical references for health studies, including seven socioeconomic dimensions: education, mobility, poverty, wealth, income, segregation and deprivation of resources and services. The index was developed using principal component analysis, and was evaluated for its construct, content and applicability components. Results: GeoSES-BR dimensions showed good association with HDI-M (above 0.85). The model with the poverty dimension best explained the relative risk of avoidable cause mortality in Brazil. In the intraurban scale, the model with GeoSES-IM was the one that best explained the relative risk of mortality from circulatory system diseases. Conclusion: GeoSES showed significant explanatory potential in the studied scales.
△ Less
Submitted 9 October, 2019;
originally announced October 2019.
-
Discriminating different classes of biological networks by analyzing the graphs spectra distribution
Authors:
Daniel Yasumasa Takahashi,
João Ricardo Sato,
Carlos Eduardo Ferreira,
André Fujita
Abstract:
The brain's structural and functional systems, protein-protein interaction, and gene networks are examples of biological systems that share some features of complex networks, such as highly connected nodes, modularity, and small-world topology. Recent studies indicate that some pathologies present topological network alterations relative to norms seen in the general population. Therefore, methods…
▽ More
The brain's structural and functional systems, protein-protein interaction, and gene networks are examples of biological systems that share some features of complex networks, such as highly connected nodes, modularity, and small-world topology. Recent studies indicate that some pathologies present topological network alterations relative to norms seen in the general population. Therefore, methods to discriminate the processes that generate the different classes of networks (e.g., normal and disease) might be crucial for the diagnosis, prognosis, and treatment of the disease. It is known that several topological properties of a network (graph) can be described by the distribution of the spectrum of its adjacency matrix. Moreover, large networks generated by the same random process have the same spectrum distribution, allowing us to use it as a "fingerprint". Based on this relationship, we introduce and propose the entropy of a graph spectrum to measure the "uncertainty" of a random graph and the Kullback-Leibler and Jensen-Shannon divergences between graph spectra to compare networks. We also introduce general methods for model selection and network model parameter estimation, as well as a statistical procedure to test the nullity of divergence between two classes of complex networks. Finally, we demonstrate the usefulness of the proposed methods by applying them on (1) protein-protein interaction networks of different species and (2) on networks derived from children diagnosed with Attention Deficit Hyperactivity Disorder (ADHD) and typically develo** children. We conclude that scale-free networks best describe all the protein-protein interactions. Also, we show that our proposed measures succeeded in the identification of topological changes in the network while other commonly used measures (number of edges, clustering coefficient, average path length) failed.
△ Less
Submitted 14 August, 2012;
originally announced August 2012.
