• The complex Adaptive Systems and Computational Intelligence (CASCI) lab at Binghamton University (State University of New York) is directed Luis M. Rocha, George Klir Professor of Systems Science, who also co­directs the Consortium for Social and Biomedical Complexity (CSBC) with Indiana University. Funded by various agencies such as NIH, NSF, FCT and others (see below), many interdisciplinary projects are pursued at CASCI and CSBC. Our expertise is in complex networks & systems, computational & systems biology, computational social science, computational intelligence, and data science, which we employ to collaborate with domain experts especially in the biomedical, health, and social good domains. Some examples are described below.

    1. Computational Social Science for Public Health Monitoring.

    Our group has been one of the first to use social media data to study collective social behavior in biomedical problems. For instance, it was the first group to use Instagram to build public health monitoring and surveillance tools for discovering drug interactions, adverse reactions, and behavior pathology, focusing on depression and epilepsy. This recent work demonstrates that the universe of social media provides a very promising source of large-scale data that can help with monitoring and understanding public health in ways that have not been hitherto possible. Indeed, given the large number of users, social media data allows us to identify under-reported, population­level pathology. Our group and collaborators have used other sources of large-scale data to tackle problems of biomedical and public-health relevance, such as: electronic health records for studying gender and age biases and comorbidity in medical care; Twitter and Google Trends data to study human reproductive behavior on a global scale; Wikipedia to automatically establish the veracity of online statements; Instagram to predict violence and social unrest situations, etc. Very important for this area is the development of biomedical corpora and dictionaries to mine social media and the literature.

    2. Redundancy and Control in Complex Networks.

    Network science has provided many insights into the organization of complex systems. The success of this approach is its ability to capture the organization of multivariate interactions as networks or graphs without explicit dynamical rules for node variables. As the field matures, however, there is a need to move from understanding to controlling complex systems. This is particularly true in systems biology, medicine, and neuroscience where increasingly accurate models of biochemical regulation have been produced. We have contributed to this goal with two mathematical concepts developed in our group which allow us to remove different forms of redundancy in networks: 1) distance closures, and 2) canalization via schema re-description. The first concept allows us to infer the invariant subgraph that is sufficient to compute all shortest paths in a weighted graph. This has demonstrated that there is massive redundancy in many networks in different domains, whereby most edges in a network are not necessary to compute shortest paths (e.g. 90% of edges in some brain networks). Removing redundant edges can facilitate computation and discovery of important pathways in many applications. The second concept is used to remove redundancy from the logical rules of biochemical regulation models in systems biology, revealing that most variables (e.g. chemical species) rely on a smaller subset of their inputs to be regulated (canalization). The removal of this redundancy simplifies and indeed enables the characterization of control in large biochemical and neural network models, which are otherwise too large to study analytically

    3. Complex Systems Approach to Biomedical Literature Mining.

    Our group has been involved in this field from its very start, having participated successfully in the first four BioCreAtIvE (Critical Assessment for Information Extraction) between 2004 and 2012. Much of the research presently conducted in the biomedical domain relies on the inference of correlations and interactions from data at multiple levels of the biological organization: from the molecular to the social. Because we ultimately want to increase our knowledge of the biochemical, functional and behavioral roles of genes and proteins in organisms, there is a clear need to integrate the associations and interactions among biological entities that have been reported and accumulated in the literature, electronic health records, and experimental databases. Our contributions to the Biomedical literature mining have been the development of novel methods based on network science or bio-inspired computing. This data-driven approach has enabled the automatic discovery, classification and annotation of protein-protein and drug-drug interactions, health risks, pharmacokinetic parameters in drug interaction and adverse reaction studies, protein sequence and structure prediction, functional annotation of transcription data, enzyme annotation publications, etc.

    4. Multivariate Time-Series Analysis and Network Inference.

    In the age of data-science, it is essential to develop methods to infer time-varying data associations such as pairwise variable interactions and subsets of variables that mostly interact with one another (modularity). Our lab’s contribution to the problem of inference on networks and multivariate dynamics has been in the area of spectral methods, statistical inference, and information theory, which has been used to uncover interactions and multiscale modularity in various domains, such as gene regulation, transcriptomics and brain activity time-series data.

    Selected Project Funding in Last few years:

    “Redundancy effects on spread and control in network dynamics: applications in computational biomedicine.” Fundação para a Ciência e Tecnologia, Portugal. PI: Rocha. ID: 2022.09122.PTDC.

    “myAURA: Personalized Web Service for Epilepsy Management”, National Institutes of Health, National Library of Medicine Program, USA. PI: Rocha. ID: 1R01LM012832-01

    “Interdisciplinary Training in Complex Networks and Systems”. National Science Foundation, Research Traineeship (NRT) Program, USA. PI: Rocha, ID: NSF-1735095

    “Factors to Promote Healthy Dialogue and Behaviours in Online School Communities”. Scientific Research and Technological Development in Data Science and Artificial Intelligence in Public Administration, Fundação para a Ciência e Tecnologia, Portugal. co-PI: Rocha. ID: DSAIPA/AI/0102/2019.

    “Identification and Forecasting Hospital Emergency Demand.” Scientific Research and Technological Development in Data Science and Artificial Intelligence in Public Administration, Fundação para a Ciência e Tecnologia, Portugal. Co-I: Rocha. ID: DSAIPA/AI/0087/2018

    “The sperm cell core genetic program: combined clinical and research approach to the diagnosis of male infertility.” Fundação para a Ciência e Tecnologia. Co-PI: Rocha. ID: PTDC/MEC-AND/30221/2017

    “Evidence-based Drug-Interaction Discovery: In-Vivo, In-Vitro and Clinical.” National Institutes of Health, National Library of Medicine Program. PI: Rocha, ID: R01LM011945-01

  • Angelo

    A human interfaces EU prize winning projects on welfare and metal health of call centre work, EU funded

    SUN

    Rebuilding social capital and protecting mental health after a natural disaster – smart community media and youth interventions, EU/MIUR funded

    IDEAS

    Emotional regulation skills interventions in borderline personality disorders, NIHR pilot

    BART II

    Interventions in youth at risk of developing bipolar disorders, MHRN funded