Biomedical Informatics: Data Mining and Simulation
Objectives and line description: The MIDAS team has large experience in Big Data preprocessing and analytics both with structure and non-structured datasets. In the last years they have also proposed new techniques for remote sensing data analytics, understanding and exploitation.
Their goal is the analysis of medical information to extract knowledge that can be the basis for designing and implementing the smarts objects to support health intelligent applications. This can be decomposed in the following lines:
- EHR (electronic health records) analysis and understanding, which involves natural language processing, indexing and knowledge discovery. The integration of the knowledge obtained from the analysis together with other information contained in the EHR can help identifying subjects, for example, for clinical trials, finding common patterns of behavior of drugs and treatment.
- MEG data analysis. Big data analytics is being applied to predict biomarkers for early stages of Alzheimer and Parkinson, as well as treatment progression identifiers for other pathologies (for instance those derived from TBI – traumatic brain injury). The analysis is performed based on Magnetoencephalography records integrated with psychological tests and clinical data.
- Postprocessing of MEG records: With the idea to transform MEG sensor data into an accurate estimation of inner deep sources. We apply soft computing techniques on hybrid optimization problems to fit candidate dipole distributions into the externally recorded sensor space. This research lean complement the previous, providing information on deep areas of the brain
- Techniques for the analysis of gene expression: Currently, these research activities cover the analysis of patient’s response to different treatments in oncology (medulloblastoma, lung and breast cancers, in particular), as well as survival and recurrence of various tumors, prevalence of neurodegenerative diseases and dementias.
- Complex bio-/neuro-simulations: In the field of simulation, we conduct several research activities in modeling complex biological processes, such as neurotransmitter release, mechanical-electrophysiological coupling in neuron transmission (to model traumatic brain injuries) or neurite growing and neurogenesis (using Monte Carlo simulations and finite element/differences solvers).
- Medical image processing, analysis and understanding: Development of image processing algorithms and tools to assist in medical diagnosis and prognosis of diseases, as well as in monitoring illnesses evolution. Annotation of images for integration with other health records, allowing search and retrieval of knowledge for supporting medical research and practice.
- Complex data visualization and interaction: Neuro- and bio-data navigation tools and representation techniques combined with interactive data analysis techniques. Interactive steering of supervised and semi-supervised data exploration and analysis.
- Neuroinformatics tools and services: Technology and toolkits to provide microscopy image storage, indexing and processing. In addition, neuroinformatics databasing and atlasing and laboratory information management systems are also part of the catalog of user-centric services and products provided by MIDAS Lab.
- Cognitive and Computational Neuroscience
- Biological Networks
- Computational System Biology
- Cajal Cortical Circuits
Contact: Ernestina Menasalvas