High-Performance Computing in Functional Biomarker Analysis Applied to Developmental Dyslexia Diagnosis and Prediction (HPC-BODDY)

Funded Research Project – PGC2018-098813-B-C32. Plan Nacional de I+D+i 2018. Ministerio de Ciencia, Innovación y Universidades

Contact: Andrés Ortiz (aortiz at ic.uma.es)


The search of effective and practical solutions for real-life problems usually requires the exploitation of large amounts of data. Fortunately, current information systems allow collecting detailed data from many diverse information sources. Nevertheless, exploiting such large amounts of data transforming them from simple raw facts to significant knowledge is a challenging task, especially in biomedical-related problems. This is the case of the diagnosis and prognosis of learning difficulties such as developmental dyslexia (DD), which presents a prevalence of between 5% and 12% and may help to explain a good part of school failure cases. The diagnosis of DD and the exploration of its neural bases could allow the development of specific, adaptive and individualized intervention methods but accurate results require not only conductual and neuropsychological data but also the use of biomedical signals.
This project can be seen as a natural extension of our previous projects. Thus, in the PSI2015-65848-R project a big data platform (LEEDUCA) has been developed where a very large amount of neuropsychological test can be conducted, stored and processed. This has then been reinforced, funded by the UNMA15-CE-3657 project, with the purchase of last generation active electroencephalography (EEG) and Functional Near Infrarred Spectroscopy (fNIRS) equipments as well as a high-performance computing (HPC) platform. As a consequence, the LEEDUCA database, with the support of the regional government (Junta de Andalucía), currently accounts test results
from more than 4000 students (4, 5, 6 and 7 year-old) at different schools, and EEG and fNIRS data have been employed to build detailed and multimodal models of DD by means of machine learning.
The current project aims to extract the knowledge contained in biomedical signals, namely EEG and fNIRS. This further requires the development of specific feature extraction and classification techniques along with the construction of models of DD based on signal processing, statistical learning and deep learning methods to: i.) leverage and outperform the accuracy provided by current neuropsychological tests in the diagnosis and prognosis of DD, and ii.) obtain DD models based on all available data
(neuropsychological, EEG and fNIRs) which provide insights into the biological causes of DD and brain processes involved during the learning-to-read process. Additionally, due to the vast amount of multi-modal data to be processed and the need to analyze a myriad of alternatives, high-performance computing methods will have to be employed.
The multidisciplinary profile of the research team of this project makes it possible the exploration of new alternatives to tackle the learning difficulties and for the design of effective intervention tools with a strong neural basis. The methods developed in this project will be transferred to the society through schools, facilitating their clinical use and moving towards more individualized interventions, representing a turning point in the treatment of learning difficulties and its corresponding socioeconomic implications.

Current Publications

  • León, J., Escobar, J.J., Ortiz, A., Ortega, J., Martín-Smith, P., Q. Gan, J., Damas, M., González, J.:”Deep learning for EEG-based Motor Imagery classification: accuracy-cost trade-off”. PLOS ONE, 2020.
  • Górriz, J.M., Ramírez, Ortiz, A., Martínez-Murcia, F.J., Segovia, F. et al.:”Artificial intelligence within the interplay between natural and artificial Computation: advances in data science, trends and applications”. Neurocomputing, 2020.
  • Martínez-Murcia, F.J., Ortiz, A., Górriz, J.M., Ramírez, López-Pérez, Pedro J. and Luque. “EEG Connectivity Analysis Using Denoising Autoencoders for the Detection of Dyslexia”. Intenational Journal of Neural Systems. 2020
  • Ortiz, A., Martínez-Murcia, F.J., Luque, J., Giménez, A., Morales-Ortega, R., Ortega, J. “Dyslexia diagnosis by EEG temporal and spectral descriptors: an anomaly detection approach”. International Journal of Neural Systems. 2020


  1. Ortiz, A., Martínez-Murcia, F.J., Luque, J., Sanchez, A.:»Dyslexia detection from EEG signals using SSA component correlation and Convolutional Neural Networks». The 15th International Conference on Hybrid Artificial Intelligence Systems (HAIS). 2020
  2. Martínez-Murcia, F.J., Ortiz, A., Lopez-Zamora, M., Luque, J., Giménez. A.:»A Neural Approach to Ordinal Regression for the Preventive Assessment of Developmental Dyslexia». The 15th International Conference on Hybrid Artificial Intelligence Systems (HAIS). 2020
  3. Luque, J.L., Ortiz, A., Cobo, M.A., Giménez, A., López-Pérez, P.J., López-Zamora, M., Sánchez, A.:»A Multiple Risk Factor study on the Prediction of Reading Learning Difficulties for Spanish Children Ages 4-7 and 5-8″. IWORDD 2019 International Workshop on Reading and Developmental Dyslexia
  4. Martínez-Murcia, F.J., Ortiz, A., Morales-Ortega, R., López, J.L., Luque, J.L., Castillo-Barnés, D., Ortega, J.:»Periodogram Connectivity of EEG Signals for the Detection of Dyslexia». 8th Work-Conference on the Interplay between. Natural and Artificial Computation (IWINAC 2019)
  5. Ortiz, A., López, P.J., Juque, J.L., Martínez-Murcia, F.J., Aquino-Britez, D., Ortega, J.:»An Anomaly Detection Approach for Dyslexia Diagnosis Using EEG Signals». 8th Work-Conference on the Interplay between. Natural and Artificial Computation (IWINAC 2019)
  6. León, J., Ortega, J., Ortiz, A.:»Convolutional Neural Networks and Feature Selection for BCI with Multiresolution Analysis». International Work-Conference on Artificial and Natural Neural Networks (IWANN 2019)