• Valencià · Castellano  I
• a · A  I
• Accesibility  I
• Map  I
• Search  I
• Staff directory

• ::  Iniciar sessió :: 

A team of researchers from the Universitat Politècnica de València (UPV) and the French National Centre for Scientific Research (CNRS) has developed the world's most advanced software to study the human cerebellum using high-resolution NMR images. Called DeepCeres, this software will help in the research and diagnosis of diseases such as ALS, schizophrenia, autism and Alzheimer's, among others. The work of the Spanish and French researchers has been published in the prestigious journal NeuroImage.

Despite its small size compared to the rest of the brain, the cerebellum contains approximately 50% of all brain neurons and plays a fundamental role in cognitive, emotional and motor functions.

As Sergio Morell-Ortega, a project researcher at the ITACA Institute of the Universitat Politècnica de València, explains, segmentation of the cerebellum has until now been a great challenge due to the complexity of its anatomy and the difficulty of differentiating its structures by means of conventional magnetic resonance imaging. 'DeepCeres overcomes all these challenges and is, today, the most accurate tool in the world for measuring such an important structure of the central nervous system as the cerebellum', emphasises Morell.

High accuracy

The DeepCeres software is capable of measuring 27 structures of the cerebellum. And it stands out above all for improving the precision of segmentation compared to what is achieved with the methods used to date, thanks mainly to the application of different artificial intelligence tools.

'Using standard resonance images of 1 cubic millimetre, these are converted into ultra-high resolution images of 0.125 mm3 using deep neural networks. This allows researchers and healthcare professionals to obtain detailed information about the anatomy of the cerebellum without the need for ultra-high-resolution data in the initial image. It's like going from a black-and-white image to a colour image. There is nothing similar currently and, moreover, it is accessible to the entire scientific community,' adds Professor José Vicente Manjón, the main researcher of the project.

Applications in neuroscience and clinical practice

According to the developers of DeepCeres, the precision in the volumetric quantification of the cerebellum will help in the study of neurological pathologies such as cerebellar ataxia, amyotrophic lateral sclerosis or psychiatric illnesses such as schizophrenia and autism.

'Furthermore, different studies published recently have demonstrated the incidence of the structure of the cerebellum in neurodegenerative diseases such as Alzheimer's', adds Sergio Morell.

15,000 cerebellums in five months

To facilitate its use, the UPV and French CNRS teams have developed an online platform that is accessible to research and medical staff. Since its launch just five months ago, DeepCeres has processed images of nearly 15,000 cerebellums. To date, it has been used by experts from many countries, with the greatest impact in the United States and China.

Researchers from the Research Institute of Industrial Control Systems and Computing and the Applied Mathematics Department at the Universitat Politècnica de València, the Department of Psychobiology at the University of Valencia, the Medical Imaging Department at La Fe University and Polytechnic Hospital and the FISABIO-Príncipe Felipe Biomedical Research Centre Joint Biomedical Imaging Unit have also participated in its development.

Reference

Sergio Morell-Ortega, Marina Ruiz-Perez, Marien Gadea, Roberto Vivo-Hernando, Gregorio Rubio, Fernando Aparici, Maria de la Iglesia-Vaya, Gwenaelle Catheline, Boris Mansencal, Pierrick Coupé, José V. Manjón, DeepCERES: A deep learning method for cerebellar lobule segmentation using ultra-high resolution multimodal MRI, NeuroImage, Volume 308, 2025, 121063, ISSN 1053-8119, https://doi.org/10.1016/j.neuroimage.2025.121063

News UPV TV
Download File