Skip to main content

The University of Bari Aldo Moro

Technical Contact

Giulio Pergola
This email address is being protected from spambots. You need JavaScript enabled to view it.
Giuseppe Stolfa
This email address is being protected from spambots. You need JavaScript enabled to view it.

Procedures

  • Procedures for the acquisition and preprocessing of multimodal neuroimaging. Acquisition of multiparametric MRI protocols (sMRI, task-based and resting-state fMRI) and whole-brain magnetoencephalography recordings with high temporal resolution; harmonized and interoperable preprocessing pipelines according to international standards for data sharing and reproducibility.
  • Procedures for the analysis of multimodal brain connectivity. Extraction of quantitative descriptors of functional and structural connectivity; integration of fMRI and MEG data for the causal analysis of neuronal processes; coupling of brain connectivity with receptor profiles derived from normative PET atlases and with genetic data, based on standard multimodal neuroimaging frameworks (e.g., neuromaps), for the identification of connectivity phenotypes and the stratification of patients into clusters of clinical interest.
  • Procedures for statistical analysis and predictive modeling on a federated platform. Implementation, within the NeuroGenWeb (NeuGeA) platform, of statistical analysis and supervised and unsupervised machine learning procedures for classification, stratification, and predictive modeling from neuroimaging and genetic data. Federated architecture, compliant with GDPR regulations, which allows local processing at individual participating centers and the return of aggregated outputs only, without transfer of raw data.

Instruments

Clinical Imaging Unit

  • 3T Magnetic Resonance Imaging System (San Giovanni Rotondo site). 3-Tesla Magnetic Resonance Imaging system installed at the Neuroradiology Unit of the IRCCS "Casa Sollievo della Sofferenza" in San Giovanni Rotondo (FG); accessible under the current agreements with the Università degli Studi di Bari Aldo Moro. The scanner allows the acquisition of multiparametric protocols for high-field structural and functional neuroimaging; including high-resolution sMRI; task-based and resting-state fMRI sequences. Acquisitions are conducted by technical staff from the IRCCS Neuroradiology Unit; in collaboration with the UNIBA-DiBraiN research team.
  • 3T Siemens MAGNETOM Vida Magnetic Resonance Imaging System (Bari site). Siemens MAGNETOM Vida 3-Tesla Magnetic Resonance Imaging system installed at the Radiology Unit of the Azienda Ospedaliero-Universitaria Consorziale Policlinico di Bari; accessible as part of the research activities of the Università degli Studi di Bari Aldo Moro. The scanner allows the acquisition of multiparametric protocols for high-field structural and functional neuroimaging; including high-resolution sMRI; task-based and resting-state fMRI sequences. The joint use of the two acquisition sites (San Giovanni Rotondo and Bari) optimizes subject recruitment based on residency and magnet availability; ensuring continuity and scalability of neuroimaging studies.
  • Elekta Neuromag TRIUX Magnetoencephalography (MEG) System. Elekta Neuromag TRIUX MEG system owned by the Università degli Studi di Bari Aldo Moro; installed at the university magnetoencephalography laboratory in the Neurological Pavilion of the A.O.U. Consorziale Policlinico di Bari. 306-channel equipment based on highly sensitive SQUID superconducting sensors (magnetometers and planar gradiometers); with whole-brain coverage; sampling frequency up to 1000 Hz and continuous head position tracking with motion correction. The system is installed inside a shielded room to reduce environmental interference.

Data Storage and Deep Learning

  • UNIBA HPC Cluster – ReCaS-Bari. The Università degli Studi di Bari "Aldo Moro" has direct access to the ReCaS-Bari Data Center; a Tier-2 HPC infrastructure co-managed with the INFN. The computing system consists of 128 servers (AMD processors; 92 allocated to UNIBA) for a total of 8;192 CPU cores; supported by an HPC cluster of 20 servers equipped with NVIDIA K40 GPUs and interconnected via a low-latency InfiniBand network. Storage capacity amounts to 3;552 TB on disk (2;400 TB UNIBA allocation) and 2;500 TB on tape; accessible at 10 Gbps via GPFS distributed file system. The infrastructure supports large-scale ML/DL training workloads and is connected to the GARR national research network at 10 Gbps.
  • GDPR-compliant Server for the NeuroGenWeb (NeuGeA) Platform. Computational resources dedicated to the NeuroGenWeb (NeuGeA) platform are hosted within the ReCaS-Bari infrastructure; ensuring full compliance with the European General Data Protection Regulation (GDPR). Data sovereignty is guaranteed through on-premises storage within the institutional Data Center of the University of Bari. The platform supports integrated remote analysis of neuroimaging and genetic data without raw data leaving the institutional perimeter; using isolated virtual environments delivered via OpenStack-based cloud provisioning.
  • Infrastructure for Cross-Site Federated Training. The ReCaS-Bari infrastructure supports federated learning architectures through its distributed cloud computing environment; with over 700 virtual machines operating simultaneously on isolated nodes. The 10 Gbps internal and external network backbone (GARR) enables efficient synchronization of model parameters between research sites; without the transfer of raw data. Job management via Torque/MAUI (PBS) and the GPFS distributed file system allow the execution of containerized federated training pipelines on local data copies; with exclusive exchange of model updates between nodes; in full compliance with cross-institutional data governance policies.

Partners