Join us at the Annual Biomedical Engineering Conference in Linköping on May 21st !
This conference is an excellent opportunity for researchers, companies, organisations, students, and clinicians to connect, collaborate, and explore the latest advancements in BME.
liu.se/en/research/...
networking opportunities, and a platform to showcase innovative contributions to the field.
liu.se/en/research/...
collaborate, and explore the latest advancements in BME technologies.
📅 Date: 21 May 2025
📍 Location: Ljusgården, Northern entrance, Linköping University Hospital
Hosted by the Department of Biomedical Engineering at Linköping University, this event offers insightful discussions,
Join us at the 6th Annual Biomedical Engineering Conference!
We are pleased to invite you to BME@LiU, where we highlight research in the field of biomedical engineering. This conference is an excellent opportunity for researchers, companies, organisations, students, and clinicians to connect,
Open PhD position in Neuroengineering at the Department of Biomedical Engineering, Linköping University. Deadline for application, March 1st.
Swedish: web103.reachmee.com/ext/I011/853...
English: web103.reachmee.com/ext/I011/853...
We are prepared to start several new master theses on deep learning for medical images. The graphics cards are hungry for data to process.
New paper by Elisabeth Klint Johan Richter Peter Milos Martin Hallbeck Karin Wårdell
Klint, E., Richter, J., Milos, P., Hallbeck, M., & Wårdell, K. (2024). In situ optical feedback in brain tumor biopsy: A multiparametric analysis. Neuro-Oncology Advances
doi.org/10.1093/noaj...
interpreting the results. In conclusion, this workflow including repeated MRI measurements could help detect changes in CBF between different measurement days and complement other conventional monitoring techniques in the NICU.
territory atlas. The test-retest data showed small variations (4.4 ml/min/100g between sessions), and the longitudinal measurement resulted in low CBF variability over 12 days. Despite the greater complexity of clinical data, the quantification and chosen visualization tools proved helpful in
To ensure accurate day-to-day comparisons between the repeated measurements, the selection of processing and analysis methods aimed to obtain CBF maps in absolute units of ml/min/100g. These CBF maps were quantified using both the FMRIB Software Library and an openly available flow
the workflow was implemented and tested using acquired test-retest data and longitudinal data from two healthy participants. Subsequently, the workflow was tested in clinical practice on an intubated and ventilated patient monitored in the NICU after suffering a subarachnoid hemorrhage.
Therefore, this work aims to develop and implement a methodological workflow for the acquisition, analysis, absolute quantification, and visualization of longitudinal arterial spin labeling MRI measurements acquired in the clinical NICU setting. At this initial stage,
enabling longitudinal CBF measurements while eliminating medical transportation risks. Arterial spin labeling is a subtraction-based MRI technique that can measure CBF globally in the brain without the use of contrast agents, and thus is suitable for repeated measurements over short time periods.
Including absolute MRI measurements of CBF in the NICU monitoring protocol could add valuable information and potentially improve patient outcomes. This is particularly feasible at Linköping University Hospital, which uniquely has an MRI scanner located in the NICU,
New paper by Sofie Tapper, Anders Tisell, Hillman Jan and Karin Wårdell
Cerebral blood flow (CBF) is carefully monitored in the Neurointensive Care Unit (NICU) to prevent secondary brain insults in patients who have suffered subarachnoid hemorrhage.
doi.org/10.1371/jour...
An important point is that common metrics for evaluating quality and diversity of synthetic images do not consider memorization, and memorization therefore needs to be tested separately.
iopscience.iop.org/article/10.1...
demonstrated that diffusion models are more likely to memorize the training images, compared to generative adversarial networks (GANs). Researchers who want to share synthetic medical images should therefore be careful before doing so.
Diffusion models have become popular for synthesizing realistic medical images. As synthetic medical images do not belong to a specific person, GDPR does not apply and these synthetic images can therefore be shared freely. @muhamadusman.bsky.social , Wuhao WANG and @wandedob.bsky.social have now
On November 27, at 9:00, Iulian Emil Tampu will defend his PhD thesis
Deep learning for medical image analysis in cancer diagnosis
liu.diva-portal.org/smash/record...
Location: Belladonna, Building 511, Campus US, Linköping, Sweden
